Systematic review and meta-analysis of the safety and efficacy of recombinant human bone morphogenetic protein-2 (rhBMP-2) for spinal fusion
Jennifer V E Brown Morag K Heirs Julian P T Higgins Richard J Mannion Mark A Rodgers Charlotte C Seneschall Mark C Simmonds Lesley A Stewart Kath Wright
Centre for Reviews and Dissemination (CRD) University of York YO10 5DD United Kingdom
May 2013
Acknowledgements We thank the following people for their help and support in this research project: We are indebted to Vanda Castle for coordinating telephone conferences and team meetings and for typing the adverse events tables from clinical study reports and to Beth Hodshon of the YODA team for her professional and efficient handling of our queries to Medtronic. We thank S D Glassman and Leah Carreon for providing of IPD from their trial. We are also grateful to Pamela Hayden, Director, Research & Quality Improvement at the North American Spine Society for sharing our call for evidence with society members and to Mark Schoene, editor of The BackLetter for publishing it and to everyone who responded. We thank Dr Eugene Carragee for providing one of his manuscripts prior to publication for inclusion in our safety analyses.
ii Authors' Contributions All authors were actively involved in this project.
Jennifer Brown was involved in eligibility screening, data extraction, assessment of risk of bias and mapping of the adverse events publications. She contributed to the writing of the report.
Morag Heirs contributed to the management and checking of data provided by Medtronic. She was involved in the eligibility screening, data extraction and risk of bias assessment processes. She contributed to the writing of the report.
Julian Higgins contributed to development of the proposal and protocol, undertook some of the analyses for safety, provided methodological advice and contributed to writing the report.
Richard Mannion contributed to development of the proposal and protocol, provided advice on clinical significance and context and contributed to writing the report.
Mark Rodgers contributed to development of the protocol. He was involved in eligibility screening, data extraction, and assessment of risk of bias and conducted some statistical analyses. He contributed to writing the report.
Charlotte Seneschall undertook the data extraction of the MedWatch forms.
Mark Simmonds was responsible for the management and checking of data provided by Medtronic and Glassman. He conducted the individual participant data meta-analyses and most of the other statistical analyses. He contributed to the writing of the report.
Lesley Stewart devised the project and developed the proposal with input from co-applicants. She was responsible for the overall direction and supervision of the project and contributed to the writing of the report.
Kath Wright developed and ran the database search strategies, set up current awareness searches, maintained the Endnote library of references and wrote up the literature searching process.
Corresponding author: Professor Lesley Stewart Centre for Reviews and Dissemination University of York Heslington York YO10 5DD UK [email protected]
iii Contents
Contents Acknowledgements ...... ii Authors' Contributions ...... iii Contents ...... iv List of Tables ...... vii List of Figures ...... ix Abbreviations and Glossary...... xi Preface ...... xii Executive Summary ...... xiv 1. Introduction ...... 1 2. General Methods ...... 3 2.1. Overview of our approach ...... 3 2.2. Establishing benefits and harms of rhBMP-2 in spinal fusion ...... 3 2.3. Overview of approach in establishing the reliability of the publicly available evidence...... 3 2.4. Objectives ...... 3 2.5. Protocol ...... 3 2.6. Inclusion criteria ...... 4 2.7. Literature searches ...... 4 2.8. Review processes ...... 5 3. Methods: Efficacy ...... 7 3.1. Data sources used ...... 7 3.2. Outcomes of interest ...... 7 3.3. Data management and checking of individual participant data ...... 9 3.4. Risk of bias ...... 9 3.5. Statistical analysis ...... 9 4. Methods: Safety ...... 13 4.1. Data sources...... 13 4.2. Outcomes of interest ...... 13 4.3. Individual participant data ...... 13 4.4. Cancer ...... 15 4.5. MedWatch forms ...... 15 4.6. Data from the wider literature ...... 15 5. Methods: Reliability of the Evidence ...... 18 5.1. Identifying the evidence ...... 18 5.2. Data collection ...... 18 5.3. Data synthesis ...... 18 6. Results: Included Studies ...... 19 6.1. Search Results ...... 19 6.2. Data supplied by Medtronic ...... 25 6.3. IPD from other sources ...... 25
iv Contents
6.4. Timeline of Medtronic trials and events relating to the use of rhBMP-2 ...... 25 7. Results: Efficacy ...... 27 7.1. Data checking and quality summary ...... 27 7.2. Pain and functionality scores ...... 32 7.3. Spinal fusion ...... 35 7.4. Impact of type of surgery on efficacy outcomes ...... 37 7.5. Impact of patient-level factors on efficacy outcomes...... 39 7.6. Sensitivity analyses ...... 40 7.7. Secondary outcomes ...... 42 7.8. Additional analyses: neurological and overall success ...... 46 7.9. The relationship between fusion and pain ...... 47 7.10. Summary of efficacy results: ...... 49 8. Results: Safety...... 50 8.1. Adverse events in the Medtronic trials ...... 50 8.2. Sensitivity and supplementary analyses of adverse events in the Medtronic RCTs ...... 57 8.3. Summary of IPD safety analyses ...... 60 8.4. Cancer ...... 60 8.5. Safety data from the wider literature ...... 63 8.6. Safety results from the MedWatch data ...... 89 9. Results: Reliability of the Evidence ...... 93 9.1. Comparison of reported outcomes ...... 93 9.2. Efficacy comparison ...... 101 9.3. Safety comparison ...... 104 9.4. Risk of bias comparison ...... 106 9.5. Summary of the reporting and reliability of Medtronic trials ...... 110 10. Discussion and Conclusions...... 111 References ...... 118 Appendix I ...... 135 Appendix II ...... 142_Toc356826544 Appendix III ...... 154 Appendix IV ...... 155 Appendix V ...... 156 Appendix VI ...... 159 Appendix VII ...... 160 Appendix VIII ...... 168 Appendix IX ...... 171 Appendix X ...... 189 Appendix XI ...... 192 Appendix XII ...... 193 Appendix XIII ...... 198
v Contents
Appendix XIV ...... 212 Appendix XV ...... 240 Appendix XVI ...... 242 Appendix XVII ...... 277 Appendix XVIII ...... 279 Appendix XIX ...... 280 Appendix XX ...... 283 Supplementary File ...... 284
vi List of Tables
List of Tables Table 1 Overview of all available sources for Medtronic trials...... 21 Table 2 Efficacy study characteristics ...... 28 Table 3 Number of included patients with recorded ODI, SF-36 PCS, and spinal fusion over time in 11 Medtronic RCTs...... 30 Table 4 Heterogeneity in meta-analyses of pain and function outcomes...... 34 Table 5 Heterogeneity in meta-analyses of spinal fusion (complete case analysis)...... 35 Table 6 Successful spinal fusion by time and treatment...... 37 Table 7 Adverse events in Medtronic trials...... 53 Table 8 Cases of cancer in 11 Medtronic trials using INFUSE®...... 61 Table 9 Cases of cancer in five Medtronic trials using AMPLIFY™ or rhBMP-2/BCP...... 62 Table 10 Lumbar PLF/PLIF/TLIF Study characteristics (data from the wider literature)...... 65 Table 11 ALF/ALIF Study characteristics (data from the wider literature)...... 68 Table 12 Cervical (ACDF/ PCF) Study characteristics (data from the wider literature) ...... 70 Table 13 Various surgical approach study characteristics (data from the wider literature) ...... 72 Table 14 Cause of event as described in PCR Investigation Summaries (spinal surgery only)...... 91 Table 15 Publication status of the 17 Medtronic trials...... 93 Table 16 Outcomes reported in peer-reviewed journal articles compared with Medtronic clinical study reports (CSRs)...... 94 Table 17 Proportion of collected outcomes presented in publicly accessible sources...... 95 Table 18 Proportion of collected outcomes reported in the peer-reviewed literature...... 96 Table 19 Reporting of adverse events in clinical study reports (CSRs) and journal articles for published Medtronic trials...... 98 Table 20 Comparison of number of adverse events reported in Medtronic IPD versus Medtronic publications...... 105 Table 21 Distribution of demographic data (age, previous surgery, smoking, alcohol, BMI), looking for outliers...... 193 Table 22 Pain scores (Oswestry. SF-36 components, back and leg pain), checking for outliers and coding errors...... 193 Table 23 Treatment coding, checking for consistency across treatment coding in different data files 194 Table 24 Previous surgery, checking for balanced randomisation across treatment groups ...... 195 Table 25 Surgery as a result of litigation, checking for balanced randomisation ...... 196 Table 26 Age, checking for balanced randomisation across treatment groups ...... 197 Table 27 Oswestry score, checking for consitency bertween reported overall score and score as calcualted from raw data ...... 197 Table 28 Primary and secondary outcomes as specified in Medtronic protocols...... 199 Table 29 Summary of interactions between patient-level factors and ODI, SF-36 PCS, back pain, and leg pain...... 205 Table 30 Summary of interactions between patient-level factors and successful fusion...... 206 Table 31 All adverse events categories in the Medtronic studies ...... 212 Table 32 Adverse events summary table from INFUSE®/LT-CAGE® pilot RCT...... 213 Table 33 Adverse events summary table from INFUSE®/LT-CAGE® open pivotal RCT...... 215 Table 34 Adverse events summary table from INFUSE®/LT-CAGE® laparoscopic pivotal single-arm study...... 216 Table 35 Adverse events summary table from INFUSE®/bone dowel pilot RCT...... 218 Table 36 Adverse events summary table from INFUSE®/bone dowel pivotal RCT...... 219 Table 37 Adverse events summary table from INFUSE®/INTER FIX™ PLIF RCT...... 220 Table 38 Adverse events summary table from INFUSE®/MASTERGRAFT® pilot RCT...... 221 Table 39 Adverse events summary table from INFUSE®/CORNERSTONE® ACDF pilot RCT...... 222 Table 40 Adverse events summary table from INFUSE®/CORNERSTONE® ACDF pivotal RCT. ... 223 Table 41 Adverse events summary table from INFUSE®/INTER FIX™ ALIF pilot RCT...... 224 Table 42 Adverse events summary table from MAVERICK™ Disc Pivotal RCT ...... 225 Table 43 Adverse events summary table from INFUSE®/TELAMON PEEK Instrumented PLIF pilot, single-arm study...... 229 Table 44 Adverse events summary table from AMPLIFY™ (rhBMP-2/CRM) pivotal RCT/ ...... 231 Table 45 Adverse events summary table from rhBMP-2/CRM 2-level pilot, single-arm study...... 235 Table 46 Adverse events summary table from rhBMP-2/BCP Mexico pilot...... 236 Table 47 Adverse events summary table from rhBMP-2/BCP US pilot RCT...... 237 Table 48 Adverse events summary table from rhBMP-2/BCP Canada pivotal RCT...... 238
vii List of Tables
Table 49 Adverse events in the Glassman trial ...... 242 Table 50 Reasons for further surgeries in Medtronic trials...... 244 Table 51 Data extracted from ALF/ALIF wider literature papers...... 246 Table 52 Data extracted from PLF/TLIF aggregate additional papers...... 248 Table 53 Data extracted from cervical wider literature papers...... 251 Table 54 Data extracted from various surgical type papers...... 254 Table 55 Quality assessment for PLF/TLIF papers from the wider literature...... 256 Table 56 Quality assessment for ALF/ALIF papers from the wider literature...... 263 Table 57 Quality assessment for cervical fusion papers from the wider literature...... 266 Table 58 Quality assessment for mixed approach fusion papers from the wider literature...... 272 Table 59 Number (%) of Medtronic/PCR adverse events reports ...... 277 Table 60 Adverse events occurring in Medtronic trials, as reported in published journal articles...... 280 Table 61 Pain and function outcome reporting in publications...... 283
viii List of Figures
List of Figures Figure 1 Flowchart displaying the flow of materials through the review process...... 20 Figure 2 Timeline of Medtronic trials and other events relating to the use of rhBMP-2...... 26 Figure 3 Risk of bias judgments for RCTs included in efficacy analyses...... 31 Figure 4 Meta-analyses of pain and function outcomes at 6 weeks, 3, 6, 12 and 24 months after surgery...... 32 Figure 5 Forest plot of mean difference in ODI 24 months after surgery...... 33 Figure 6 Forest plot of mean difference in SF-36 PCS score 24 months after surgery...... 34 Figure 7 Forest plot of relative risk of successful fusion 24 months after surgery...... 36 Figure 8 Meta-analyses of successful spinal fusion at 6, 12, and 24 months after surgery...... 37 Figure 9 Meta-analysis of ODI at 24 months according to type of surgery received...... 38 Figure 10 Meta-analysis of fusion at 24 months according to type of surgery received...... 39 Figure 11 Meta-analyses of pain and function outcomes for Medtronic trials only...... 40 Figure 12 Reduction in pain scores from pre-operative results by treatment received...... 41 Figure 13 Forest plot of mean difference in duration of hospital stay...... 42 Figure 14 Forest plot of mean difference in operating time...... 43 Figure 15 Meta-analysis of successful return to work...... 44 Figure 16 Meta-analysis for use of pain-relief medication...... 45 Figure 17 Meta-analyses of neurological and overall success...... 46 Figure 18 Relationship between relative risk of fusion and mean difference in effectiveness outcomes across trials 24 months after surgery...... 47 Figure 19 Mean improvement in pain and function according to treatment and fusion status at six, 12, and 24 months...... 48 Figure 20 Risk of bias judgments of trials included in safety analyses...... 51 Figure 21 Meta-analysis of adverse events by category at time of surgery in 11 Medtronic RCTs...... 54 Figure 22 Meta-analysis of adverse events by category at any time during the first 24 months after surgery in 11 Medtronic RCTs...... 55 Figure 23 Meta-analyses of adverse events according to adverse event category...... 56 Figure 24 Meta-analyses of adverse events according to severity in 11 Medtronic RCTs...... 58 Figure 25 Meta-analyses of adverse events according to relatedness to device use or surgery in 11 Medtronic RCTs...... 59 Figure 26 Forest plot of second surgeries...... 60 Figure 27 Forest plot of cancer incidence in the Medtronic RCTs...... 63 Figure 28 Relative risk of heterotopic bone formation in five studies...... 75 Figure 29 Relative risk of osteolysis in two studies...... 76 Figure 30 Relative risk of infection in four studies...... 77 Figure 31 Relative risk of neurologic events in three studies...... 78 Figure 32 Relative risk of hardware failure in six studies...... 79 Figure 33 Relative risk of wound complications in 10 studies...... 80 Figure 34 Relative risk of retrograde ejaculation in one study...... 81 Figure 35 Relative risk of urinary retention in one study...... 82 Figure 36 Relative risk of dysphagia in six studies...... 83 Figure 37 Relative risk of neck pain in one study...... 84 Figure 38 Relative risk of recurrent laryngeal neck palsy in one study...... 84 Figure 39 Relative risk of leg pain or radiculitis in four studies...... 85 Figure 40 Relative risk of inflammatory cyst formation in two studies...... 86 Figure 41 Distribution of pre-specified adverse events reported in MedWatch forms in spinal use (08/2003 to 04/2011)...... 90 Figure 42 Meta-analyses of pain and function outcomes comparing complete IPD and clinical study report (CSR) data with publications ...... 102 Figure 43 Meta-analyses of successful fusion comparing complete IPD and clinical study report (CSR) data with publications ...... 103 Figure 44 Risk of bias judgment of fusion outcomes in trials included in reliability analyses comparing full vs. public access...... 107 Figure 45 Risk of bias judgment of patient reported outcomes in trials included in reliability analyses comparing full vs. public access...... 108 Figure 46 Risk of bias judgment of adverse events in trials included in reliability analyses comparing full vs. public access...... 109 Figure 47 Forest plot of back pain 24 months after surgery...... 203
ix List of Figures
Figure 48 Forest plot of leg pain 24 months after surgery...... 203 Figure 49 Meta-analyses of the eight sub-sections of the SF-36 questionnaire...... 204 Figure 50 Sensitivity analysis: Oswestry score at 24 months including additional patients from BCP US trial...... 207 Figure 51 Sensitivity analysis: Fusion success at 24 months including additional patients from BCP US trial...... 207 Figure 52 Sensitivity analysis: ODI score at 24 months excluding cervical surgery trial (Cornerstone pilot)...... 208 Figure 53 Sensitivity analysis: Fusion at 24 months excluding cervical surgery trial (Cornerstone pilot)...... 208 Figure 54 ODI at 24 months according to pilot or pivotal study...... 209 Figure 55 ODI at 24 months INFUSE® vs AMPLIFY™/BCP...... 209 Figure 56 Spinal fusion at 24 months INFUSE® vs AMPLIFY™/BCP...... 210 Figure 57 One-stage meta-analysis results for pain and functionality outcomes...... 210 Figure 58 Analysis of efficacy outcomes comparing complete case and multiple imputation analyses...... 211 Figure 59 Sensitivity analysis of spinal fusion comparing complete case and multiple imputation analyses...... 211 Figure 60 Adverse events for INFUSE trials...... 243 Figure 61 Severity of adverse events in INFUSE trials...... 244 Figure 62 Forest plot for relative risk of cancer including additional cases from INFUSE®/LT-CAGE® open pivotal RCT...... 245
x Abbreviations and Glossary
Abbreviations and Glossary ACDF Anterior cervical discectomy and fusion ACS Absorbable collagen sponge ALF Anterior lumbar fusion ALIF Anterior lumbar interbody fusion AMPLIFY™ 2.0mg/cm3 of rhBMP-2 + hard collagen matrix imbedded with calcium phosphate ATLANTIS Anterior cervical plate system BCP Biphasic calcium phosphate CD HORIZON® Cotrel-Dubousset posterior spinal fixation system CORNERSTONE-SR® Machined allograft bone rings for cervical interbody fusion surgery CRD Centre for Reviews and Dissemination, York, UK CRM Compression resistant matrix (absorbable collagen sponge imbedded with biphasic calcium phosphate) CSR Clinical study report FDA US Food and Drug Administration ICBG Iliac crest bone graft INFUSE® 1.5mg/cm3 of rhBMP-2 + absorbable collagen sponge IPD Individual participant data LT-CAGE® Lumbar Tapered Fusion device used in interbody fusion surgery MASTERGRAFT® Ceramic granules used as a space filler, placed on top of an absorbable collagen sponge MAVERICK™ Total disc replacement system MCS Mental component score of the SF-36 Medtronic Medtronic, Inc. – manufacturer of INFUSE® and AMPLIFY™ rhBMP-2 compounds MedWatch The US Food and Drug Administration safety information and adverse event reporting programme NDI Neck Disability Index NIHR National Institute for Health Research (UK) NOVUS LC Interbody fusion cage ODI Oswestry Disability Index PCS Physical component score of the SF-36 PLF Posterolateral fusion PLIF Posterior Lumbar Interbody Fusion RCT Randomised controlled trial rhBMP-2 Recombinant human bone-morphogenetic protein type 2 SF-36 Short Form (36) Health Survey TELAMON P Impacted Implant Tapered lumbar interbody fusion implant (made from polyether-ether-ketone) containing BMP or bone graft TLIF Transforaminal lumbar interbody fusion YODA Yale University Open Data Access
xi Preface
Preface The Centre for Reviews and Dissemination (CRD) was commissioned by the Yale University Open Data Access (YODA) initiative to undertake this research. It forms part of an overarching YODA project to review the safety and effectiveness of the bone morphogenetic protein rhBMP-2 as used in spinal fusion surgery, including re-analysis of individual participant data (IPD) made available by the manufacturer (Medtronic). This is a high profile and controversial topic so we preface our report with brief background information outlining the context for our research and provide a short rationale for our approach. The YODA initiative was established to facilitate independent assessment and dissemination of data relevant to the benefits and harms of industry products. The team at Yale set up an infrastructure within which data from clinical trials, in the form of de-identified IPD, can be deposited and subject to independent scrutiny and re-analysis. For this project two independent evaluation teams, CRD and the Oregon Evidence-based Practice Center at Oregon Health & Science University (OHSU), were commissioned to undertake separate evaluations of rhBMP-2. YODA's aim is to promote wider access to clinical trial programme data, increase transparency, protect against industry influence and accelerate the generation of new knowledge (http://medicine.yale.edu/core/projects/yodap/). This is the first project to be undertaken as part of the YODA initiative. CRD, which is part of the National Institute for Health Research (NIHR) and an academic department of the University of York in the UK, submitted a research proposal to undertake this project in response to a competitive call. We did so for a number of reasons. The project provided an opportunity to address an important high profile clinical question and to explore methodological areas of interest. The chance to contribute to the ongoing debate around data disclosure and be part of the overarching YODA initiative to pilot a process for opening up access to data from industry-sponsored clinical trials was an important factor. We believe the project as a whole to be proof of concept that could help set new standards in disclosure and transparency. From the outset our research has been conducted independently and with complete academic freedom. CRD had no direct contact with Medtronic and all questions and queries about data were mediated by the YODA team at Yale. Nor have we had direct contact with the Oregon evaluation team. CRD had regular telephone progress meetings with the YODA team throughout the project but these were focussed on administrative issues and progress updates. Results were not shared with the YODA team or with anyone outside of CRD prior to submission of this report. Beyond approval of the CRD proposal and agreement of the core project objectives and deliverables, the YODA team have had no input to the conduct of the project and have not influenced the design, conduct, analyses, interpretation or reporting of findings in any way. In order to avoid financial conflict of interest, CRD has a policy not to undertake work for the pharmaceutical or medical devices industry. Therefore the project was agreed only when we were clear that our contract was with Yale and that there was no direct link with Medtronic. The research grant awarded was conditional only on completion of the work agreed in the core objectives and delivery of a final report. There are no financial consequences for CRD of the findings of our research. There were two inter-related research questions to be addressed. Firstly, is rhBMP-2 a safe and effective intervention in spinal fusion surgery? Secondly, has academic reporting of industry- sponsored trials lacked rigour, particularly in relation to adverse effects, and if so has this undermined the integrity of the publicly available evidence on which clinical decisions are based? To address these questions, we undertook a research project that embeds a systematic review and IPD meta-analysis of the benefits and harms of rhBMP-2 in spinal fusion within a wider exploration of data reporting and publishing practice. The intention of taking this wider view was to establish whether the published literature presents a true reflection of the underlying research data and whether it and any syntheses that might be derived from it provide reliable evidence for decision making. Our intention to address both questions was approved by Yale but, recognising that time and resources were limited, only the former was included in the core objectives that were common to both evaluation teams. Therefore, we describe these themes in separate sections of this report. It is important to note that we were provided with a large amount of data and documentation in this time and resource limited project and that our research proposal was developed before we were aware of the exact nature of the data and information to be provided by Medtronic. Our intention was
xii Preface to consider the Medtronic sponsored trials in the context of a full systematic review of all relevant clinical studies. Although some information recorded in the supplied documents (such as the summary patient narratives from the adverse event sections of the clinical study reports) provide more detailed qualitative information than was available in the IPD, it was neither our intention nor was it possible to undertake a detailed examination of every piece of information supplied by Medtronic. As part of the overarching process this report was submitted to YODA in August 2012 for peer review prior to release. We amended the report in light of peer review comments and logged substantive changes in a supplementary file. We also received comments from Medtronic and in consequence we corrected inaccuracies. These changes are logged in a supplementary file at the end of this report. In October 2012, results of the analyses of effectiveness and safety were submitted to and accepted for publication by Annals of Internal Medicine. Changes to this report made in keeping with amendments to the manuscript following peer review by the journal are noted in the supplementary file. Results of our evaluation of the reliability of the published evidence will be submitted for publication in due course.
xiii Executive Summary
Executive Summary
Background In 2002 the US Food and Drug Administration (FDA) approved the use of rhBMP-2 (INFUSE®) in anterior lumbar interbody fusion (ALIF) spinal fusion surgery. Since then, use of this product for spinal surgery – including in "off label" procedures – has increased rapidly. Preliminary industry-sponsored trials of rhBMP-2 in ALIF surgery published in academic journals from 2000 reported clinical benefits and stated that there were no (or no unanticipated) adverse events relating to rhBMP-2 treatment. Subsequent publications of further industry-sponsored trials that used rhBMP-2 in a variety of spinal fusion techniques similarly described positive clinical outcomes and reported no adverse events. From 2006 a number of published studies, predominantly case series or case-control studies, reported potential adverse events associated with the use of rhBMP-2. In 2008 the FDA issued a Public Health Notification of potentially life-threatening complications associated with rhBMP-2 use in the cervical spine. A subsequent review of publicly available data suggested that study designs were biased and suggested an increased risk of complications and adverse events for patients receiving rhBMP-2 that was 10 to 50 times higher than the original estimates. Amidst this controversy, the Yale University Open Data Access (YODA) team reached an agreement with Medtronic, the manufacturer of rhBMP-2 compounds, for the latter to provide full individual participant data from all their trials of rhBMP-2 and allow unrestricted independent re-analysis of these data. Here we report the finding of an independent analysis commissioned by YODA and conducted by the Centre for Reviews and Dissemination at the University of York.
Methods The review focused exclusively on the use of rhBMP-2 in patients undergoing spinal fusion surgery for treatment of degenerative disc disease, spondylolisthesis or any other relevant spinal condition. In addition to receiving individual participant data (IPD) and related documentation provided to YODA by Medtronic we conducted comprehensive literature searches to identify any other eligible randomised controlled trials (RCTs) and any other clinical studies reporting adverse events. The three objectives of the review were to: 1. Examine the potential benefits of rhBMP-2 We conducted a systematic review and meta-analysis of IPD of all RCTs (whether Medtronic funded or not) comparing rhBMP-2 with iliac crest bone graft (ICBG) in spinal fusion surgery irrespective of spinal level, surgical approach and dose of rhBMP-2. We checked IPD and used the Cochrane Collaboration’s risk of bias assessment tool to investigate risk of bias in the studies. Our primary outcomes were those most likely to be important to patients: Oswestry Disability Index (ODI); Neck Disability Index (NDI); SF-36 Physical Component Score (SF-36 PCS); back and leg pain; and successful spinal fusion (as defined radiographically by Medtronic). Data were analysed at six weeks and three, six, 12 and 24 months after surgery. We also analysed other outcomes such as neurological status, operation time, length of hospital stay, whether patients returned to work or usual activity and post-operative use of pain medication. We used standard two-stage meta-analytic techniques: IPD from each trial were analysed separately using the same methods across trials; and the resulting summary statistics were combined. Separate meta-analyses were performed for each of the specified time points. Where events were sparse we used one-stage models. For continuously distributed outcomes (such as ODI) we used mean differences between treatment arms in the change in score from pre-operative values. For dichotomous outcomes (such as successful fusion) we used relative risks and odds ratios, as appropriate. Heterogeneity was assessed in all meta-analyses using Higgins’ I2 statistic and Cochran’s Q test. We used subgroup analysis to examine whether the effect of rhBMP-2 on outcomes varied according to the type of spinal surgery. We used a one-stage random-effects regression model to investigate whether patient-level factors (age, sex, smoking, alcohol consumption, body mass index, diabetic status and history of spinal surgery for back pain) were associated with efficacy of rhBMP-2 surgery.
xiv Executive Summary
We further investigated the association between treatment received, successful fusion and long-term pain using a one-stage regression model. 2. Examine the potential harms of rhBMP-2 We investigated whether there was any difference in incidence of adverse events between patients receiving rhBMP-2 and ICBG by performing meta-analyses of adverse event data from the RCTs, using a one-stage meta-analysis model. We additionally considered IPD from Medtronic single-arm studies and summary adverse event data from other quality-assessed published comparative clinical studies. We also collated information from MedWatch/Product Comment Report (PCR) forms provided by Medtronic. 3. Assess the reliability of the published evidence base Evaluation of benefits and harms was informed by formal assessments of potential risks of bias that might influence study results. We also compared outcomes known to be collected for each study (as described in Medtronic’s original study protocols and clinical study reports) against outcomes reported in publicly available journal articles and abstracts so that we could establish the completeness of clinical outcome reporting from Medtronic sponsored studies. We performed meta-analyses of efficacy outcomes using data in the published literature and in clinical study reports (CSRs) and compared these with the results of the IPD analyses. These comparisons enabled us to assess potential for bias or under-reporting in the published literature. We compared study publications against their corresponding CSR on the number of events reported for each of the key categories to evaluate the completeness and consistency of adverse event reporting in the research literature.
Results Included studies Medtronic submitted data from 17 studies including 12 RCTs. Eleven of the RCTs used relevant comparators that made them eligible for the planned IPD meta-analysis. We obtained additional data from one of two further (non-Medtronic) RCTs identified by literature searches. In total, we analysed IPD from 1,409 participants enrolled in 12 RCTs that compared rhBMP-2 and ICBG. Searches also identified 43 publications reporting on 35 additional controlled studies that reported adverse events. We also examined 913 post-marketing MedWatch/PCR forms. Potential benefits of rhBMP-2 The use of rhBMP-2 in spinal surgery had modest benefits when compared with ICBG surgery 24 months after surgery. The use of rhBMP-2 reduced back pain measured on the ODI by 3.5 percentage points (95% confidence interval 0.5 to 6.5) and improved pain and function as measured by SF-36 PCS by 1.9 percentage points (95% CI 0.6 to 3.2) more than ICBG did. There was clear evidence that rhBMP-2 surgery improved fusion rates by 12% (95% CI 2% to 23%) 24 months after surgery; approximately 70% of patients who received ICBG had achieved fusion by 24 months. Typical improvements in ODI, SF-36 PCS and back pain at 24 months after surgery were not substantially different among ICBG patients irrespective of whether they had achieved fusion. This suggests that successful fusion does not in itself improve pain at 24 months post-surgery. Over the same period, rhBMP-2 patients with fusion reported greater (and statistically significant) benefits in ODI and back pain (but not SF-36 PCS) than patients without fusion. However, rhBMP-2 patients with fusion had slightly better outcomes than ICBG patients with fusion and rhBMP-2 patients without fusion had worse outcomes than ICBG patients without fusion. A possible explanation for this is that patients’ knowledge of treatment and fusion status led to bias in reporting pain and function outcomes. There was substantial heterogeneity between trials in many meta-analyses, particularly for spinal fusion (I2=76% 24 months after surgery). We found no evidence that treatment effects varied according to surgical approach and patient factors such as age or sex. Use of rhBMP-2 reduced operating times by about 20 minutes. We found no evidence of any benefit in terms of duration of hospital stay, whether patients returned to work and on post-operative use of pain medication. Overall, the trials show that rhBMP-2 surgery led to more patients achieving successful spinal fusion and a modest improvement in pain scores when compared with ICBG surgery. The improvement in
xv Executive Summary pain does not seem to be a result of the increase in fusion and the possibility that the moderate benefits of rhBMP-2 are primarily a consequence of bias due to patients’ knowledge of the surgical intervention cannot be excluded.
Potential harms of rhBMP-2 Analyses of adverse event IPD from the Medtronic sponsored trials showed some complications to be more common among rhBMP-2 patients. Back and leg pain were more commonly reported as an adverse event in the early post-operative period in the rhBMP-2 group (odds ratio 1.92, 95% CI 1.14 to 3.25). Arthritis, implant-related events, retrograde ejaculation, wound complications and neurological, urogenital and vascular events were also more common among rhBMP-2 patients but the lower incidence of these events meant that results were not statistically significant. Incidence of severe adverse events was similar across treatment groups. There was no overall difference in the need for second surgical procedures between rhBMP-2 and ICBG patients. Analyses of the Medtronic data indicated a near doubling in the number of cancers with rhBMP-2 (relative risk 1.8, 95% CI 0.8 to 4.2). There was no clear evidence of a dose response relationship and no indication that increased cancer risk was restricted to trials that used the higher dose preparation (AMPLIFY™). The overall absolute risk of cancer was low in both groups. Studies published in the wider literature and post-marketing data raised concerns about other adverse events not captured or easily apparent in the IPD provided. These included possible increased risk of heterotopic bone formation, osteolysis, retrograde ejaculation, urinary retention and dysphagia. Many of these observational studies had methodological limitations. Reliability of the published evidence base Sixty-five per cent (11/17) of the known Medtronic trials were published in scientific journals. Three of the six unpublished studies were RCTs. Around 70% of data collected at a study level on the key outcomes of fusion success, back pain and function were available in the peer-reviewed academic literature. Examination of the trial protocols and CSRs supplied by Medtronic suggested that even among the published studies a substantial proportion of outcomes collected in trials were not reported. Key efficacy outcome data relating to pain, function and fusion were consistently reported and we found no substantial differences between meta-analysis based on the published aggregate data and those based on the IPD. Therefore, reporting of efficacy in these trials was incomplete but does not appear to be substantially biased. In contrast, we found adverse events to be incompletely and inadequately described in the trial publications. Only two of 20 peer-reviewed journal publications (AMPLIFY™ pivotal RCT and MAVERICK™ Disc pivotal RCT) reported a comprehensive table of all adverse events that occurred during the study, similar to those provided in the CSRs. Publications from three trials did not report any data collected on adverse events. The remaining publications limited their reporting to a small proportion of the adverse events noted in the CSRs and the pattern of categories reported differed across trials. The way in which adverse event data were presented in the literature was inconsistent and the rationale for presenting some adverse events but not others was rarely clear. This suggests that the published academic literature alone is an inadequate resource for evaluating the safety of rhBMP-2. This problem is not unique to the Medtronic trials and several authors have noted poor reporting of adverse events from trials in the wider academic literature. This evidence suggests that, when compared with the IPD analysis, a systematic review of the published literature alone would produce similar findings in relation to efficacy outcomes but would likely have more uncertainty around these findings. Any conclusions about adverse effects from the published literature alone would be based on extremely limited and inconsistently reported data that were likely to result in uncertain and potentially misleading conclusions. To achieve even this would require a large amount of time and effort identifying, extracting and assessing all relevant publications.
Conclusion For the period up to 24 months after surgery, treatment with rhBMP-2 increases the probability of successful fusion (according to Medtronic definitions) but this does not appear to translate to clinically meaningful benefits in pain reduction, function or quality of life. The small benefits in these outcomes
xvi Executive Summary observed from six months onwards come at the expense of more pain in the immediate post-operative period and a possible increased risk of cancer, although the absolute risk of cancer is small. Chances of experiencing other complications may also be increased, although these findings are derived from less robust study designs. We suggest it is very important that these findings are expressed clearly and discussed with patients so that they can make informed choices about the type of surgery they would prefer.
xvii 1 Introduction
1. Introduction Bone morphogenetic proteins (BMPs) are a family of naturally occurring growth factors that are involved in a variety of physiological processes during development and following tissue injury,3 where they are implicated in inflammation and repair. Some bone morphogenetic proteins such as BMP-2 and BMP-7 have potent osteogenic properties encouraging mesenchymal stem cells to differentiate into bone forming osteoblasts4 and therefore are of potential clinical value in promoting bone fusion. Bony union usually occurs naturally during healing, for example following a long bone fracture. Fusion is not always achieved and non-union with pseudarthrosis formation can result in ongoing problems for patients. Thus, techniques or compounds that promote bony fusion are of potential value in a range of clinical scenarios including traumatic bone fractures, some types of spinal surgery and a number of maxillofacial and dental procedures. Patients with any of a wide range of spinal problems – including cervical and lumbar disc prolapse, neck pain, low back pain, degenerative and idiopathic spinal deformity and traumatic spinal fractures – may undergo surgery with an aim of attaining fusion at the operated spinal level. An established surgical technique for promoting local bone formation and fusion is to use fresh bone graft harvested from the patients themselves (autograft) either locally at the site of the fusion surgery or remotely from a region of healthy bone. A common site for bone graft harvesting is the iliac crest, although this has its own associated morbidity including donor site pain.5, 6 Consequently, a number of commercial products that act as bone graft substitutes have been developed. These include osteoconductive compounds such as bicalcium phosphate, demineralised bone matrix harvested from cadavers and most recently, recombinant bone morphogenetic proteins. One of these, recombinant human bone morphogenetic protein 2 (rhBMP-2) is produced and marketed by Medtronic under the trade name INFUSE®. INFUSE® contains 1.5mg/cm3 of rhBMP-2 in varying application sizes. At surgery, freeze-dried rhBMP-2 is reconstituted in water and bonded to a collagen sponge. This sponge is then applied to the site at which bony fusion is required. More recently, another formulation of rhBMP-2 known as AMPLIFY™ has been produced. This product contains 2.0mg/cm3 of rhBMP-2 in a 40mg application size supplied on a hard collagen matrix imbedded with calcium phosphate (20cm3 total). This is designed to resist deformation in vivo, therefore acting as an osteoconductive scaffold and providing an osteogenic local environment. In 2002, the US Food and Drug Administration (FDA) approved the use of rhBMP-2 INFUSE® in spinal fusion surgery. The license referred specifically to its use in anterior lumbar interbody fusion (ALIF) with a spacer device known as the LT-CAGE®. Any other use of rhBMP-2 in spinal surgery is off-label and outside of the licensed indication. Nevertheless, the numbers of spinal procedures using the product grew rapidly. In the United States, spinal fusion surgery involving the use of rhBMP-2 increased from less than 1% of all fusions in 2000 to almost 25% of all fusions in 2006.7 It has been reported that at least 85% of rhBMP-2 use is off-label.8 Preliminary industry-sponsored trials of rhBMP-2 in ALIF surgery published in academic journals reported benefits and stated that there were no (or no unanticipated) adverse events relating to rhBMP-2 treatment.9, 10 Subsequent publications of further industry-sponsored trials that used rhBMP- 2 in a variety of spinal fusion techniques similarly described positive clinical outcomes and reported no adverse events. However, from 2006 a number of published studies, predominantly case series or case-control studies, have reported potential adverse events associated with the use of rhBMP-2. In 2008, the FDA issued a Public Health Notification of potentially life-threatening complications associated with rhBMP-2 use in the cervical spine.11 Investigations by the US Justice Department regarding off-label use added to the controversy surrounding the use of this product. Prompted by this background and by claims of inadequate oversight of academic publication, the Editor-in-Chief of The Spine Journal (which published some of the original rhBMP-2 studies) led a comprehensive review of the publicly available data. This review suggested that the way studies were designed was inherently biased in favour of rhBMP-2 and that there was an increased risk of complications and adverse events for patients receiving rhBMP-2 that was 10 to 50 times higher than the original estimates.11 Amidst this controversy, the Yale University Open Data Access (YODA) team reached an agreement with Medtronic that they would provide full individual participant data from all their trials of rhBMP-2 and allow unrestricted independent re-analysis of these data. This provided an opportunity to carry out a systematic review and IPD meta-analysis which together are regarded as a gold standard
1 1 Introduction approach to synthesis and enables more in-depth analysis than is possible in a review of the published literature.
2 General Methods
2. General Methods
2.1. Overview of our approach We aimed to embed a systematic review and IPD meta-analysis of the benefits and harms of rhBMP- 2 in spinal fusion within a wider exploration of data reporting and publishing practice. Our core aim was to provide a robust and fair assessment of the benefits and harms of rhBMP-2 in spinal surgery so that health care professionals and patients can make informed decisions about its use. The intention of a wider exploration of publication practice was to establish whether the published literature presents a true reflection of the underlying research data and whether it, and any syntheses that might be derived from it, provide reliable evidence for decision making.
2.2. Establishing benefits and harms of rhBMP-2 in spinal fusion To limit the potential for bias we aimed to conduct a full systematic review and IPD meta-analysis of all relevant randomised controlled trials (RCTs) that compared rhBMP-2 with ICBG, including both published and unpublished studies. We did not restrict our review to Medtronic trials. We did not address the use of rhBMP-2 outside of the spine and we did not consider the use of other bone morphogenetic protein compounds. Our analysis of efficacy was restricted to RCTs. A major consideration from the outset was to focus on outcomes that are meaningful to patients. We additionally aimed to identify all published studies that reported information about adverse effects of rhBMP-2 in spinal surgery and in this respect did not restrict our consideration of adverse effects to randomised trials or to data from the Medtronic sponsored trials. We thus evaluated adverse events using IPD obtained from randomised trials and from Medtronic non-randomised studies. Aggregate data from other relevant independent non-randomised studies were considered and MedWatch/Product Comment Report forms (hereafter referred to as MedWatch forms) provided to us as part of the Medtronic data package were collated and summarised.
2.3. Overview of approach in establishing the reliability of the publicly available evidence Our second aim was to explore issues relating to the reporting and publication of the Medtronic trials. We were particularly interested in exploring the impact that any differences between trial publications and the underlying data might have on the results of systematic reviews and meta-analyses. To do this we compared the findings of our IPD meta-analyses with the findings of meta-analyses based on the corresponding data that were available in the published literature. Specifically, we extracted data from all publications of Medtronic trials and undertook analyses of these aggregate data in parallel to the IPD analyses. The intention of this was to establish whether there were important differences between the conclusions that could be drawn from the publicly available research evidence and from the underlying data.
2.4. Objectives 1. To examine the potential benefits of rhBMP-2 (focusing on meta-analysis of IPD from RCTs and on outcomes that are meaningful to patients). 2. To examine the potential harms of rhBMP-2 (focussing on meta-analysis of IPD from RCTs but also considering IPD from Medtronic single-arm studies, aggregate data from other published clinical studies, and information from MedWatch forms provided by Medtronic). 3. To assess the reliability of the published evidence base (by comparing the findings of our IPD meta-analyses with meta-analyses based on the data that were available in the published literature).
2.5. Protocol A protocol based on our original research proposal was written at the outset of the project. A copy of this was lodged with YODA and can be found in Appendix I.
3 General Methods
We registered the protocol in PROSPERO (an international prospective register of systematic review protocols) on 7 February 2012 (registration number CRD42012001907). We registered brief information to avoid sharing detailed research plans with a team who were undertaking a parallel, independent evaluation of the same evidence (Oregon Health & Science University).
2.6. Inclusion criteria The inclusion criteria for our review were formulated using the standard PICOS approach and are detailed below. Participants Studies that included patients who underwent spinal fusion surgery for treatment of degenerative disc disease, spondylolisthesis and any other relevant spinal conditions were included. Interventions Studies that evaluated rhBMP-2 in spinal fusion were included, including both Medtronic’s INFUSE® and AMPLIFY™ rhBMP-2 preparations. Other recombinant forms of BMP (such as rhBMP-7) were not included. Inclusion was not restricted by operative approach. Anterior lumbar interbody fusion (ALIF), posterolateral lumbar fusion (PLF), posterolateral lumbar interbody fusion (PLIF), transforaminal lumbar interbody fusion (TLIF) and anterior cervical discectomy and fusion (ACDF) approaches were all eligible. Inclusion was not restricted by type of surgery (open, minimally invasive or laparoscopic). Studies of rhBMP-2 used outside spinal fusion surgery (such as in long bone fractures) were excluded. All animal and in vitro studies were excluded. Comparators For the evaluation of efficacy, RCTs that compared rhBMP-2 with ICBG were eligible for inclusion. The single-arm studies provided by Medtronic were described but not included in our analyses. For the evaluation of adverse events we included RCTs as for efficacy comparisons. We also considered non-randomised studies in the wider literature that compared rhBMP-2 with any other spinal fusion techniques. Single-arm studies that reported adverse events were identified and are listed in Appendix IX but were not considered further. Outcomes We considered a range of clinical and radiological, patient-reported and investigator-derived outcomes. Further details are provided in the relevant sections that follow. Study designs For the evaluation of efficacy only RCTs were analysed. Single-arm studies provided by Medtronic were described but not included in the analyses. For the evaluation of safety, randomised and non- randomised comparative studies were included. Further details are provided in the relevant sections that follow.
2.7. Literature searches We searched the following databases for eligible literature: BIOSIS Previews, Cochrane Central Register of Controlled Trials (CENTRAL), Database of Abstracts of Reviews of Effects (DARE), EMBASE, MEDLINE, MEDLINE In-Process and Other Non-Indexed Citations, PubMed, Science Citation Index, TOXLINE and the FDA website. The search strategy was designed to retrieve any studies relevant to the efficacy or adverse events of rhBMP-2 in spinal fusion. We searched for synonyms in title and abstract and selected appropriate indexing/keywords. We did not use search filters for specific study designs due to the need to search beyond RCTs for adverse events data.12 Searches were not restricted by publication status or date. We checked references of relevant papers for further relevant studies. ClinicalTrials.gov was searched. A summary of the search strategies used are presented in Appendix II.
4 General Methods
Current awareness searches We used a range of current awareness services from 25 January 2012 to 30 June 2012 to identify emerging publications. Zetoc Alert (a current awareness service provided by the British Library that sends an email alert whenever new data are loaded into the database matching the prespecified search criteria) was set up to notify us whenever “spinal fusion”, “rhBMP” or “recombinant human bone morphogenetic” appeared in the title of an article or paper. We enabled OvidAutoAlerts in MEDLINE to notify us when new records that matched the original search criteria were added to the database. Call for evidence In addition to the database searches, we submitted a call for notification of potentially relevant evidence to The Spine Journal and The Back Letter newsletter on 1 February 2012 following completion of our initial literature searches and screening. The call was published by The Back Letter13 and placed on the website of the YODA project at Yale University in March 2012. (http://medicine.yale.edu/core/projects/yodap/research_groups.aspx). The text of this call is presented in Appendix III.
2.8. Review processes Two reviewers independently screened all titles and abstracts retrieved from electronic database and other searches. Full paper manuscripts of any potentially relevant publications were obtained and the relevance of each was independently assessed by two reviewers according to detailed criteria. A copy of our full-text screening criteria is presented in Appendix IV. Any discrepancies were resolved by consensus and where necessary a third reviewer was consulted. The data extraction and checking processes are described in more detail in the relevant sections that follow. Data collection Provision of individual participant data The YODA team secured the release of all clinical trial data (published and unpublished), post- marketing surveillance data and spontaneous adverse event data from Medtronic and made these data available for inclusion in this systematic review. We also sought IPD for any additional RCTs identified by the literature searches. Investigators of trials for which IPD were not supplied by Medtronic were contacted and asked to participate in the review by providing individual participant data for inclusion and re-analysis. If they agreed to participate, fully anonymised data on all randomised patients relating to the outcomes and trial and patient characteristics described above were requested. We checked these data to ensure that they were correctly coded, that missing data were correctly identified and to ensure that data were consistent with published results. Extraction of data from MedWatch forms As part of their original data submission, Medtronic provided us with 913 individual adverse event report forms (in PDF format) relating to the post-marketing use of rhBMP-2. These had been collected as part of the FDA safety information and adverse event reporting program (MedWatch). An additional 87 MedWatch forms were provided on 26 April 2012 and a further 116 in October 2012. Due to time and resource constraints we were not able to address the two additional submissions of data. We summarised the original 913 MedWatch forms alongside other safety-related data sources. Data storage and confidentiality All IPD and adverse event data from Medtronic were transferred to CRD via password-protected memory stick (thumb drive). Data from another investigator were in a de-identified format and received via email. All data were held in a password-protected area of the CRD server. Access was limited to staff working directly on the project. Copying of data to laptop computers or memory sticks was prohibited. Data from study publications We developed and piloted two separate data extraction forms for collecting information from the published reports of RCTs (Appendix V) and adverse events studies (Appendix VI). Data extracted
5 General Methods included details of study design, setting, and sponsor as well as trial and patient characteristics and outcome data. One researcher extracted data into a Microsoft Excel spreadsheet and a second reviewer checked the data extraction. We resolved any discrepancies via consensus or recourse to a third reviewer where necessary. Clinical study reports (CSRs) We did not systematically assess the CSRs but we used information from summary data tables of adverse events in our evaluation of reliability, for information on numbers of withdrawals after randomisation in our assessments of risk of bias and generally to provide clarification as required. We extracted dates of first and last surgery from these reports, where available.
6 3 Methods: Efficacy
3. Methods: Efficacy
3.1. Data sources used Analyses of efficacy were restricted to RCTs in spinal fusion surgery where rhBMP-2 was compared with conventional ICBG. Single-arm trials of rhBMP-2 and trials with comparators other than IBCG were excluded. One trial (rhBMP-2/BCP US pilot RCT) had two rhBMP-2 arms using different fixation procedures. Only the primary arm was used in these analyses. The second rhBMP-2 arm (of 11 patients) was combined with the first in sensitivity analyses. We performed all analyses using the patient-level data supplied by Medtronic and similar patient-level data from other identified relevant trials. Although intention-to-treat analyses were intended this was not possible as a number of randomised patients withdrew before surgery and no outcome data were available for them.
3.2. Outcomes of interest Investigation of efficacy focused on measures likely to be meaningful to patients, primarily on clinically relevant outcomes that assessed reduction in pain or improvements in general physical function before looking at radiological outcomes such as whether fusion was achieved. Primary outcomes Intended primary effectiveness outcomes were: • Disease-specific pain (patient-assessed by questionnaire, e.g. Oswestry Disability Index (ODI), Neck Disability Index (NDI), back pain) • General quality of life and functionality (patient-assessed by questionnaire e.g. SF-36) • Post-operative pain at surgical site and bone graft donor site • Successful spinal fusion The outcomes available in the Medtronic trials were: • ODI low back pain score (or NDI for cervical spine surgery): score ranges from zero to 100% (lower is better) • SF-36 Physical Component Score (SF-36 PCS): score ranges from zero to 100% (higher is better) • Other components of the SF-36 questionnaire • Back pain: score ranges from zero to 20 (lower is better) • Leg pain: score ranges from zero to 20 (lower is better) • Successful spinal fusion (according to Medtronic definition) The ODI and SF-36 questionnaires are widely used in assessing pain and have been externally validated.14, 15 The back pain and leg pain scores are those used by Medtronic. Validation data were not available. These outcomes were available at a range of different time points after surgery, including: • Six weeks • Three months • Six months • 12 months • 24 months Data on successful fusion were available from six months after surgery onwards. Data on post-operative pain were not provided as an efficacy outcome by Medtronic. Post-operative pain at the donor graft site was provided but only for ICBG patients. Pain was not assessed at the corresponding anatomical site for the rhBMP-2 patients so comparative analysis was not possible.
7 3 Methods: Efficacy
Secondary outcomes Secondary outcomes specified in the protocol and available in the Medtronic data were: • duration of hospital stay • operating time • successful return to work or usual activity We also analysed the supplied data on post-operative use of pain-relief medication as a surrogate for the post-operative pain outcome for which data were not available. Successful return to work and use of pain-relief medication were assessed at all time points listed above. Our protocol specified time to fusion as an outcome but no data were available for analysis. Additional outcomes For completeness, we also considered some outcomes that we had not specified in our protocol but were consistently reported in all the trial data supplied by Medtronic. These were: • Neurological success: defined as maintenance or improvement in neurological status (consisting of motor, sensory, reflex and leg raising measurements) • Overall success, requiring all of: Successful fusion Maintenance or improvement in neurological status ODI score improvement of 15 percentage points or more No associated adverse events No additional surgery These outcomes were analysed at all the time points listed above. Outcomes provided but not analysed Data were provided on bovine serum antibody response. We did not specify analysis of this in our protocol. We judged it to be of limited clinical interest and did not use these data. Patient-level and trial-level factors affecting efficacy We intended to investigate how the efficacy of rhBMP-2 might be influenced by trial- and patient-level characteristics. We intended to investigate the following trial-level factors (where available): • spinal location of surgery (e.g. cervical or lumbosacral) • surgical approach (e.g. anterior lumbar fusion, posterior lumbar interbody fusion) • nature of spinal condition (e.g. degenerative disc disease, spondylolisthesis) Investigation of the nature of spinal condition was not possible because there was no variation in spinal conditions considered across the randomised trials. We also investigated the following patient-level factors (where available): • previous spinal surgical interventions • age • sex • smoking status • diabetic status • body mass index
Diabetic status and body mass index were not prespecified in our protocol. We included these analyses as they were available and of clinical relevance.
8 3 Methods: Efficacy
3.3. Data management and checking of individual participant data Data were provided by Medtronic for each trial in a range of separate SAS-format data files for each trial according to the types of outcomes reported. From these, we collated individual-level data on all available efficacy outcomes as well as trial-level and participant-level for all outcomes at all the time points listed earlier. Data from the single non-Medtronic trial was provided as a single Excel spreadsheet. For each trial for which IPD were provided we checked the consistency of the data to ensure that the data as provided were valid and to check for errors in our data collation process. Data checking procedures included: • Uniqueness and consistency of patient identification numbers • Consistency of treatment allocation records • Ensuring all demographic data (e.g. ages) were within plausible ranges • Ensuring all pain and function measurements (e.g. ODI score, SF36 Physical Component Score) were within range with no outliers • Checking for balanced randomisation in terms of age, sex and other patient-level factors • Checking that summary scores (e.g. ODI score) agreed with the raw scores from each question from the questionnaire It was not possible to check judgements of fusion status without access to the raw radiological data and a radiologist. However, if fusion status was assessed by blinded experts any possible reporting bias or inconsistencies in judgment should be shared across both rhBMP-2 and the control arms. We have assumed that the assessments of spinal fusion provided by Medtronic are valid.
3.4. Risk of bias In addition to the data-checking procedures we assessed each trial using the Cochrane risk of bias tool16. This was developed by The Cochrane Collaboration to assess those aspects of trial design and conduct that have been demonstrated empirically to affect treatment effect estimates. The tool does not address aspects of trial design that relate to applicability or generalisability. The risk of bias tool covers six key areas of potential bias: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data and selective reporting of outcomes. Each area was given a judgement of high risk of bias, unclear risk of bias or low risk of bias and a reason for each judgement was recorded in the main data extraction spreadsheet. We used guidelines from The Cochrane Collaboration on what constituted high, low and unclear risk of bias with additional details based on discussions with the clinical member of the team (RM), see Appendix VII for details. Two reviewers independently completed risk of bias assessments for all trials included in the efficacy analyses for the first four domains. Judgements were made for each type of outcome reported in the trials: fusion, patient-reported and adverse events. These decisions were based on the full trial protocols provided by Medtronic and the brief protocol provided by the authors of the independent Glassman study. To address incomplete outcome data we used standard data-checking procedures described in the previous section to compare loss to follow-up in each arm. We checked that we had been provided with IPD for all outcomes that were listed in the trial protocols. We requested from Medtronic any available data for patients who were recruited but not reported in the trials. Some tabulated information was provided on why these patients did not receive surgery and so were excluded from the IPD but further data were unavailable. No post-randomisation data had been collected for these individuals. We checked loss to follow-up and missing outcome data (by treatment arm) for main outcomes at each analysis time point.
3.5. Statistical analysis All analyses of efficacy outcomes were performed using only RCTs that compared rhBMP-2 to ICBG in spinal surgery. Single-arm trials and trials that did not use ICBG were excluded. We performed all analyses using both IPD provided by Medtronic and IPD from other non-Medtronic trials for which data were available.
9 3 Methods: Efficacy
We performed separate analyses at each of the following time points after spinal surgery: • Six weeks • Three, six, 12 and 24 months We did not perform any analyses at time points beyond 24 months after surgery because most trials were designed to assess the effects of rhBMP-2 up to 24 months after surgery and follow-up data beyond that were available only for a subset of the patients. Our primary statistical method for estimating the efficacy of rhBMP-2 surgery for all the specified outcomes was to use standard two-stage meta-analytic techniques.17, 18 IPD from each trial were analysed separately for all the efficacy outcomes. We used the same methods across the trials. Separate meta-analyses were performed at each of the specified time points. We also used a one-stage meta-analysis approach in some analyses, primarily as a sensitivity analysis to confirm the results of two-stage analyses. These approaches are described in more detail below. All main analyses used a complete-case approach. Participants with missing data were excluded from the analysis. Estimates of effect Continuously distributed outcomes For the continuously distributed outcomes (ODI, SF-36, back pain, leg pain) we assessed efficacy in terms of the mean difference (MD) in outcome between the rhBMP-2 and ICBG arms. This was calculated by finding the change in score from baseline to the time point of interest for each patient, then calculating the mean (average) of these changes in each arm within each trial, along with its standard deviation. The difference in these mean changes between the two treatment arms is the mean difference at that time. This mean difference along with its associated standard error was calculated for each trial. Mean differences were combined in a meta-analysis and results were reported as mean differences between rhBMP-2 and ICBG arms, with their associated 95% confidence intervals. Dichotomous outcomes For dichotomous outcomes (successful fusion, successful return to work, use of pain-relief medication, neurological success, overall success) we assessed efficacy in terms of the relative risk (RR) for the outcome between the rhBMP-2 and ICBG arms. For each arm of each trial the risk of the outcome is the number of patients with the outcome divided by the total number of patients in that arm. The relative risk is the ratio of the risk in the rhBMP-2 arm to that in the ICBG arm. The logarithm of the relative risk, and its standard error, were calculated for each trial and these log-relative risks were combined in a meta-analysis. Results were transformed back into standard relative risks with 95% confidence intervals for presentation in the report. In the one-stage random-effects meta-analyses relative risks could not be calculated because algorithms did not converge successfully (they crashed). For these models, results were calculated in terms of the odds ratio (OR), with its corresponding 95% confidence interval. Two-stage meta-analyses We combined the effect estimates from each trial (mean difference or relative risk) across trials using a standard DerSimonian and Laird random-effects meta-analysis to account for potential heterogeneity in effects across trials.19 Separate analyses were performed at each time point. We present the results as summary plots across all times and as forest plots at individual time points. This is called a two-stage approach because it is performed in two stages: first we estimate effects within trials, and then combine results across trials in a meta-analysis. One-stage meta-analysis We performed one-stage meta-analyses of the pain and function outcomes as a comparison in order to confirm the validity of the two-stage analyses. In a one-stage analysis all patient data from all trials are combined simultaneously in a single regression model that is stratified by trial (hence one stage). For ODI, for example, we used a
10 3 Methods: Efficacy random-effects linear regression model of change in ODI from baseline against treatment received.
This model included data at all the time points simultaneously but with a separate treatment effect estimated at each time point. The model does not assume any particular model for changes in effects over time. However, it does assume the same amount of between-patient variation at every time point and the same amount of between-study variation in treatment effects (heterogeneity) at every time point. The model was also stratified according to the trial to which each patient belonged.20 This model had the form:
11 3 Methods: Efficacy imputation analysis for the main pain outcomes and for spinal fusion. This was achieved by
performing a regression of the outcome at each time point against the outcome at the previous time point, stratified by trial, using the complete-case data. Where a participant had no recorded outcome at some time point the predicted outcome from the model and its standard error were used to impute 10 possible outcomes, assuming that outcomes were normally distributed. The resulting 10 compete data sets were analysed and the results averaged to obtain a summary result after imputation. Exploratory analyses We performed some further analyses of efficacy not specified in our protocol. The main such analysis was an investigation of the association between pain outcomes, successful fusion and treatment to investigate whether successful or unsuccessful fusion affected a patient’s experience of pain. This analysis was conducted using a one-stage meta-analysis approach. For example, for ODI we performed a random-effects linear regression of change in ODI from baseline against both treatment received and whether or not fusion had been successful, with an interaction term between fusion and treatment. This model had the form:
12 4 Methods: Safety
4. Methods: Safety In this section we describe the sources of data used in the analysis of the safety of rhBMP-2 surgery, adverse events we considered and statistical methods used in the analysis.
4.1. Data sources • IPD from RCTs • Data from the wider literature • Data collated from the MedWatch forms supplied by Medtronic.
4.2. Outcomes of interest As prespecified in our protocol, we were interested in the following adverse events: General events • Heterotopic bone formation • Osteolysis • Infection • Neurological events (new/worse leg pain, sensory disturbance, reflex changes, bladder disturbance) • Cancer • Hardware failure (e.g. cage subsidence, implant breakages) Surgery-specific events • ALIF: major vascular injury, retrograde ejaculation, urinary retention • ACDF: dysphagia, airway obstruction, neck pain, recurrent laryngeal nerve palsy • PLIF/TLIF/PLF: leg pain/radiculitis, leg weakness, inflammatory cyst formation
4.3. Individual participant data In the analyses of IPD we summarised and tabulated data from all trials (excluding patients who received neither rhBMP-2 nor ICBG). Statistical analyses and meta-analyses were restricted to RCTs where rhBMP-2 was compared with ICBG. Single-arm trials of rhBMP-2 and trials with comparators other than ICBG were excluded from these comparative analyses. Classification of adverse events Adverse events in the Medtronic trials had not been recorded in a way that matched easily with our prespecified categorisations. Some of our prespecified adverse events were not recorded in the Medtronic trials (particularly, heterotopic bone formation, osteolysis and radiculitis). We analysed those adverse events supplied by Medtronic that we considered to be potentially related to spinal surgery. We considered the following categories of adverse events: • Pain: back, leg, lower extremity, arm, neck and upper extremity pain • Implant related (hardware failure): displacement, breakage, loosening, malpositioning and subsidence • Infection • Spinal events: not clearly specified by Medtronic but included new stenosis and adjacent segment events • Gastrointestinal events: including nausea and constipation • Neurological events: not clearly specified by Medtronic but including numbness, tingling and “pins and needles”
Additional information in Medtronic trials The Medtronic trials also categorised adverse events according to their perceived severity using four categories: mild, moderate, severe and life-threatening. Adverse events were also classified according to whether they were considered to be related to the medical device used (e.g. related to rhBMP-2 or the type of cage used), related to the surgery itself, unrelated to device or surgery or of
13 4 Methods: Safety unknown status. We analysed the adverse events in the Medtronic trials according to these severity
and relatedness classifications.
The Medtronic trials included data on whether patients required further spinal surgery. These data were extracted and combined in a meta-analysis across trials. Risk of bias We used the Cochrane risk of bias tool16 to assess the risk of bias in the assessment of adverse events. We followed the same processes as outlined in section 3.4 but based our judgments of risk of bias exclusively on adverse events outcomes and their assessment. Analysis We performed all analyses of incidence of adverse events using IPD provided by Medtronic. All analyses were performed using only RCTs that compared rhBMP-2 and ICBG. Single-arm trials and trials that did not use ICBG were excluded. We performed separate analyses at each of the following time points after spinal surgery: • At or immediately after surgery • Six weeks • Three, six, 12 and 24 months We did not perform any analyses at time points beyond 24 months after surgery because most trials were designed to assess the effects of rhBMP-2 up to 24 months after surgery and follow-up data were available only for a subset of the patients at time points beyond 24 months. One-stage meta-analyses The numbers of events in any particular adverse event category in any trial at any time point were often small. Some trials had no adverse events in a particular category. In such a situation (where events are rare) two-stage meta-analyses of relative risk may be inaccurate because of corrections required to adjust for trials with no events. Therefore, for all analyses of adverse events we used one- stage meta-analysis models to combine all data from all trials in one model as our primary method of analysis. For example, for an analysis of spinal events at the time of surgery the random-effects, trial stratified, one-stage logistic regression model used had the form:
14 4 Methods: Safety
4.4. Cancer Due to specific concerns that rhBMP-2 might increase the incidence of cancer we considered this as a distinct adverse event category and analysed it separately from other adverse events. Individual participant data We report all cases of cancer recorded in all Medtronic trials, including cancers in the single-arm trials, pre-existing cancers and cancers identified through extended follow-up. Meta-analyses were restricted to the RCTs conducted by Medtronic that compared rhBMP-2 with ICBG patients. Pre-existing cancers and cancers in the single-arm trials and the MAVERICK™ Disc pivotal RCT trial were excluded. Because the numbers of cancers were small we used a one-stage approach and carried out a two-stage meta-analysis as a confirmatory analysis. As a sensitivity analysis we analysed cancers in trials that used INFUSE® trials separately from those that used AMPLIFY™ or BCP preparations because of the differences in rhBMP-2 dose.
4.5. MedWatch forms Data were extracted from the FDA MedWatch forms provided by Medtronic and summarised as described below. Analysis Due to resource constraints, one clinically trained researcher extracted key information from the 913 MedWatch forms received in December 2011, including the source of complaint (e.g. healthcare professional, Medtronic employee, legal, literature), product name (e.g. INFUSE®), type of surgery (e.g. ALIF, PLIF, TLIF, PLF, ACDF, other), adverse event description (see General Methods) and the recorded cause of the event (“Unable to determine”, “Device may have caused or contributed”, “Not believed to be related to the product”, Other). Data were not extracted from the 83 forms that were received in April 2012 or from the 116 forms received in October 2012. We extracted these MedWatch-reported events in patients who received rhBMP-2 into a Microsoft Excel spreadsheet and used descriptive statistics to summarise data.
4.6. Data from the wider literature We aimed to identify all published studies in which any patients received rhBMP-2 during spinal fusion for degenerative disc disease, spondylolisthesis or any other relevant spinal condition. As specified in the protocol, we initially conducted a mapping exercise to describe these studies and to allow us to decide (based on the volume and level of evidence) which types of study we would include in our consideration of adverse events. The mapping exercise included published studies that ranged from small case series through to large observational cohort studies. We decided to extract detailed information only from comparative studies that reported on at least 10 adult patients (children were excluded because rhBMP-2 was used in different indications in the immature spine) to limit the investigation to sources of adverse event data that were potentially useful and reliable. Eligible comparator treatments were not restricted to ICBG alone. We did not seek IPD from these additional studies given the sparse data, the analyses of dichotomous data planned (event/no event) and because it was not feasible within the time and resources available. Classification of adverse events The publications we identified did not report the adverse events specified in our protocol in a standardised way. We developed a taxonomy based on synonymous and related terms in collaboration with our clinical co-author (RM) who has expertise in spinal fusion. This framework allowed us to capture events of interest as they were reported in the literature in a clear and transparent manner. The classification process was partly led by the data as we updated the categories during data extraction. For example, we added the category “wound complications” to the list of general adverse events, a category which included wound-related infections to distinguish between these and
15 4 Methods: Safety systemic infections. We also broadly classified adverse events reported in publications that did not closely fit our categories as “Other – potentially related to rhBMP-2” or “Other – probably not related to rhBMP-2”. We made all decisions on possible relatedness with the same clinical co-author. Appendix VIII shows the full list of categories and the individual adverse events they contain. We also recorded any reported adverse events that we could not classify according to this taxonomy. Quality assessment RCTs were assessed using the Cochrane risk of bias tool as described previously. Non-randomised controlled studies of interventions are generally considered to be at risk of the same biases as RCTs (selection, performance, detection, attrition and reporting biases). However, there is a greater risk of selection bias where comparison groups may differ in important characteristics. We attempted to focus on the more reliable studies by excluding very small studies and single-arm studies. We describe the non-randomised studies in terms of study design, purpose and sample characteristics within tables and text. The classification of study designs was based on the wording used by the authors of the article or abstract. We did not contact authors for clarification of these details. Quality assessment of non-randomised studies is an ongoing challenge as there is less evidence to support inclusion of particular dimensions when considering risk of bias for these designs. The non- randomised studies were assessed using a domain-based approach, using a modified version of the Newcastle-Ottawa scale25 which incorporates most of the areas suggested by the Cochrane Handbook.26 However, we avoided scoring or assigning star values to the studies as this runs the risk of producing an uninformative summary score.27 Instead, we tabulated the relevant information for each domain to allow for greater transparency. The Newcastle-Ottawa scale was designed for the assessment of observational studies of aetiology where the exposure is considerably less well defined than our interventional exposure of spinal fusion surgery. This made our judgements for some of the domains more straightforward. We considered the following domains: • Representativeness of the exposed cohort (did the patients require spinal fusion surgery? did they form a particular subgroup e.g. all smokers?) • Selection of the control group (were participants drawn from the same source e.g. same hospital or database?) • Ascertainment of exposure (given the nature of the topic this was usually via secure record in which the exposure to treatment was clearly known) • Outcome of interest not present at start of the study (was this explicitly checked for e.g. for pre- existing cancer?) • Confounders or other factors used to match or controlled for in analysis • Outcome assessment (independent/blind assessment, via secure medical record, self-report or no details) • Follow-up (we considered adequacy of duration in relation to the specific adverse events reported and whether this was similar across the two groups) We ensured that all data extraction was checked at least once, if possible by a second reviewer. We also carefully scrutinised all included studies to identify any possible overlap of patient populations across studies and discussed any uncertainty in the team. In cases where overlapping populations across publications seemed likely, we requested clarification from the authors to avoid double- counting of data. Analysis We broadly grouped studies by surgical approach: anterior lumbar, posterior lumbar and cervical spinal fusion. We added a fourth category (various) for publications that reported on more than one type of surgery or where type of surgery could not be determined. From the publications of each study we extracted data on the number of patients who received each type of surgery and the number of patients reported as having adverse events. Data were extracted
16 4 Methods: Safety for all prespecified adverse events, where these were reported in the studies. Where events were reported only as percentages of a known sample size, we imputed the number of patients experiencing the adverse event based on the reported data. For the primary adverse events of interest, within each study, we calculated the relative risk of having each specified adverse event for rhBMP-2 patients compared with the other type of surgery considered. Many studies had no adverse events in one surgery arm; we used the conventional adjustment of adding 0.5 to the numbers of events and patients in that study to estimate the relative risk and its 95% confidence interval.26 Relative risks with 95% confidence intervals were plotted in forest plots for each type of adverse event. Studies with no reported adverse events were excluded from these plots. Type of surgery, risk of event and comparator varied considerably across studies so we did not combine relative risks across studies in meta-analyses.
17 5 Methods: Reliability of the Evidence
5. Methods: Reliability of the Evidence As described in our protocol, as part of this project we intended to establish whether analyses based on data reported in scientific publications, data reported for regulatory purposes and the underlying trial participant data led to the same results and conclusions.
5.1. Identifying the evidence We obtained relevant IPD from Medtronic and identified published data from comprehensive searches of the literature. These processes are more fully described in section 2. Medtronic provided all related original protocols and CSRs in addition to IPD from their studies.
5.2. Data collection For each included study, we matched all known published papers and abstracts to the protocols and clinical trial reports provided by Medtronic. For each publication, we extracted key trial and patient characteristics as well as outcome data. Where trial data were reported in more than one publication, we used the most complete dataset available across all known publications. One researcher extracted data into a Microsoft Excel spreadsheet. A second reviewer checked the data extraction. We resolved any discrepancies via consensus or recourse to a third reviewer where necessary.
5.3. Data synthesis Comparison of reported outcomes To establish the completeness of clinical outcome reporting in publications of Medtronic studies, we compared the list of known collected outcomes for each study (as described in Medtronic’s original study protocols and CSRs) against the outcomes reported in publicly available publications and abstracts. To evaluate the completeness and consistency of adverse event reporting in the research literature, we compared study publications against their corresponding CSR on the number of events reported for each of the key categories. Efficacy comparison To establish the extent to which conclusions about efficacy might depend upon the source of study data analysed, we conducted separate meta-analyses of aggregate outcome data reported in CSRs and in publications and compared these against the results of our IPD meta-analyses. Meta-analyses were performed for ODI, SF-36 PCS, back pain, leg pain and fusion success. Meta- analyses based on data provided in CSRs were conducted using standard DerSimonian and Laird random-effects meta-analyses, combining the reported mean differences or relative risks across trials. A similar approach was used in meta-analyses using aggregate data extracted from trial publications. Not all publications provided data on standard deviations of pain scores required for an analysis. We calculated the average standard deviation in pain scores across trials where these were reported and used these averages as standard deviations in trials where these were not reported. Safety comparison To determine the extent to which conclusions about safety might depend upon the source of study data analysed, we compared the reporting of adverse events between publicly available and confidential data sources. Risk of bias comparison We assessed risk of bias using publicly available information from the trial publications and abridged protocols available on ClinicalTrials.gov and from the most complete source of information for each trial (confidential full trial protocols and CSRs provided by Medtronic). We used the Cochrane risk of bias tool as described previously.
18 6: Results: Included Studies
6. Results: Included Studies
6.1. Search Results We searched 10 key databases as described in section 2 as well as handsearching named sources and setting up current awareness alerts. We also searched ClinicalTrials.gov for completed and ongoing RCTs. Figure 1 shows that our searches of bibliographic databases returned 6,807 references. After de- duplication we screened 2,734 titles and abstracts and 311 of these we judged to be potentially eligible for inclusion. We obtained copies of these and on examination of the full papers we found 233 to be ineligible. These studies are listed with our reasons for exclusion in Appendix IX. We were unable to retrieve the full text for three studies that we considered potentially relevant based on their title and/or abstract. Details of these three publications are also listed in Appendix IX. When extracting data from the CSRs we encountered several conference abstracts of the trials that had not been retrieved by our searches. After consultation with our information specialist we agreed that this was not due to a flaw in the search strategy. Conference abstracts are often difficult to identify as they are not routinely indexed in bibliographic databases. We identified a total of 41 publications relating to 17 RCTs reported in the literature. Many of these were updated, partial or duplicate reports of the same trial or papers that reported findings combined across a number of trials. Three RCTs were not eligible for inclusion in our assessment of efficacy because they used comparators other than ICBG but they were included in our exploration of the wider adverse event literature. Ten of the 14 remaining RCTs identified from the searches were Medtronic trials. Two further Medtronic RCTs did not appear in the searches as they were not published but we had IPD available for analysis. Table 1 summarises all available data sources for the Medtronic trials. Two RCTs were conducted independently of Medtronic and were reported as a full paper and an abstract.1, 2 In addition to the RCTs that were eligible for inclusion in our analyses of efficacy, we identified a further 43 publications reporting 35 controlled studies that were included in our assessment of the wider adverse event literature. Figure 1 also shows the flow of data, published papers and other materials into the three main sections of the review (Efficacy, Safety and Reliability). It demonstrates where IPD were provided by Medtronic and where additional information (such as abridged protocols publically available at ClinicalTrials.gov) were downloaded by the evaluation team. We identified one further Medtronic-sponsored RCT from ClinicalTrials.gov.28 This RCT is ongoing so it was not eligible for inclusion in our review and is not shown in figure 1.
19 References identified in database search: n=6,807
Excluded during de-duplication: n=4,073 6: Results: Included Studies
Library of n=2,734 records
Additional records identified from bibliography hand-searches and electronic update searches: n=103
Library of n=2,837 records
Excluded during screening of titles and abstracts: n=2,521
Ordered for full-text screening: n=311
Materials provided by Medtronic: Full text could not be retrieved: n=3 • 17 protocols for 17 studies Excluded during full-text screening due to unmet inclusion criteria: n=233 • IPD for 17 studies • 17 CSRs reporting on 17 studies • Updated IPD cancer table
17 protocols for 17 Medtronic studies retrieved from clinicaltrials.gov
Efficacy Safety Reliability of the evidence • IPD from 11 Medtronic RCTs • IPD from 17 Medtronic studies • IPD from 17 Medtronic studies • IPD from 1 additional RCT • IPD from 1 additional RCT (Glassman • 32 publications reporting on 12 Medtronic studies 1 1 (Glassman 2008 ) 2008 ) • 17 Medtronic CSRs reporting on 17 studies 2 1 further RCT (Shapiro 2005 ) eligible • 43 publications reporting on 35 additional • 17 Medtronic protocols for 17 studies but IPD not provided controlled adverse events studies • 17 clinicaltrials.gov protocols for 17 Medtronic studies 1 further RCT (Shapiro 20052) eligible but IPD not provided
Figure 1 Flowchart displaying the flow of materials through the review process.
20 6: Results: Included Studies Table 1 Overview of all available sources for Medtronic trials.
Type of Trial Available Sources Comments source INFUSE®/LT-CAGE pilot Boden et al 2000 9 Full paper
RCT Original Medtronic Confidential protocol29 CSR: Volume 130, Volume Confidential 2 31 Clinicaltrials.gov Online protocol32 register INFUSE®/LT-CAGE open Dickman 200133 Abstract Reports extra 2 rhBMP-2 pivotal only patients unaccounted for
RCT elsewhere Gornet et al 200134 Abstract only Burkus et al 200210 Full paper Burkus et al 2003a35 Full paper Reports patients receiving open surgery from LT Cage Open trial and those receiving laparoscopic surgery from LT Cage Laparoscopic trial Burkus et al 2003b36 Full paper Single-centre report Burkus 200437 Full paper Reports Bone Dowel Pilot trial, pooled data from LT Cage Open and LT Cage Laparoscopic, and some patients from Interfix PLIF Chhabra et al 200638 Abstract Economic evaluation* only Chhabra et al 200739 Abstract Economic evaluation* only Van Genugten et al Abstract Economic evaluation* 200840 only Alt et al 200941 Full paper Economic evaluation* Burkus et al 200942 Full paper Reports patients receiving open surgery from LT Cage Open and those receiving laparoscopic surgery from LT Cage Laparoscopic Burkus et al 201143 Full paper Reports arms from LT Cage Open, LT Cage Laparoscopic , and Maverick Original Medtronic Confidential protocol44 CSR45 Confidential Clinicaltrials.gov Online protocol46 register INFUSE®/LT-CAGE Kleeman et al 200147 Full paper Reports a sub-group of patients laparoscopic pivotal 48 Zdeblick et al 2001 Abstract Single-arm trial only Burkus et al 2003a35 Full paper Reports patients receiving open surgery from LT Cage Open trial and those receiving laparoscopic surgery from LT Cage Laparoscopic trial Chhabra et al 200638 Abstract Economic evaluation* only
21 6: Results: Included Studies
Type of Trial Available Sources Comments source Chhabra et al 200739 Abstract Economic evaluation* only Van Genugten et al Abstract Economic evaluation* 200840 only Alt et al 200941 Full paper Economic evaluation* Burkus et al 200942 Full paper Reports patients receiving open surgery from LT Cage Open and those receiving laparoscopic surgery from LT Cage Laparoscopic Original Medtronic Confidential protocol49 CSR: Volume 250 Confidential Clinicaltrials.gov Online protocol51 register INFUSE®/Bone dowel pilot Burkus et al 200152 Abstract only RCT Burkus et al 200253 Full paper Burkus et al 200437 Full paper Reports Bone Dowel Pilot trial, pooled data from LT Cage Open and LT Cage Laparoscopic, and some patients from Interfix PLIFl Burkus et al 200554 Full paper Combined data from Bone Dowel Pilot and Bone Dowel Pivotal Burkus et al 200655 Full paper Combined data from Bone Dowel Pilot and Bone Dowel Pivotal Original Medtronic Confidential protocol56 CSR57 Confidential Joint report for Bone Dowel Pilot and Bone Dowel Pivotal Clinicaltrials.gov Online protocol58 register INFUSE®/Bone dowel Burkus et al 200459 Abstract pivotal only 54 RCT Burkus 2005 Full paper Combined data from Bone Dowel Pilot and Bone Dowel Pivotal Burkus et al 200655 Full paper Combined data from Bone Dowel Pilot and Bone Dowel Pivotal Original Medtronic Confidential protocol60 CSR57 Confidential Joint report for Bone Dowel Pilot and Bone Dowel Pivotal Clinicaltrials.gov Online protocol61 register INFUSE®/INTER FIX™ PLIF Alexander et al 200262 Abstract only RCT Burkus et al 200437 Full paper Reports Bone Dowel Pilot trial, pooled data from LT Cage Open and LT Cage Laparoscopic, and
22 6: Results: Included Studies
Type of Trial Available Sources Comments source some patients from Interfix PLIFl Haid et al 200463 Full paper Original Medtronic Confidential protocol64 CSR65 Confidential Clinicaltrials.gov Online protocol66 register INFUSE®/CORNERSTONE® Baskin et al 200367 Full paper ACDF pilot Original Medtronic Confidential 68 RCT protocol CSR69 Confidential Clinicaltrials.gov Online protocol70 register INFUSE®/MASTERGRAFT® Bae et al 200771 Full paper pilot 72 Dawson et al 2009 Abstract RCT only Original Medtronic Confidential protocol73 CSR74 Confidential Clinicaltrials.gov Online protocol75 register INFUSE®/INTER FIX™ ALIF Not published pilot Original Medtronic Confidential 76 RCT protocol CSR77 Confidential Clinicaltrials.gov Online protocol78 register MAVERICK™ Disc pivotal Gornet et al 200779 Abstract only RCT Gornet et al 2011 80 Full paper Original Medtronic Confidential protocol81 CSR82 Confidential Clinicaltrials.gov Online protocol83 register INFUSE®/TELAMON PEEK Not published instrumented PLIF pilot Original Medtronic Confidential 84 Single-arm trial protocol CSR85 Confidential Clinicaltrials.gov Online protocol86 register rhBMP-2/BCP US pilot Boden et al 200287 Full paper
RCT Original Medtronic Confidential protocol88 CSR89 Confidential Clinicaltrials.gov Online protocol90 register rhBMP-2/BCP Canada Assiri 200491 Abstract Single-centre report pivotal only 92 RCT Alexander et al 2007 Abstract Reports results from one arm only
23 6: Results: Included Studies
Type of Trial Available Sources Comments source Abraham et al 200893 Abstract Reports results from one arm only Abraham et al 201094 Abstract Reports results from one arm only Original Medtronic Confidential protocol95 CSR96 Confidential Clinicaltrials.gov Online protocol97 register AMPLIFY™ pivotal Glassman et al 200598 Full paper Single-centre report 99 RCT Dimar et al 2006 Full paper Dimar et al 2006100 Abstract only Glassman et al 2007101 Full paper Reports a sub-group of patients Dimar et al 2009102 Full paper McInnis 2010103 Abstract Budget impact model only Original Medtronic Confidential protocol104 CSR105 Confidential Clinicaltrials.gov Online protocol106 register rhBMP-2/CRM Two-level Not published pilot Original Medtronic Confidential 107 Single-arm trial protocol CSR108 Confidential Clinicaltrials.gov Online protocol109 register rhBMP-2/BCP Mexico pilot Not published
Single-arm trial Original Medtronic Confidential protocol110 CSR111 Confidential Clinicaltrials.gov Online protocol82 register INFUSE®/CORNERSTONE® Not published Stopped after recruiting three ACDF pivotal patients Original Medtronic Confidential 112 RCT protocol CSR113 Confidential Clinicaltrials.gov Online protocol114 register *Data from economic evaluations were not used in analyses. Ongoing trials Our searches of ClinicalTrials.gov identified a large (estimated number of patients to be recruited: 534) ongoing Medtronic trial of rhBMP-2 versus ICBG in TLIF surgery which, because it is not yet completed, could not be included in any analyses.28
24 6: Results: Included Studies
6.2. Data supplied by Medtronic Individual participant data Medtronic provided IPD in a series of SAS data files for each of their 17 completed trials: 13 were RCTs and four were single-arm studies. Further details of these data can be found in Appendix X. Trial documentation Medtronic also provided the following documentation and any revision documents for each one of their 17 trials. • Document index • Data dictionary • Trial protocol • Imaging protocol • Statistic considerations redacted • Descriptions of derived endpoint variables • Clinical report form • Adverse events form • Pre-market approval CSR redacted (where relevant) • Pre-market approval PAS final report redacted (where relevant) • Final report redacted • CSR Medtronic provided tabulated data on all cases of cancer observed in the 17 Medtronic trials and the MedWatch forms as described in section 4.5.
6.3. IPD from other sources We identified two RCTs that were conducted independently of Medtronic.1, 2 We contacted the authors of each of these trials and invited them to collaborate and provide IPD for inclusion in our meta- analyses. The author of one study published only as an abstract declined to participate.2 The authors of the other trial provided IPD (as a Microsoft Excel spreadsheet), further details of which can be found in Appendix XI. This additional trial is referred to as the “Glassman” trial throughout this report.
6.4. Timeline of Medtronic trials and events relating to the use of rhBMP-2 Since 1996 Medtronic has sponsored 17 completed trials of rhBMP-2. Figure 2 presents an overview of their chronological order alongside other key events relating to the use of rhBMP-2. In the figure, light grey shading indicates unpublished trials and published trials are displayed in dark grey. Start dates were determined by the date of the first surgery as reported in the CSR and completion dates were determined by the CSR submission date. P indicates the date of the protocol that was provided by Medtronic. These were the latest versions of the protocol for each trial. If there is no P the date of the protocol could not be determined from the documentation. A green X indicates a deviation in start date with the X indicating the start date reported in the ClinicalTrials.gov protocol. A red X indicates a deviation in completion date with the X indicating the completion date reported in the ClinicalTrials.gov protocol. Blue letters relate to the investigational device exemption (IDE) process of the FDA as detailed in additional documentation provided to us by Medtronic. For the INFUSE®/LT- CAGE® trials Medtronic requested an IDE from the FDA (indicated by the letter R), approval dates were not provided. A blue letter A indicates the time of IDE approval where data were available. For the MAVERICK™ trial, an IDE was initiated (represented by the letter I) by the FDA. No approval date was provided. No approval date was available for the rhBMP-2/BCP Mexico pilot. The ongoing Medtronic trial is displayed in blue. Black dots in the figure indicate publications of the respective trials. Details of these publications are provided in table 1 above. The numbers of MedWatch forms are included in this figure as an indicator of adverse events occurrence and are based on the documents supplied by Medtronic (see section 4.5). The MedWatch forms provided by Medtronic in 2012 differed in layout from the ones sent in 2011. We have attempted to extract the same date from all forms for inclusion in this figure even though the fields were labelled differently in the more recently submitted forms. Up to and including the forms that were provided in October 2012, 111 MedWatch forms were filed in 2012.
25 6: Results: Included Studies
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 201
1
Carragee FDA paper FDA Approval Studies begin Events to report warning for of INFUSE complications cervical use FDA denies
AMPLIFY YODA of Start project
BCP Mexico LT-CAGE® R •
pilot Bone dowel A • • • pilot LT-CAGE® R P • • • • • • • • open LT-CAGE® R P • • • • • laparoscopic IINTER A FIX™ ALIF pilot INTER FIX™ A • • • PLIF BCP US A • CORNERST A • ONE® pilot BCP Canada A P • • X • • Bone dowel A P • • • pivotal CORNERST A ONE® P X pivotal AMPLIFY™ A • P • • • X • A MASTERGR X • • AFT® pilot P MAVERICK I P • • X ™ Telamon A P CRM Two- P A level INFUSE® TLIF (estimated end: 2017) Number of MedWatch None None None None None None None 4 9 47 69 52 73 231 303 200 forms filed
Figure 2 Timeline of Medtronic trials and other events relating to the use of rhBMP-2.
26 7 Results: Efficacy
7. Results: Efficacy In this section we focus on the analysis of the efficacy of rhBMP-2 compared with ICBG surgery using individual participant data from 11 RCTs sponsored by Medtronic plus one additional randomised trial that was not sponsored by Medtronic1. We excluded single-arm studies. We excluded the MAVERICK™ Disc pivotal RCT because the comparator was not ICBG. Also excluded were 11 patients from the rhBMP-2/BCP US pilot RCT who received rhBMP-2 but with a different fixation procedure from other patients in the trial. Table 2 provides details of study characteristics for the included RCTs.
7.1. Data checking and quality summary As described in section 3.3 we checked the IPD provided by Medtronic for errors in coding of factors, errors in calculation of outcomes (such as ODI) and for potential imbalances in randomisation. Example SAS output from this checking for the INFUSE®/LT-CAGE® Pivotal RCT is given in Appendix XII. We found no evidence of any errors in the coding of patient factors or outcomes and no evidence of imbalance in the randomisation process for any demographic factor (such as age or sex). We were unable to check the validity of the randomisation process because dates of randomisation were not available. We were unable to check the internal consistency of classification of successful spinal fusion because we were unable to convert the component radiological data provided by Medtronic into a definition of successful fusion. We have assumed throughout this analysis that the Medtronic assignment of successful fusion according to their criteria is correct. Some patients in the Medtronic trials were randomised but did not receive surgery. There were 211 such patients across all 17 Medtronic trials: 96 were in the MAVERICK™ Disc pivotal RCT (not included in these efficacy analyses) and 55 were in the AMPLIFY™ pivotal RCT. Medtronic supplied information on the patients from the AMPLIFY™ pivotal RCT: 23 were randomised to rhBMP-2 and 32 to ICBG. The main reasons for exclusion were that a different surgery was thought to be required, refusal to participate or issues relating to insurance. Arguably these criteria should have been checked prior to randomisation but there was no evidence of bias between arms in the exclusions. Nine patients were excluded from rhBMP-2 arm of the INFUSE®/Bone dowel pivotal RCT; only one patient was excluded from the ICBG arm. Data were not available for all included patients at all the time points. We investigated the proportion of patients lost to follow-up for three main outcome measures to determine whether attrition bias may have been present. Table 3 summarises the numbers of patients with recorded ODI, SF36 PCS and spinal fusion over time for the 11 included Medtronic RCTs. Loss to follow-up is acceptable up to 24 months for ODI, although more data are missing (13.5%) for ICBG patients than for rhBMP-2 patients (9.5%). More data are missing for SF36 PCS and for spinal fusion outcomes. Again there appear to be slightly more data missing for ICBG patients. This pattern was similar across trials and for all outcomes and therefore does not suggest attrition bias. We did not have information on reasons why data were missing. Some Medtronic trials included follow-up data beyond 24 months; these were available for only a small proportion of individuals. Proportions of participants without data beyond 24 months after surgery are summarised in Table 3. Only a minority of participants were followed up beyond 24 months so no analyses were performed beyond that time.
27 7 Results: Efficacy
Table 2 Efficacy study characteristics
Trial name Previous Intervention Mean Patients Patients Patients Percentage spinal Surgical Approach age Follow-up Date of 1st patient randomised excluded evaluated male surgery Control enrolment (%) rhBMP-2/ACS + LT-CAGE INFUSE®/LT- Single-level 11 0 11 42 45 0 CAGE pilot (1.5 mg/ml) open/laparoscopic 2 years
January 1997 ALIF ICBG + LT-CAGE 3 0 3 40 67 0 rhBMP-2/ACS + LT-CAGE 2 years INFUSE®/LT- 145 2 143 43 55 38 CAGE open pivotal Single-level open (1.5mg/ml) control, 6 ALIF years ICBG + LT-CAGE 137 1 136 42 50 40 August 1998 investigational* rhBMP-2/ACS + cortical INFUSE®/Bone 24 0 24 41 33 46 dowel pilot Single-level open bone dowels (1.5mg/ml) 4 years ALIF ICBG + cortical bone 23 1 22 45 45 32 April 1998 dowels rhBMP-2/ACS + cortical INFUSE®/Bone 64 9 55 39 44 33 dowel pivotal Single-level open bone dowels (1.5mg/ml) 2 years ALIF ICBG + cortical bone 31 1 30 42 30 33 September 2000 dowels INFUSE®/INTER rhBMP-2/ACS + INTER 35 1 34 46 50 35 FIX™ PLIF Single-level PLIF FIX cages (1.5mg/ml) 2 years
March 1999 ICBG + INTER FIX cage 36 3 33 46 45 39 18 10/18 rhBMP-2/ACS + had Cornerstone implant + 18 0 51 44 6 single- INFUSE®/ Atlantis plate (1.5mg/ml) CORNERSTONE® One or two level level ACDF pilot anterior cervical surgery 2 years interbody fusion 15 September 1999 8/15 had ICBG/Cornerstone implant 15 0 single- 47 47 0 + Atlantis plate level surgery INFUSE®/ rhBMP- MASTERGRAFT® Single-level PLF 2/ACS/MASTERGRAFT/C 27 2 25 56 0 24 3 years** pilot D HORIZON (12mg)
28 7 Results: Efficacy
Trial name Previous Intervention Mean Patients Patients Patients Percentage spinal Surgical Approach age Follow-up Date of 1st patient randomised excluded evaluated male surgery Control enrolment (%)
April 2003 ICBG + CD HORIZON 23 2 21 57 0 29 rhBMP-2/ACS+NOVUS LC INFUSE®/INTER 25 0 25 45 44 44 FIX™ ALIF pilot*** Single-level open Device 2 years ALIF February 1999 ICBG+NOVUS LC 20 0 20 45 45 35 rhBMP-2/BCP (3mg/ml) 11 rhBMP-2/BCP US rhBMP-2/BCP + TSRH 22 0 50 55 18 pilot Single-level PLF 11 (2mg/ml) 2 years (not interbody) May 1999 ICBG + TSRH 5 0 5 52 40 0
rhBMP-2/BCP + CD rhBMP-2/BCP One or two level HORIZON or rhBMP- 102 4 98 53 36 19 Cohort 1 4 Canada pivotal posterolateral 2/BCP + TSRH years, cohort 2
lumbar fusion ICBG + CD HORIZON or 2 years**** September 1999 105 6 99 53 48 20 ICBG + TSRH
rhBMP-2/CRM/CD Horizon 262 23 239 53 45 31 AMPLIFY™ Open bilateral (2mg/ml) 5 years March 2002 PLF ICBG + CD Horizon 256 32 224 52 42 28
Glassman***** Single or rhBMP-2/ACS + fixation n/a n/a 52 69 30 32 multilevel PLF (no 2 years ICBG + fixation May 2004 interbody fusion) n/a n/a 54 70 33 37
* In their review comments on our draft report Medtronic clarified that in the INFUSE®/LT-CAGE® open pivotal RCT the investigational patients were followed up to 6 years while control patients were followed for 2 years. ** In their review comments on our draft report Medtronic clarified that in the INFUSE®/MASTERGRAFT® pilot RCT most of the patients were followed up to 2 years. Only seven patients were followed up to 3 years. *** The Interfix ALIF pilot CSR notes one rhBMP-2 patient who did not receive rhBMP-2 as having been treated as control, but this patient appears to have appropriately remain assigned to rhBMP-2 in the IPD. Two patients in the ICBG group are noted in the CSR as being not included beyond pre-operative and surgical analysis. **** In their review comments on our draft report Medtronic clarified that in the rhBMP-2/BCP Canada pivotal RCT cohort 1 was followed up for 4 years, only 3 patients had data for 6-year follow-up. Cohort 2 had 2 years of follow-up. *****The IPD provided for the Glassman trial did not contain information on any post-randomisation exclusions.
29 7 Results: Efficacy
Table 3 Number of included patients with recorded ODI, SF-36 PCS, and spinal fusion over time in 11 Medtronic RCTs.
Percentage of records missing Total Outcome Treatment randomised 6 3 6 1 2 3 4 5 weeks months months year years years years years
rhBMP-2 696 2.9 2.2 3.3 5.5 9.5 68.7 77.3 89.5 ODI ICBG 609 4.1 3.3 5.3 7.2 13.5 66.7 87.4 100
rhBMP-2 696 13.4 13.4 8.2 9.9 15.2 80.3 67.7 86.5 SF36 PCS ICBG 609 9.9 9.7 6.9 10 17.6 78.5 75.4 93.9
Spinal rhBMP-2 696 n/a n/a 14.5 13.5 17 81.5 71.8 87.6 fusion ICBG 609 n/a n/a 16.3 14.6 17.9 81.8 80.5 94.6
Similar data checks were performed on the Glassman trial data but these were more limited given the more limited data provided. We found no evidence of any errors in the coding of patient factors or outcomes and no evidence of imbalance in the randomisation process for any demographic factor. Summary of risk of bias We assessed Medtronic trials and the trial conducted by Glassman for potential risk of bias based on the supplied trial protocols and CSRs and IPD. The results are summarised in Figure 3. We found that randomisation and allocation concealment procedures were generally adequate, consistent across trials and unlikely to lead to bias. This is consistent with our IPD checks which revealed no evidence of imbalance in participant characteristics between treatment arms. Patients and physicians were not blind to the treatment received. This is stated clearly in trial protocols, reports and publications and is inevitable because ICBG requires an additional incision. Assessors of fusion were blinded to the treatment received (although it may have been possible to determine treatment from radiological images). It is unlikely that patient knowledge of treatment could affect spinal fusion. However, most efficacy outcomes (such as pain and function scores) were patient-reported and may potentially have been influenced by knowledge of the treatment received. This could lead to bias in these outcomes if, for example, a patient’s opinion of their pain was influenced by which treatment they received. IPD were provided for all outcomes listed in the full trial protocols and so there was no issue of selective reporting relating to our review and meta-analysis of the IPD provided (but see section 9.1 for a discussion of selective reporting relating to trial publications). A list of outcomes specified in each trial protocol is provided in Appendix XIII Table 28. In general trials excluded few randomised patients and where they did so this was usually balanced across treatment arms (except INFUSE® Bone Dowel Pivotal). The proportion of patients followed up in each arm was generally high across all trials and there was no apparent imbalance between arms so risk of attrition bias was considered to be low (see Table 3 above). Taken together, the risk of bias assessment of trial design and detailed checking of IPD suggest that Medtronic randomised trials and the Glassman trial were at generally low risk of selection, attrition and reporting biases. The risk of performance and detection biases was typically low for fusion outcomes, although these biases are more likely to influence non-blinded subjective outcomes such as pain and disability scores.
30 7 Results: Efficacy
Figure 3 Risk of bias judgments for RCTs included in efficacy analyses.
31 7 Results: Efficacy
7.2. Pain and functionality scores In this section we present results of meta-analyses of pain and functionality scores, namely: • Oswestry Disability Index (ODI; or Neck Disability Index, NDI, for cervical spinal surgery in the Cornerstone pilot trial): score ranges from zero to 100%, lower is better • SF-36 Physical Component Score (SF-36 PCS): score ranges from zero to 100%, higher is better • Back pain: score ranges from zero to 20, lower is better • Leg pain: score ranges from zero to 20, lower is better Analyses were performed using the 12 available RCTs (11 Medtronic and the Glassman trial1). Two- stage meta-analyses were performed at six weeks, three and six months and 12 and 24 months after surgery. Figure 4 summarises the results of the meta-analyses for all four outcomes (ODI, SF-36 PCS, back and leg pain) at each time point. The scales for ODI and SF-36 PCS are given as percentage points. For back and leg pain the results are shown in terms of the absolute 20-point pain scale. The points on this plot represent the mean difference (in change from baseline) between rhBMP-2 and ICBG at each time. The vertical lines show the 95% confidence interval of this mean difference. Points below the horizontal red line of no difference indicate greater efficacy of rhBMP-2, except for SF-36 PCS where points above the line indicate greater efficacy of rhBMP-2. Confidence intervals that do not cross the red line indicate a statistically significant result (p-value < 0.05).
Figure 4 Meta-analyses of pain and function outcomes at 6 weeks, 3, 6, 12 and 24 months after surgery. From three months to 24 months all estimates of effectiveness lie in the graph sector which illustrates advantage of rhBMP-2. For back pain, ODI and SF-36 PCS the corresponding 95% confidence intervals also lie within these sectors and do not cross the line of no difference (they are statistically significant at the p=0.05 level). Although results for back pain and ODI estimates at six weeks fall in the graph sector that favours ICBG, the confidence intervals span the line of no difference and are consistent with greater efficacy of either treatment. Similarly, there was no clear evidence of a difference in leg pain between treatment groups as the 95% confidence intervals cross the line of no difference. The observed benefits of rhBMP-2 were small. Results indicate that from three months to 24 months ODI was approximately two to four percentage points lower (better); back pain was better by about one point on the 20-point scale and SF36-PCS was about two percentage points higher than for people receiving rhBMP-2.
32 7 Results: Efficacy
We illustrate the effects of rhBMP-2 in each trial using forest plots for meta-analyses of pain and function outcomes. A forest plot for the mean difference in ODI at 24 months between rhBMP-2 and ICBG is shown in Figure 5 and for mean difference in SF-36 PCS at 24 months in Figure 6. In these forest plots each horizontal bar represents a trial. The square shows the effect estimate (e.g. mean difference in change from baseline for ODI) and the width of the horizontal line gives the 95% confidence interval. The larger the square, the greater the precision of the estimate (approximately equivalent to the size of the trial). The diamond and the dotted line show the summary effect estimate from the meta-analysis. The width of the diamond gives the 95% confidence interval for the summary effect estimate. The solid vertical line shows where there is no difference between rhBMP-2 and ICBG. Trials whose bars do not intersect this line have statistically significant results (p-value < 0.05).
Figure 5 Forest plot of mean difference in ODI 24 months after surgery. Figure 5 illustrates that for ODI at 24 months although most trials favour rhBMP-2, they have wide confidence intervals that do not rule out an advantage of ICBG. Individually, only two small pilot trials (Bone Dowel and Cornerstone) had statistically significant results. The combined results show that rhBMP-2 reduced lower back pain on the ODI by about 3.5 percentage points more than ICBG with the 95% confidence interval consistent with a 0.5% to 6.5% benefit.
33 7 Results: Efficacy
There is no square representing the INFUSE®/LT-CAGE® pilot RCT as this trial did not provide usable data for this analysis. Figure 6 Forest plot of mean difference in SF-36 PCS score 24 months after surgery. Figure 6 illustrates that for SF36 PCS at 24 months all trials showed greater efficacy of rhBMP-2 surgery but that all but one trial had 95% confidence intervals that crossed the line of no difference. The combined estimate indicates that rhBMP-2 increased functionality, according to SF-36 PCS, by about 1.9 percentage points more than ICBG, with a 95% confidence interval that is consistent with a 0.6% to 3.2% increase. Forest plots for back pain and leg pain at 24 months after surgery are given in Appendix XIII, Figure 47 and Figure 48. Back pain with rhBMP-2 was reduced relative to ICBG surgery by 1.58 points (95% CI 0.51 to 2.65); leg pain by 0.63 points (95% CI reduction of 1.37 points better to 0.12 worse). Heterogeneity is summarised in terms of the I2 statistic and the p-value from Cochran’s Q test in Table 4. Overall we found little evidence of heterogeneity in these analyses. However, some heterogeneity was present for back pain and for ODI at 24 months after surgery. These were the only two analyses where Cochran’s Q test had a p-value of less than 0.1, the commonly used cut-off for significant heterogeneity.
Table 4 Heterogeneity in meta-analyses of pain and function outcomes.
Outcome I2 (%) and p-value from Cochran’s Q test 6 weeks 3 months 6 months 12 months 24 months ODI 29 p=0.17 0 p=0.57 1 p=0.43 0 p=0.51 38 p=0.08 SF-36 PCS 0 p=0.82 0 p=0.71 0 p=0.60 7 p =0.37 0 p=0.67 Back pain 17 p=0.29 10 p=0.35 0 p=0.56 33 p=0.14 44 p=0.07 Leg pain 0 p=0.43 0 p=0.86 0 p=0.58 0 p=0.54 0 p=0.83
We do not include an analysis of immediate post-operative pain here as data were not provided by Medtronic. Later in this section we consider an analysis of post-operative medication use (which may
34 7 Results: Efficacy be considered a proxy for pain) and in the safety analysis we consider experience of pain where this was reported as an adverse event. Analysis of SF-36 components The 11 Medtronic RCTs provided full data on all responses to the 36 questions in the SF-36 questionnaire. This enabled us to analyse SF-36 across all eight of its included categories, namely the four categories that compose the Physical Component Score (General Health, Pain, Physical function and Physical role) and the Mental Component Score (Mental health, Emotional role, Social function and Vitality). We performed two-stage meta-analyses to investigate the differences between rhBMP-2 and ICBG surgery for these eight components. The results are given in Appendix XIII, Figure 49. Although tending to favour rhBMP-2, there was no clear evidence of differences in mental health outcomes between rhBMP-2 and ICBG (confidence intervals crossed the line of no difference at all time points). In general, rhBMP-2 improved pain and physical function by more than ICBG, with statistically significant differences for the Pain, Physical function, and Physical role categories of about 5 percentage points from three months after surgery onwards. There was no evidence of a benefit of rhBMP-2 for overall general health and so the additional improvement observed in SF-36 PCS score appears to be caused by a reduction in pain.
7.3. Spinal fusion We also investigated whether rhBMP-2 improved the rate of successful spinal fusion. Spinal fusion was defined by Medtronic as requiring all of the following: • Evidence of bridging trabeculae • No evidence of motion (<3mm difference in translation, less than 5 difference in angular motion) • No evidence of radiolucency ⁰ We assumed that fusion success was reported correctly in the data supplied. It was not possible to check whether reported fusion success agreed with the supplied radiological data. We also note that these radiological assessments were subjective so it is not possible to confirm whether reported successful fusions truly were successful. We did not include the Glassman or LT Cage pilot trials in these analyses because they did not provide analysable data on fusion. For each trial at each time period we calculated the relative risk of successful fusion with its associated standard error (on the log scale) and combined these results across trials in two-stage meta-analyses. Note that in the trials fusion was assessed only from six months after surgery. We identified substantial heterogeneity in the relative risk of successful fusion across trials. The I2 statistics for heterogeneity and p-values for Cochran’s Q test are summarised in Table 5, showing that heterogeneity was particularly high at six months.
Table 5 Heterogeneity in meta-analyses of spinal fusion (complete case analysis).
Outcome I2 (%) and p-value for Cochran’s Q test 6 months 12 months 24 months Spinal fusion 97 p<0.0001 80 p<0.0001 76 p<0.0001
This heterogeneity is apparent in Figure 7, which shows the forest plot for successful fusion 24 months after surgery. Several trials have significant results that favour rhBMP-2. Two trials (INFUSE®/INTER FIX™ ALIF pilot RCT and rhBMP-2/BCP US pilot RCT) favour ICBG. Across all trials rhBMP-2 increased the rate of successful fusion at 24 months after surgery by about 12% with a 95% confidence interval consistent with between a 2% and 23% increase compared with ICBG surgery.
35 7 Results: Efficacy
Figure 7 Forest plot of relative risk of successful fusion 24 months after surgery.
36 7 Results: Efficacy
The results of meta-analyses of successful fusion at six, 12 and 24 months are shown in Figure 8. The rate of spinal fusion in the rhBMP-2 group is increased by 10-20% compared with the ICBG group (from the baseline rate of fusion of 69% after 24 months). These rates were consistent over time and were all statistically significant.
Figure 8 Meta-analyses of successful spinal fusion at 6, 12, and 24 months after surgery. Table 6 summarises the proportion of successful fusions by time and treatment across all 11 Medtronic RCTs. Most patients achieved fusion regardless of treatment (69% by 24 months among ICBG patients) but there were more successful fusions among rhBMP-2 patients (81% by 24 months).
Table 6 Successful spinal fusion by time and treatment.
Treatment Percentage of patients with successful fusion 6 months 12 months 24 months rhBMP-2 79 82 81 ICBG 62 70 69
7.4. Impact of type of surgery on efficacy outcomes We investigated whether the efficacy of rhBMP-2 varied with the type of spinal surgery using subgroup analysis. We divided the trials into three categories: • ALIF: Anterior lumbar interbody fusion (open or laparoscopic, five trials) • PLF/PLIF: Posterior lumbar fusion (including interbody fusion, six trials) • Cervical: Cervical spinal surgery (one trial) We performed meta-analyses of all the pain and function outcomes and the dichotomous success outcomes within each category. Forest plots of the results of these analyses are given for ODI at 24 months in Figure 9 and for successful fusion at 24 months in Figure 10. There was moderate evidence of a difference between surgery types for ODI (p-value 0.065) but this was primarily due to the very large benefit in the NDI in
37 7 Results: Efficacy the single small cervical surgery trial. There was no statistically significant difference between ALIF and PLF/PLIF approaches (p = 0.17). There was also heterogeneity among the anterior lumbar trials (I2 = 65%) where the effects in smaller pilot trials are larger than in the pivotal trials.
Figure 9 Meta-analysis of ODI at 24 months according to type of surgery received.
38 7 Results: Efficacy
Figure 10 Meta-analysis of fusion at 24 months according to type of surgery received. There is no square representing the INFUSE®/LT-CAGE® pilot RCT as this trial did not provide usable data for this analysis or for the INFUSE®/CORNERSTONE® ACDF pilot RCT as all patients for whom we had data achieved fusion.
There was no evidence of a difference in the rates of fusion across surgery types (p-value 0.88) and no evidence of any difference between surgery types for any other outcome. We note that in the cervical surgery trial (INFUSE®/CORNERSTONE® ACDF pilot RCT) all patients on both arms achieved fusion at 24 months and yet the trial showed a substantial difference between rhBMP-2 and ICBG in terms of lowering the NDI (by 16 percentage points more than in the ICBG arm). This is surprising as we might predict that reduction in pain is related to successful fusion. We will return to this issue at the end of this section.
7.5. Impact of patient-level factors on efficacy outcomes We considered whether a range of patient-level factors could alter the effectiveness of rhBMP-2 in surgery: • Age • Sex • Smoking • Alcohol consumption • Body mass index • Diabetic status • History of spinal surgery We performed one-stage analyses to investigate whether these factors were associated with effects of rhBMP-2 on ODI, SF-36 PCS and back pain and leg pain. For each outcome and each factor we fitted one-stage random-effects linear regression models as described in the methods section.
39 7 Results: Efficacy
The results of these regression models are summarised for ODI, SF-36 PCS, back pain and leg pain in Appendix XIII, Table 29. In general there was little evidence for any interactions between patient- level factors and the effectiveness of rhBMP-2 surgery. No particular type of patient (e.g. older or younger, male or female, smoker or non-smoker) appeared to benefit any more or less than average after rhBMP-2 surgery. We performed similar one-stage regression models to investigate potential interactions between patient-level factors and rates of fusion. These results are summarised in Appendix XIII, Table 30. In general there is little evidence of an impact of patient-level factors on the efficacy of rhBMP-2. One possible exception is that in people with a history of spinal surgery, rhBMP-2 is no more effective than ICBG surgery in terms of reduction in ODI or in improved fusion rates. (The relative risk of an interaction between previous surgery and rhBMP-2 surgery was +2.35 percentage points, 95% CI 0.35 to 4.34; very similar to the overall benefit of rhBMP-2). Given the number of analyses performed this result is compatible with a chance finding.
7.6. Sensitivity analyses We performed sensitivity analyses to investigate the effects of including or excluding certain patients or trials from our analyses. We investigated whether excluding the trial that was not sponsored by Medtronic altered the results. Figure 11 shows the results of meta-analyses of only the 11 Medtronic RCTs, excluding the Glassman trial, in a similar fashion to Figure 4. These results were very similar to those including the Glassman trial.
Figure 11 Meta-analyses of pain and function outcomes for Medtronic trials only. We re-performed the primary meta-analyses with the 11 patients in the rhBMP-2/BCP US pilot RCT who received a different fixation procedure incorporated into the rhBMP-2 arm. This had minimal effect on the analyses. A forest plot of mean difference in ODI at 24 months after surgery is shown in Appendix XIII, Figure 50, and for successful fusion in Figure 51. None of our results would be altered significantly by including these 11 patients. We also performed meta-analyses of only the lumbar spine trials, excluding the single cervical spine trial (INFUSE®/CORNERSTONE® ACDF pilot). This avoids making the assumption, for example, that NDI is equivalent to ODI. This also had minimal effect on the analyses. Forest plots of mean
40 7 Results: Efficacy difference in ODI and successful fusion at 24 months after surgery are shown in Appendix XIII Figure 52 and Figure 53. In response to peer review comments we carried out additional sensitivity analyses to investigate whether results of pilot studies differed from the later “pivotal” studies. The results for ODI at 24 months after surgery are shown in Appendix XIII, Figure 54. For pilot studies the mean improvement in ODI at 24 months is almost seven percentage points greater for the rhBMP-2 group than for the ICBG group but for pivotal studies the difference is less than two percentage points. The test for difference between pilot and pivotal studies was not statistically significant (p = 0.21). There were no differences between pilot and pivotal trials for any other outcome. We carried out an additional sensitivity analysis to explore whether findings differed between the INFUSE® and AMPLIFY™/BCP preparations of rhBMP-2. The results for ODI at 24 months after surgery are shown in Appendix XIII, Figure 55. The INFUSE® trials showed a greater benefit of rhBMP-2 with a difference in improvement in ODI between treatment arms of 6.26 percentage points, compared to 1.02 in the AMPLIFY™ and BCP trials. The difference between the groups was not quite statistically significant (p=0.068). As all the pilot trials used INFUSE® it is not clear whether this difference is due to a difference between INFUSE® and AMPLIFY™ or between pilot and pivotal trials. The comparison of INFUSE® and AMPLIFY™ or BCP trials for spinal fusion at 24 months is shown in Appendix XIII, Figure 56. There was no evidence of a difference between trials (p = 0.63). One-stage analyses We performed one-stage meta-analyses for the pain and function outcomes on the 11 Medtronic RCTs with the Glassman trial. Results were similar to the two-stage analyses presented above. The exception was ODI, where the mean difference between arms was four percentage points from three months onwards rather than the two to four percentage points observed in the two-stage analysis. (see Appendix XIII). We used the results from the one-stage models to calculate the typical absolute changes in pain scores rather than the differences between treatment groups. Figure 12 shows the observed improvement from pre-operative levels in ODI, SF-36 PCS and back and leg pain scores over time for both rhBMP-2 and ICBG patients. Patients in both groups improved considerably over time. For example, ODI scores for ICBG patients improved by 25 percentage points by 24 months after surgery. The extra benefit of rhBMP-2 over ICBG was small compared with the typical improvements in pain and function experienced.
Figure 12 Reduction in pain scores from pre-operative results by treatment received.
41 7 Results: Efficacy
Missing data As some outcome data were missing, particularly at 24 months (see Table 3), we performed a multiple imputation analysis to investigate the sensitivity of our results to these missing data. The results from the multiple imputation for the pain outcomes are shown in Appendix XIII, Figure 58 and for fusion in Appendix XIII, Figure 59. These analyses and the multiple imputation analysis for spinal fusion gave almost identical results to the complete case analyses so there was no evidence that missing data were causing any bias in the results.
7.7. Secondary outcomes We performed meta-analyses for the following secondary outcomes: • Duration of hospital stay • Operating time • Successful return to work or usual activity • Use of pain relief medication These analyses are based only on the 11 RCTs supplied by Medtronic because these outcomes were not available for the Glassman trial. Duration of hospital stay Mean duration of hospital stay was 4.2 days among ICBG surgery patients and 3.9 days for rhBMP-2 patients. Figure 13 shows a forest plot from a two-stage meta-analysis of mean difference in duration of hospital stay between patients in the rhBMP-2 and ICBG arms. This shows no clear evidence of any difference in hospital stay in any trial. Correspondingly, the overall difference in duration of hospital stay was not statistically significantly different between the two groups (mean difference -0.15 days, 95%CI -0.33 to 0.03 days). Because duration of hospital stay was measured in whole days and was not normally distributed we performed a one-stage Poisson regression for this data to better represent this distribution. This model showed no clear evidence of any difference in duration of hospital stay (mean difference -0.05 days, 95%CI -1.11 to 0.003 days).
There is no square representing the INFUSE®/LT-CAGE® pilot RCT or INFUSE®/MASTERGRAFT® pilot RCT as they did not provide usable data for this analysis. Figure 13 Forest plot of mean difference in duration of hospital stay.
42 7 Results: Efficacy
Operating time The median operating time across all patients receiving ICBG surgery was 135 minutes and across all patients receiving rhBMP-2 was 117 minutes. Figure 14 shows a forest plot from a two-stage meta- analysis of mean difference in operating time. There was clear evidence that using rhBMP2 shortened operating times by 21 minutes with 95% confidence intervals consistent with 15 to 27 minutes. To account for non-normality in operating times we performed a one-stage random-effects log-linear regression of the logarithm of the operating time. This also had a statistically significant result, with operating times reduced by 15% (95% CI 9% to 21%); that is, by approximately 18 minutes. We conclude that using rhBMP2 does reduce operating time, presumably because it does not require extraction of bone from the iliac crest.
There is no square representing the INFUSE®/LT-CAGE® pilot RCT or INFUSE®/MASTERGRAFT® pilot RCT as they did not provide usable data for this analysis. Figure 14 Forest plot of mean difference in operating time.
43 7 Results: Efficacy
Successful return to work or normal activity We performed a two-stage meta-analysis of successful return to work or normal activity at each of the time periods provided in the Medtronic data. The results of the meta-analyses across time periods are shown in Figure 15. Approximately 24% of all patients had returned to work or normal activity by 24 months after surgery. There was no clear evidence that rhBMP-2 surgery increased a patient’s chance of returning to work when compared with receiving ICBG surgery at any point between three and 24 months after surgery (relative risk at 24 months 1.01, 95% CI, 0.88 to 1.17).
Figure 15 Meta-analysis of successful return to work.
44 7 Results: Efficacy
Use of pain relief medication Medtronic provided data on the use of pain relief medication after surgery. Drugs were classified as weak or strong narcotics, non-narcotics or muscle relaxants. We performed a two-stage meta-analysis for use of these drugs at each time period. Data were supplied giving an approximate level of use of each drug (not at all, once per week, three or more per day). We performed an analysis of any use of pain-relief drugs regardless of frequency of use. The results are shown in Figure 16. This showed no evidence that use of pain-relief medications differed between rhBMP-2 and ICBG arms at any time. If rhBMP-2 surgery were reducing pain levels we might expect to see a reduced level of drug use over time but this does not seem to be the case. These results contrast with earlier analyses of pain where modest improvements in ODI or back pain were observed with rhBMP-2. It may be that these small improvements in pain are insufficient to alter a patient’s use of pain-relief medication.
Figure 16 Meta-analysis for use of pain-relief medication.
45 7 Results: Efficacy
7.8. Additional analyses: neurological and overall success We performed analyses on two further outcomes: neurological success and overall success. These were not part of the planned analyses set out in our protocol but these data were reported in detail in all the Medtronic trials, were judged to be of clinical relevance and were analysed for completeness. Both were composite outcomes defined by Medtronic as follows: • Neurological success: maintenance or improvement in neurological status (consisting of motor, sensory, reflex and leg raising measurements) • Overall success (requiring all of): Successful fusion Maintenance or improvement in neurological status ODI improvement of 15 percentage points or more No associated adverse events No additional surgery The results of meta-analyses of neurological and overall success are shown in Figure 17. Overall success was reported only from six months after surgery onwards. There was no evidence that rhBMP-2 improves neurological function. The results for overall success were very similar to those for successful fusion in Figure 8.
Figure 17 Meta-analyses of neurological and overall success.
46 7 Results: Efficacy
7.9. The relationship between fusion and pain We noted above that the cervical surgery trial (INFUSE®/CORNERSTONE® ACDF pilot RCT) showed no difference in fusion rates between the rhBMP-2 and ICBG arms (all patients for whom data were available achieved fusion) but did show a substantial benefit of rhBMP-2 in reducing the NDI score. This is surprising because if, as is generally believed, fusion reduces pain we would not have expected to see such a difference in NDI. We therefore explored in unplanned analyses (at trial level) the relationship between relative risk of successful fusion and improvement in ODI, SF-36 PCS, back pain and leg pain across all the Medtronic RCTs at 24 months after surgery. Results are illustrated in Figure 18.
Figure 18 Relationship between relative risk of fusion and mean difference in effectiveness outcomes across trials 24 months after surgery. This figure does not indicate a consistent relationship between improvements in fusion due to rhBMP- 2 and improvements in pain or function. If successful fusion resulted in reduced pain we would expect trials with higher fusion rates on rhBMP-2 to show greater improvement in pain scores but this was not the case. In particular, trials where fusion was less common in the rhBMP-2 patients (INFUSE®/INTER FIX™ ALIF pilot RCT and rhBMP-2/BCP US pilot RCT) still showed a benefit of rhBMP-2 in terms of ODI and SF-36. Therefore, the apparent benefits of rhBMP-2 in pain reduction do not appear to be due to increased fusion rates. We investigated this further by performing a one-stage regression of change in ODI from baseline against treatment and whether or not a patient had achieved fusion at six, 12 or 24 months after surgery. This analysis takes account of trial by assuming random effects across trials for the intercept term in the model (see Section 3.5). From this model we calculated the predicted average improvement in ODI for each of the following categories: • Received ICBG surgery and achieved fusion • Received ICBG surgery: no fusion • Received rhBMP-2 and achieved fusion • Received rhBMP-2: no fusion
47 7 Results: Efficacy
We repeated the analysis for SF-36 PCS and back pain. The results are shown in Figure 19.
Figure 19 Mean improvement in pain and function according to treatment and fusion status at six, 12, and 24 months. For patients receiving ICBG surgery, successful fusion had almost no impact on back pain or SF-36 PCS. Thus it is not apparent from these data that achieving fusion necessarily translates to improved patient outcomes. Twelve months after surgery there was a statistically significant 11 percentage point greater improvement on ODI among patients with successful fusion than for those who did not achieve fusion. At 24 months after surgery this reduced to 4.5 points and the difference was not statistically significant. For rhBMP-2 patients there were significantly larger differences in ODI and back pain (but not SF-36 PCS) between patients who did and did not achieve fusion. Greater (and statistically significant) benefits were reported by those who fused compared with those who did not. Also, rhBMP-2 patients with fusion had slightly better outcomes than patients with fusion who received ICBG and rhBMP-2 patients without fusion had worse outcomes than patients without fusion who received ICBG. If successful fusion was all that was required for improvements in pain and function then there should be no difference between these groups. A possible explanation for this is that patient knowledge of both treatment and fusion status led to bias in the reporting of pain and function outcomes. We repeated this analysis without the outlying INFUSE®/CORNERSTONE® cervical spinal trial: this did not alter the results.
48 7 Results: Efficacy
7.10. Summary of efficacy results: In general, the use of rhBMP-2 in spinal fusion surgery appears to have modest benefits when compared with ICBG surgery at up to 24 months of follow-up. Use of rhBMP-2 reduced back pain measured on the ODI by two to three percentage points more than ICBG and improved function as measured by SF-36 PCS by about an additional two percentage points (or five points on the pain sub- scale). Whether or not these modest benefits are clinically relevant is discussed in section 10. There was clear evidence that rhBMP-2 surgery improves fusion rates by around 12% more than ICBG within 24 months after surgery; approximately 70% of patients who received ICBG surgery have achieved fusion by that time. There was substantial heterogeneity across trials in many meta-analyses and particularly for spinal fusion. We found no evidence that treatment effects varied according to the surgical approach used or according to patient factors such as age or sex. The use of rhBMP-2 reduced operating times by about 20 minutes. We found no evidence of any benefit in terms of duration of hospital stay, whether patients returned to work or on post-operative use of pain medication. There was no evidence of a consistent relationship between improvements in fusion due to rhBMP-2 and improvements in pain or function. If successful fusion resulted in reduced pain, we would expect trials that showed higher fusion rates with rhBMP-2 to show greater improvement in pain scores but this did not seem to be the case. In particular, trials where fusion was less common in the rhBMP-2 recipients (Interfix ALIF pilot and BCP US) still showed a benefit of rhBMP-2 in terms of improved SF- 36 PCS score. The BCP US trial also showed a benefit of rhBMP-2 on ODI score and back pain. Therefore, the apparent benefits of rhBMP-2 in pain reduction do not seem to be due to increased fusion rates. A possible explanation for this is that patients’ knowledge of both treatment and fusion status led to bias in the reporting of pain and function outcomes. Overall, the trials show that rhBMP-2 surgery leads to an increase in the rate of successful spinal fusion and to a modest improvement in level of pain reduction when compared with ICBG surgery. The improvement in pain does not seem to be as a result of the increase in fusion and the possibility that the moderate benefits of rhBMP-2 are primarily a consequence of bias due to patients’ knowledge of the surgical intervention cannot be excluded.
49 8 Results: Safety
8. Results: Safety In this section we present results of analyses of the safety of rhBMP-2 surgery in terms of the numbers of adverse events experienced by patients receiving rhBMP-2 when compared with other spinal surgeries. We first examine adverse events in the Medtronic trials in an IPD analysis and then investigate adverse events reported in non-randomised studies of spinal surgery published in the wider literature and in MedWatch report forms.
8.1. Adverse events in the Medtronic trials The IPD provided by Medtronic included data on adverse events at or shortly after (up to four weeks) the time of surgery and at six weeks, three, six, 12 and 24 months after surgery. Data were available from all 17 trials. Data were limited to a categorisation of each adverse event (as back pain, gastrointestinal, neurological and so on) and an indication of its severity and relatedness to spinal surgery as judged by the assessing physician. Some trials provided data at time points beyond 24 months but as these data were not available for all patients in any trial we do not consider adverse events beyond 24 months after surgery. The Medtronic trials used structured data collection forms to collect up to 22 adverse event categories. The IPD provided included 56 categories as described in Appendix XIV, Table 31. Some categories match or map directly between the two and others do not. For example, vertigo is included as a category within the IPD but is not a category used in the adverse event forms. It is not clear whether the additional adverse event data were recorded as “other” events on the data collection forms during routine monitoring or whether they were reported spontaneously. Individual-level data on adverse events were provided for the Glassman trial but these data did not indicate when the adverse events occurred. The adverse event tables included in each of the Medtronic CSRs and a summary table of adverse events reported in the Medtronic IPD are reproduced in Appendix XIV, Tables 32 to 48 and Appendix XV. Checking of individual participant data and risk of bias The large numbers of adverse events recorded across the trials meant that it was not feasible to check the adverse event IPD against the information provided in the CSR (this was never our intention and until these data were received we were unaware that summary patient narratives would be included in the CSRs). However, owing to differences between the published paper and the IPD for the INFUSE®/INTER FIX™ PLIF RCT we compared the supplied IPD with the tabulated adverse event classifications provided for each patient in the CSRs. We also compared informally the summary patient narratives with the adverse events table provided for each patient in the CSR for this trial. We found that IPD were consistent with the tabulations of adverse events for each patient in this trial and with the summary tables of adverse events presented in the CSR. The definitions of severity of the event and its relatedness to the surgery or device used appeared reasonable. We found some inconsistencies in reporting. In particular, ongoing adverse events were generally only reported once in the IPD (at the time they were first identified). Pain appeared to be underreported, particularly where it coincided with neurological problems (such as numbness or tingling) or with spinal events such as stenosis. In those circumstances a separate adverse event for pain was not generally recorded. As previously discussed under efficacy, randomisation and allocation concealment was judged to be robust and there was a low risk of selection, attrition or reporting biases. However, the lack of patient and assessor blinding regarding the reporting of adverse events left these trials at high risk of performance and detection bias, see Figure 20.
50 8 Results: Safety
Figure 20 Risk of bias judgments of trials included in safety analyses.
51 8 Results: Safety
Analysis of adverse events Data on the types of adverse event specified in our protocol were not generally available in the Medtronic-supplied IPD. Therefore, categorisation of adverse events used in this analysis does not precisely match our protocol-specified set of adverse events. In particular, data relating to heterotopic bone formation, osteolysis and radiculitis were not generally available or discernible from the IPD. It was not clear whether these events did not occur or occurred but were not recorded. Radiculitis, for example, may have been recorded as back/leg pain or as a neurological event. Data from analyses described as outside the protocol were provided for three trials. Heterotopic bone formation (described as anterior bone growth) was provided for the INFUSE®/INTER FIX™ PLIF RCT. Data on osteolysis (described as transient localised areas of bone remodelling or resorption in the vertebral body adjacent to the allograft dowel) was provided for the rhBMP-2 arm of the INFUSE®/Bone dowel pilot and INFUSE®/Bone dowel pivotal RCTs. These have been included within our analyses, where possible. We consider here only adverse events that are potentially related to spinal surgery. The categorisation of adverse events is that used by Medtronic, with some simplification because coding of adverse events varied across trials. Prevalence of adverse events in the 17 Medtronic trials is summarised in Table 7. This table shows numbers of patients who experienced each type of adverse event at any time during the first 24 months after surgery and also just at or shortly after surgery. Some patients experienced multiple adverse events (either different types of adverse event at a particular time or multiple incidences of the same type of event) and these patients are counted only once in each cell of the table. All 17 trials are included in the table but the MAVERICK™ disk system arm of the MAVERICK™ Disc pivotal RCT is excluded. Table 7 distinguishes between patients on rhBMP-2 in the RCTs and in the single-arm trials (including the MAVERICK™ Disc pivotal RCT) to enable a direct comparison of adverse events within the RCTs.
52 8 Results: Safety
Table 7 Adverse events in Medtronic trials. Nature of adverse event At or shortly after initial Across all times surgery
Single - Single - arm arm trials RCTs trials RCTs rhBMP- rhBMP- rhBMP- 2 2 ICBG rhBMP-2 2 ICBG Total number of patients 381 696 610 381 696 610
Arm & neck (upper extremity) pain 22 13 12 2 1 2 Arthritis or bursitis 0 24 16 0 3 1 Back & leg (lower extremity) pain 142 268 207 26 39 24 Cardiovascular 18 93 99 7 60 65 Death 1 6 10 0 0 2 Dural injury 1 13 12 1 11 10 Dysphagia 0 3 2 0 1 2 Gastrointestinal 52 117 115 31 67 73 Graft related ------45 ------17 Heterotopic bone formation* --- 24 4 ------Implant 18 22 10 4 10 6 Infection 34 104 90 13 53 47 Neurological 89 152 119 13 29 20 Osteolysis** --- 12 ------Other pain 25 97 63 1 12 9 Respiratory 7 34 27 3 19 16 Retrograde ejaculation 8 5 1 3 3 1 Spinal 50 70 63 2 4 7 Trauma 93 153 137 8 8 8 Urogenital 40 81 63 17 42 35 Vascular 4 6 4 1 6 4 Vertebral fracture 0 3 1 0 3 0 Wound complication 0 6 4 0 4 3 * Only available in INFUSE®/INTER FIX™ PLIF RCT ** Only reported in rhBMP-2 arms of INFUSE®/Bone dowel trials as additional data beyond protocol. Data on adverse events were provided for the Glassman trial but were reported differently and did not indicate the times when the adverse events occurred. The results are summarised in Appendix XVI, Table 49. The trial is too small for any patterns in adverse events to be detected. Adverse events in the eleven Medtronic RCTs In order to compare formally the prevalence of adverse events in patients receiving rhBMP-2 and ICBG surgery we conducted meta-analyses only on the 11 Medtronic RCTs excluding the single-arm trials, the MAVERICK™ Disc pivotal RCT (no ICBG group) and the INFUSE®/CORNERSTONE® ACDF pivotal RCT (due to insufficient adverse event data; only three participants were recruited). We performed all analyses of the Medtronic IPD in this section using one-stage meta-analysis methods, combining all data from all trials in a single random-effects logistic regression model (stratified by trial). This was because many trials had few or no adverse events in any specified category and the use of one-stage models avoided the need to make adjustments to data in such
53 8 Results: Safety cases. Results are presented as odds ratios (OR) as these are more robust than relative risks when there are few adverse events in some of the trials. We performed meta-analyses across the trials to determine the odds ratio of experiencing each adverse event at least once during the first 24 months after surgery (i.e. patients who experienced that event) and specifically at or shortly after surgery (up to four weeks after surgery). Figure 21 gives the results of the meta-analyses of adverse events at or shortly after surgery on a forest plot. Each bar on the plot gives the result of the meta-analysis for the specified category of adverse event. Figure 22 similarly shows the results of these meta-analyses at any time during the first 24 months after surgery. Figure 21 shows that at or shortly after surgery some adverse events were more common among rhBMP-2 patients: arthritis/bursitis, back and leg pain, implant-related events, neurological events, other pain, retrograde ejaculation, wound complications and vascular events all had at least a 50% increase in risk of occurring among rhBMP-2 patients. Because of the rarity of most events, these differences were statistically significant only for back and leg pain (OR 1.92, 95% CI 1.14 to 3.25). Only arm and neck pain, spinal events and dysphagia were more common among ICBG patients but again events were few and none of these results were statistically significant.
Figure 21 Meta-analysis of adverse events by category at time of surgery in 11 Medtronic RCTs. Figure 22 indicates that for most categories of adverse events the risk of experiencing at least one event across the first 24 months after surgery was similar for rhBMP-2 and ICBG surgery patients. Back and leg pain was more common in rhBMP-2 patients (OR 1.20, 95% CI 0.99 to 1.46) as was “other pain” (OR 1.39, 95% CI 1.00 to 1.92), which was statistically significant at the 5% level. Heterotopic bone formation, data for which were from one trial, does not appear on the forest plot as the odds ratio is extreme and lies beyond the axis of the plot (OR 27.4, 95% CI 1.5 to 498). This result is also statistically significant. Implant-related events (OR 1.94, 95% CI 0.91 to 4.11), retrograde ejaculation (OR 4.76, 95% CI 0.55 to 40.95) and vertebral fractures (OR 2.79, 95% CI 0.29 to 26.89) were also more common among
54 8 Results: Safety rhBMP-2 patients; there was a small number of these events and the results were not statistically significant. No adverse events were statistically significantly less common among rhBMP-2 patients. The pattern of results is similar for INFUSE trials alone, excluding the AMPLIFY and BCP trials (Appendix XVI, Figure 60).
Figure 22 Meta-analysis of adverse events by category at any time during the first 24 months after surgery in 11 Medtronic RCTs.
55 8 Results: Safety
Specific categories of adverse event We also considered the prevalence of specific categories of adverse events. We performed meta- analyses for the following categories of adverse event: gastrointestinal, implant related infection, neurological, pain (combining back, leg, arm and neck pain) and spinal events to investigate how adverse events varied over the course of the trials. As in the analysis above, a one-stage meta-analysis approach was used and odds ratios calculated to avoid the need to adjust the data when there were few or no adverse events. This analysis is based on the number of patients experiencing an adverse event at a specified time; some patients may have had more than one incidence of an adverse event in any category at a specified time point and some events may be ongoing occurrences of an adverse event from a previous time. We performed separate analyses for adverse events at each of the time periods: at or shortly after surgery; six weeks after surgery; and three, six, 12 and 24 months after surgery. Figure 23 shows the results of these meta-analyses for each category of adverse event at all time periods.
Figure 23 Meta-analyses of adverse events according to adverse event category. At or shortly after surgery there was inconclusive evidence that implant-related events were more common in rhBMP-2 patients (OR 1.87, 95% CI 0.67 to 5.21), as were neurological events (OR 1.58, 95% CI 0.88 to 2.83). Only for pain was there clear evidence that this was more common in the rhBMP-2 patients at or shortly after surgery (OR 1.78, 95% CI 1.06 to 2.95, p = 0.007). This contrasts with the results seen in the analyses of ODI, SF-36 PCS and back pain in section 7 where pain was generally lower in the rhBMP-2 patients from three months after surgery onwards. The meta-analyses did not identify evidence of any difference in the risk of experiencing at least one adverse event from three months after surgery onwards, except for spinal events. Spinal events were more common among rhBMP-2 patients 24 months after surgery but as there were no differences in prevalence of spinal events at earlier times this result may be due to chance.
56 8 Results: Safety
8.2. Sensitivity and supplementary analyses of adverse events in the Medtronic RCTs In response to peer review comments, we carried out sensitivity analyses that were restricted to trials that used the INFUSE® preparation of rhBMP-2. This is shown in Appendix XVI, Figure 60 for adverse events observed at any time during the first 24 months after surgery. There are no apparent differences in the pattern of results observed between this and figure 21, which shows the same analyses based on all trials including AMPLIFY™ and BCP preparations. As numbers of events are reduced by excluding trials, the confidence intervals are wider and no categories of event show statistically significant differences between rhBMP-2 and ICBG. This should be interpreted bearing in mind that the pattern of results is the same and that estimates of adverse effect are very similar between the two plots. Severity of adverse events Reporting of adverse events in the Medtronic trials included classification of events according to the perceived severity of events and their relatedness to the type of surgery performed. These classifications were the judgement of the reporting physician and the reliability of these categorisations could not be fully assessed but examination of the CSR of the INFUSE®/INTER FIX® PLIF RCT suggested that these categorisations were reasonable. These were not prespecified in our protocol but for completeness we performed meta-analyses of adverse events according to these categorisations. Adverse events were classified in the Medtronic trials as being one of: • Mild • Moderate • Severe • Life-threatening We performed one-stage meta-analyses of adverse events at each time point according to these categorisations. For each patient we considered whether they had at least one adverse event of each severity at each time, for example a patient was classified as having a “moderate or worse” event if they had at least one adverse event classified as moderate, severe or life-threatening. All these analyses were based on the 11 Medtronic RCTs, excluding the single-arm trials, the MAVERICK™ Disc pivotal RCT and the INFUSE®/CORNERSTONE® ACDF pivotal RCT. The results of these meta-analyses are shown in Figure 24. The results show that, in general, adverse events were more common at the time of surgery for patients receiving rhBMP-2 but incidence was similar thereafter. Only “mild” and “moderate” adverse event were more common in the rhBMP-2 group; there was no evidence of any difference in numbers of “severe” or “life-threatening” events between treatments.
57 8 Results: Safety
Figure 24 Meta-analyses of adverse events according to severity in 11 Medtronic RCTs. As a sensitivity analysis we repeated this analysis of adverse event severity for the trials using INFUSE®. The results are shown in Appendix XVI, Figure 61. There is no evidence that the pattern of incidence of adverse events in the INFUSE® trials is different from that including the AMPLIFY™ and rhBMP-2/BCP trials (in figure 24). Device and surgery-related adverse events Medtronic also classified adverse events as being related to the device use, related to the surgery performed or as being unknown or unrelated. These categorisations were determined by the physician reporting the event. We performed meta-analyses of only these “related” events with results shown in Figure 25. This analysis shows that surgery-related events were more frequent in rhBMP-2 patients at the time of surgery (OR 1.41, 95% CI 1.04 to 1.91), which was statistically significant. Device or surgery-related events combined were also more common among rhBMP-2 patients at the time of surgery (OR 1.32, 95% CI 1.02 to 1.71). Related adverse events were generally less common among rhBMP-2 patients at later times. There were very few of such events (only five surgery-related events and 14 device-related events at 24 months across both treatment arms) and consequently results at later times had very wide confidence intervals and were not statistically significant.
58 8 Results: Safety
Figure 25 Meta-analyses of adverse events according to relatedness to device use or surgery in 11 Medtronic RCTs. Further surgery The Medtronic data also reported whether patients received any further spinal surgery because of perceived implant failure, failure to fuse or for other reasons. A summary of the numbers of additional surgeries performed and the reasons given for the surgeries is given in Appendix XVI, Table 50. We performed a meta-analysis by classifying each patient according to whether or not they had further spinal surgery at any time after the original surgery. This analysis ignores the fact that many patients had multiple further surgeries. The results of this analysis are shown in Figure 26. There is no clear evidence of a difference in the proportion of patients receiving second surgery between rhBMP- 2 and ICBG surgery patients as all the confidence intervals cross the line of no difference.
59 8 Results: Safety
Figure 26 Forest plot of second surgeries.
8.3. Summary of IPD safety analyses These analyses found that some types of adverse events were more common among patients receiving rhBMP-2 than among those receiving ICBG. In particular, pain was more common. There was a 20% increase in risk of back and leg pain and a 39% increase in risk of other pain (i.e. not back, leg, arm or neck pain). Some other adverse events were also more common among rhBMP-2 patients, including arthritis, implant-related events and retrograde ejaculation. All these events were rare so the increases in risk were not statistically significant. It appears that this increased risk of adverse events among rhBMP-2 patients is primarily due to an increased risk immediately after surgery (up to four weeks after). Back and leg pain was almost twice as common among rhBMP-2 patients immediately after surgery. Most other categories of event were also more common among rhBMP-2 patients, although none was statistically significantly so. From six weeks after surgery onwards rates of adverse events were similar across treatment groups. When considering the severity of events (as defined by Medtronic) mild and moderate events were significantly more common among rhBMP-2 patients immediately after surgery but there was no difference in the incidence of serious adverse events. Adverse events related to surgery (as defined by Medtronic) were more common among rhBMP-2 patients immediately after surgery. Device and surgery-related events were less common among rhBMP-2 patients from six weeks after surgery onwards but the differences were not generally statistically significant. The use of rhBMP-2 did not reduce the need for further spinal surgery.
8.4. Cancer Medtronic provided information on all cases of cancer observed in their 17 trials as part of the supplied IPD and a table that gave details of the types of cancer that occurred in the trials.
60 8 Results: Safety
Table 8 provides a summary of the cancers observed in the 11 Medtronic trials that used the INFUSE® preparation of rhBMP-2. We did not include the MAVERICK™ disk system arm from the MAVERICK™ Disc pivotal RCT as it did not meet our inclusion criteria; however, the rhBMP-2 arm of this trial is included with the single-arm studies. No cases of cancer were reported in the Glassman trial.
Table 8 Cases of cancer in 11 Medtronic trials using INFUSE®. Cancer Single-arm RCTs trials (8 trials) (3 trials) rhBMP-2 rhBMP-2 ICBG
Skin cancers Melanoma 1 1 0 Squamous cell carcinoma 1 0 0
Breast 1 2 1 Carcinoid tumour (appendix) 1 0 0 Liver 1 0 0 Pancreatic 0 1 0 Testicular 0 1 0 Thyroid 0 2 0
Total cancers 5 7 1 Total patients 337 335 280
This table excludes one cancer identified before surgery (a pancreatic cancer in the INFUSE®/MASTERGAFT® pilot RCT, in the rhBMP-2 arm). It shows three cancers in the rhBMP-2 arm of the INFUSE®/LT-CAGE® Open pivotal RCT (one thyroid, one testicular and one melanoma) that were identified at 36 months during extended follow-up of only the rhBMP-2 recipients. We did not include these three cancers in our quantitative analyses as this would have biased against rhBMP-2 as some cancers may have occurred among ICBG patients after the end of primary follow- up at 24 months and would not have been identified. Three RCTs recorded cancer events (INFUSE®/LT-CAGE® Open pivotal, INFUSE®/Bone dowel pilot and INFUSE®/Bone dowel pivotal). Two single-arm studies (INFUSE®/LT-CAGE® Laparoscopic pivotal and MAVERICK™ Disc pivotal) also recorded cancer events. Trials that observed no cancers are included in the total number of patients. Table 9 summarises the cancer cases in the AMPLIFYTM and rhBMP-2/BCP trials (reported separately because the rhBMP-2 preparation differs from the INFUSE® trials) where a higher dose was used. Two RCTs recorded cancer events (AMPLIFY™ and rhBMP-2/BCP Canada pivotal); the two single-arm trials using AMPLIFY™ (rhBMP-2/CRM Two-level pilot) and rhBMP-2/BCP (rhBMP- 2/BCP Mexico pilot) did not record cancer events.
61 8 Results: Safety
Table 9 Cases of cancer in five Medtronic trials using AMPLIFY™ or rhBMP-2/BCP. Cancer Single-arm trials RCTs (2 trials) (3 trials) rhBMP-2 rhBMP-2 ICBG Skin cancers Basal cell carcinoma 0 0 1 Melanoma 0 1 0 Squamous cell carcinoma 0 1 1
Breast 0 0 1 Colon 0 1 1 Larynx 0 1 0 Leukaemia 0 1 0 Lung 0 1 0 Non-Hodgkin’s lymphoma 0 1 1 Ovarian 0 1 0 Pancreatic 0 1 0 Prostate 0 2 1 Stomach 0 1 0 Thyroid 0 1 1
Total cancers 0 13 7 Total patients 44 359 328
Because events were sparse we performed a one-stage meta-analysis of cancer incidence. This analysis considered all 11 Medtronic RCTs. The pre-existing cancer and the three cancers identified through additional follow-up that occurred only on the investigational arm of the INFUSE®/LT-CAGE Open pivotal RCT have been excluded. No data from single-arm studies are included. The overall relative risk of cancer among rhBMP-2 patients was almost double that of the ICBG patients but the result was not statistically significant (RR 1.98, 95% CI 0.86 to 4.54) with the confidence interval consistent with a 14% reduction in risk and a 454% increase. We performed a two-stage meta-analysis excluding the same cancers as in the one-stage analysis above and this gave similar results (RR 1.84, 95% CI 0.81 to 4.16). We performed a subgroup analysis to compare cancer risk between those trials using INFUSE® and those using AMPLIFY™ or rhBMP-2/BCP. The forest plot for this analysis is given in Figure 27. It shows a higher risk of cancer in the rhBMP-2 patients in the INFUSE® trials (RR 2.03, 95% CI 0.40 to 10.38) than in the AMPLIFYTM or BCP trials (RR 1.60, 95% CI 0.47 to 5.47). There was no evidence of any heterogeneity in the relative risk of cancer across trials (I2 = 0). In particular, there was no evidence of that the risk of cancer in the AMPLIFYTM trial differed from that in the INFUSE® trials (test for interaction p = 0.82).
62 8 Results: Safety
Figure 27 Forest plot of cancer incidence in the Medtronic RCTs. As an additional analysis we included the extra three cancer patients in the INFUSE®/LT-CAGE Open pivotal RCT identified through extended follow-up in our one-stage analysis. This increased the relative risk to 2.33 and led to a statistically significant result (95% CI 1.04 to 5.25) consistent with an increased risk of cancer of between 4% and 524%. We note that this result will overestimate the relative risk if there were additional cancers in the ICBG arm of this trial that were not identified. A forest plot for a two-stage analysis including these three cases is given in Appendix XVI, Figure 62.
8.5. Safety data from the wider literature RCTs are unlikely to provide much information on rare or long-term adverse effects so we did not restrict our consideration of adverse events to RCTs but additionally considered the wider research evidence on possible complications of rhBMP-2 from published comparative studies. We conducted a systematic mapping exercise which identified 107 publications that reported adverse effects of rhBMP-2. To focus on studies most likely to be informative and to keep the project manageable, we restricted further investigation to 35 comparative studies (reported in 43 publications) that reported on at least 10 adult patients (children were excluded because rhBMP-2 is used in different indications in the immature spine). Bibliographic details of all 64 studies excluded at this stage are given in Appendix IX. Although information from studies that compared rhBMP-2 with ICBG is most directly meaningful to our research question and is consistent with our analyses of comparative effectiveness, we considered studies with other comparators to be informative for adverse effects. Included studies varied in their stated purpose and approach. Some reviewed patient records or re- interviewed patients to identify specified adverse events; others focused on reporting fusion rates with some adverse event data provided alongside. We also identified five large population-based studies that utilised insurance claims or submitted cases to investigate associations between use of rhBMP-2 and subsequent intervention attributable to adverse effects. We tabulated studies according to surgical approach (Table 10 to 13) and then looked across all studies to explore the body of evidence relating to each of the adverse effects that we specified in our protocol and for which data were available. Publications described adverse events using variable
63 8 Results: Safety terminology and we developed a dictionary approach to help non-clinical researchers to extract and record adverse event data appropriately (Appendix VIII). Appendix XVI, tables 51 to 54 show the data extracted from these papers. As described in section 4.6, we used a modified version of the Newcastle-Ottawa-Scale to assess the methodological quality of the non-randomised studies included in these analyses. Any RCTs were quality assessed using the Cochrane risk of bias tool. We did not seek IPD from these additional studies given the sparse data, the analyses of dichotomous data planned (event/no event) and because it was not feasible within the time and resources available. Included studies Each group of studies is briefly described and tabulated below. One paper appears in all three surgical groupings as it separately reported data for anterior cervical fusion, posterior lumbar fusion and overall for spinal fusion (excluding cervical).115 Any RCTs detailed here were not included in the efficacy analyses due to an ineligible comparator arm. Some studies did not specify which form of bone morphogenetic protein was used; these are described as rhBMP in the tabulated data but when describing the body of evidence across studies we have used the shorthand of rhBMP-2 (as most are likely to be rhBMP-2). Lumbar PLIF/TLIF/PLF studies Fourteen studies were in PLIF/TLIF/PLF surgery115-128 and are summarised in Table 10. Quality assessments are tabulated in Appendix XVI, Table 55. Three of the studies were cohort studies and nine were retrospective reviews of patients. One reviewed all spinal fusion cases submitted to the Scoliosis Research Society, which included degenerative disc disease and related spinal conditions. The final study was a small RCT (with a comparator arm that made it ineligible for our review of efficacy).124 Seven studies compared rhBMP-2 with ICBG, two with autologous bone and one each with autograft instrumented fusion, bone marrow aspirate and silicated calcium phosphate). Lumbar ALF/ALIF studies Five studies were in ALIF surgery129-133 and are summarised in Table 11. Quality assessments are tabulated in Appendix XVI, Table 56. Two non-randomised studies compared rhBMP-2 with ICBG and one with bone marrow aspirate. One RCT compared rhBMP-2 with bone marrow aspirate. The Maverick RCT (for which we had IPD) compared rhBMP-2 with a complete disc replacement system (neither RCT was eligible for inclusion in our comparison of effectiveness). Four papers (two conference abstracts and two full publications) reported on multiple retrospective cohorts of patients from the same surgical centre.129, 134-136 We used only the most recent and complete data to avoid double counting of patients.129 Cervical studies Ten studies were in ACDF/PCF surgery115, 137-145 and are summarised in Table 12. Quality assessments are tabulated in Appendix XVI, Table 57. Two of the studies compared rhBMP-2 with ICBG, one with allograft and one with mixed approaches that included ICBG and local bone. The other studies reported only that no rhBMP-2 was used in the control group. Studies that included a range of surgical approaches Eight non-randomised studies reported global findings for more than one type of surgery where it was not possible to split the reported data according to surgical intervention.7, 115, 146-151 These are summarised in Table 13 and quality assessments are tabulated in Appendix XVI, Table 58. Two of the studies compared rhBMP-2 with ICBG, one with allograft, one with allograft or autograft and four reported only that no rhBMP-2 was used in the control group.
64 8 Results: Safety Table 10 Lumbar PLF/PLIF/TLIF Study characteristics (data from the wider literature).
Study ID rhBMP-2 Control Surgery Design/Setting Patient Details AE defined Follow-up Design: Prospective cohort study Anand (2006) 116 4.2mg rhBMP- Allocation: consecutive ICBG Mean 30 2 TLIF patients, last 85 received 100% with DDD No N=15 months N=85 BMP USA Purpose: comparing fusion rates Design: retrospective Glassman review DDD and related (2007) 117 12mg rhBMP-2 ICBG Allocation: control data spinal conditions, Mean 27 PLF Not reported N=91 N=35 taken from AMPLIFY trial including 16 months USA Purpose: assess fusion revision surgeries rate for INFUSE® 118 Design: Consecutive Gray (2010) rhBMP-2 Autologous cohort study DDD and related Unclear 4.2mg per level bone TLIF Allocation: not reported 2yrs Country not spinal conditions (abstract only) N=82 N=39 Purpose: compare reported complication rates Design: observational study Mean 25 Joseph (2007) Defined in 119 4.2 to 8.4mg Autologous Allocation: use of BMP 100% DDD and months relation to rhBMP-2 bone PLIF & TLIF was patient decision related spinal (clinical) and heterotopic N=23 N=10 Purpose: assess conditions 7.9 months Canada bone formation incidence of heterotopic (radiological) bone formation Lee (2010) 120 Design: retrospective rhBMP-2 BMP: 38.3- review USA 4.2mg per level 39.2mths ICBG Allocation: not reported DDD and related Total n=86 PLF Not reported N=41 Purpose: evaluate spinal conditions AEs reported >65yrs n=34 Control: mean efficacy in patients only for >65yrs <65yrs n=52 34.7mths >65yrs patients
65 8 Results: Safety
Study ID rhBMP-2 Control Surgery Design/Setting Patient Details AE defined Follow-up Mannion (2010) Design: comparison of 121 two surgical groups, no rhBMP-2 ICBG further details Requiring lumbar 1.4mg PLIF/TLIF Not defined 12 months UK N=19 Allocation: not reported fusion n=17 Purpose: compare rates (abstract only) of fusion Design: retrospective Mindea (2009) chart review DDD, mechanical 122 4.2 mg/level Autograft + Allocation: reason for back pain, Partially rhBMP-2 allograft TLIF Not reported BMP use not reported recurrent disc described N=35 N=8 USA Purpose: look at post- herniation operative radiculitis 8.4mg rhBMP- 9 months Design: retrospective Mummaneni 2 + cage (mean) 123 review (2004) (adjacent to ICBG + cage DDD and related *mentions only TLIF Allocation: not reported No ICBG + BMP in N=19 spinal conditions having follow- Purpose: comparing USA 14 patients) up data for 40 fusion rates N=25 patients* Pimenta (2011) 124 Silicated Design: RCT rhBMP-2 (dose Calcium Allocation: randomised not reported) Lateral interbody fusion DDD Not defined 36 months Brazil Phosphate Purpose: compare two n=15 N=15 bone substitutes (abstract only) Design: retrospective radiographic review (likely to be follow-up 125 rhBMP-2 Rihn (2009) consecutive patients) DDD or previous mean (dose not ICBG Defined and TLIF Allocation: not reported unsuccessful 19.1mths, reported) N=33 detailed USA Purpose: compare lumbar surgery clinical follow- N=86 complications between up mean ICBG and rhBMP-2 27.6mths Rowan (2011) Primary 126 Design: retrospective instrumented review over 4 years rhBMP-2 lumbar fusion Interested only PLF Allocation: not reported No details reported Not reported UK N=64 without rhBMP- in leg pain Purpose: incidence and 2 cause of leg pain (abstract only) N=40
66 8 Results: Safety
Study ID rhBMP-2 Control Surgery Design/Setting Patient Details AE defined Follow-up 127 Design: consecutive sex- Singh (2006) 12mg rhBMP-2 ICBG matched cohort study DDD and related + ICBG PLF No 24 months N=11 Allocation: not reported spinal conditions USA N=41 Purpose: compare fusion Bone marrow Design: retrospective aspirate with Taghavi (2010) review of consecutive All patients 128 allograft rhBMP-2 12mg patients required Up to 24 N=18 PLF Not reported n=24 Allocation: not reported instrumented months Autograft USA Purpose: compare fusion revision surgery N=20 rates Total n=38 Design: retrospective review of all fusion cases Williams (2011) 115 rhBMP-2 (dose submitted by Scoliosis No rhBMP DDD and related Not defined but not reported) PLF/TLIF/PLIF Research Society Not reported N=18267 spinal disease, examples given n=6049 Allocation: unknown USA Purpose: compare complication rates DDD – degenerative disc disease
67 8 Results: Safety Table 11 ALF/ALIF Study characteristics (data from the wider literature).
Patient Study ID rhBMP-2 Control Surgery Design/Setting AE defined Follow-up Details 129 Design: retrospective Carragee (2012) cohort review of
consecutive patients at USA one surgical centre over
10 years Linked papers rhBMP-2 DDD and Minimum of 1 134 ICBG Allocation: cohorts Carragee (2011) 4.2mg ALIF related spinal Not defined year follow-up N=233 described in detail, based full paper n=239 conditions for each patient. 135 on use of rhBMP-2 Carragee (2007) Purpose: compare abstract 136 incidence of retrograde Carragee (2011) ejaculation, fusion and abstract subsidence Design: matched cohort across two sites Gerszten (2011) 130 rhBMP-2 Allocation: Site DDD and Bone marrow aspirate 8.4mg ALIF dependent related spinal Not defined 2 years N=54 USA n=45 Purpose: comparing conditions fusion rates/clinical outcomes McConnell (2011) 131 rhBMP-2 Design: RCT DDD and (dose not Bone marrow aspirate Allocation: randomised USA ALIF related spinal Not defined 24 months reported) N=32 Purpose: compare conditions n=34 outcomes (abstract only) rhBMP- MAVERICK™ Disc Design: RCT 2/ACS + LT- Maverick total disc DDD and pivotal RCT Allocation: randomised Broadly CAGE replacement ALIF related spinal 24 months Purpose: compare defined (1.5mg/ml) N=405 conditions IPD provided fusion/clinical outcomes n=172
68 8 Results: Safety
Patient Study ID rhBMP-2 Control Surgery Design/Setting AE defined Follow-up Details Design: consecutive cohort of rhBMP-2 compared with historical 132 rhBMP-2 rhBMP-2: mean Pradhan (2006) control DDD and (dose not ICBG 26mths ALIF Allocation: only 9 patients related spinal Not defined reported) N=27 Control: mean USA treated with rhBMP-2 due conditions n=9 36mths to poor results Purpose: fusion comparison Design: consecutively recruited controlled study 133 Allocation: first 30 100% DDD Slosar (2007) 3 to 9mg Allograft patients allocated to and related Minimum of 12 rhBMP-2 ALIF Not defined N=30 control, remainder to spinal months USA N=45 BMP conditions Purpose: compare fusion/clinical outcomes DDD – degenerative disc disease
69 8 Results: Safety Table 12 Cervical (ACDF/ PCF) Study characteristics (data from the wider literature)
Study ID rhBMP-2 Control Surgery Design/Setting Patient Details AE defined Follow-up Design: consecutive patients Butterman over 3years 137 rhBMP-2 0.9 (2008) ICBG Allocation: decision to use BMP DDD and related 24 to 36 to 1.8mg ACDF Not defined N=36 vs. IGBG made by patient spinal conditions months n=30 USA following counselling Purpose: compare fusion rates Design: retrospective review of Crawford (2009) Stenosis, 138 rhBMP-2 4.2 consecutive patients over 3years ICBG spondylosis or to 12mg PCF Allocation: surgeon’s discretion Not defined 3 months N=36 repair of previous n=41 Purpose: compare risk of USA non-union complications Design: retrospective review of rhBMP-2 Various Hiremath (2009) all patients who underwent 139 average (local, 38% rhBMP-2 posterior cervical fusion during a 1.3mg per ICBG, PCF patients previous Not defined Not reported 1yr period level other) failed ACDF USA Allocation: clinical choice n=16 N=67 Purpose: evaluate safety Longley (2009) 140 Design: retrospective case- rhBMP-2: mean
rhBMP-2 No control study 4.8
(dose not rhBMP- Allocation: unclear Control: mean Country ACDF Not reported Not defined reported) 2 Purpose: compare revision 5.8 unknown n=84 N=84 surgeries and adverse event Units not
rates reported (abstract) Lu (2010) 141 Design: retrospective review of rhBMP-2 rhBMP-2: mean matched case-control (dose not Allograft 35 months USA ACDF Allocation: not reported Not reported Defined reported) N=50 Control: mean Purpose: compare dysphagia n=100 25 months (abstract) rates Design: retrospective cohort Smucker (2006) rhBMP-2 No study of consecutive patients Pre- 142 (dose not rhBMP- over 2 years vertebral ACDF Not reported 6 weeks reported) 2 Allocation: surgeon discretion swelling USA n=69 N=165 Purpose: compare prevertebral defined swelling
70 8 Results: Safety
Study ID rhBMP-2 Control Surgery Design/Setting Patient Details AE defined Follow-up Design: retrospective review of consecutive patients Vaidya (2007a) rhBMP-2 No 143 Allocation: first 22 given rhBMP-2 (dose not rhBMP- Cervical disc Up to 24 ACDF Purpose: compare clinical Defined reported) 2 disease months outcomes USA n=22 N=24
Design: retrospective review of Williams (2011) rhBMP-2 all fusion cases submitted by Not defined 115 No (dose not ACDF (also reports PCF but Scoliosis Research Society DDD and related but rhBMP Not reported reported) mixed with thoracolumbar) Allocation: unknown spinal disease, examples N=652 USA n=4532 Purpose: compare complication given rates Design: retrospective review of 144 rhBMP-2 Non- Xu (2011) consecutive patients over 4 yrs Degenerative (dose not rhBMP- Mean 24.2 PCF Allocation: unclear cervical spinal Not defined reported) 2 months USA Purpose: assess safety and conditions n=48 N=156 efficacy Design: retrospective case Yaremchuk rhBMP-2 Non- 145 control study 1mth for (2010) (dose not rhBMP- Cervical spinal arthrodesis Allocation: surgical choice Not reported Defined complications, reported) 2 Purpose: compare complication 3mths for death USA n=260 N=515 rates DDD – degenerative disc disease
71 8 Results: Safety Table 13 Various surgical approach study characteristics (data from the wider literature)
Patient Study ID Intervention Control Surgery Design/Setting AE defined Follow-up Characteristics Design: retrospective cohort of database Cahill (2009) 7 rhBMP-2 (dose No patients DDD or related In patient Not not reported) rhBMP Spinal fusion (all) Allocation: clinical disease AEs defined reported USA n= 17623 N=53026 decision Purpose: compare complications Design: retrospective analysis of insurance Defined need Cahill (2011) 146 rhBMP-2 claims data set, matched DDD and for repeat Minimum ICBG (dose unknown) PLF, ALIF by propensity score related spinal fusion, no 12 N=2372 N=2372 Allocation: unknown conditions other AEs months USA Purpose: comparison of defined repeat fusion risk Design: Medicare patient data retrospective cohort Deyo (2012) 147 rhBMP-2 (dose Not review DDD and not reported) rhBMP-2 Lumbar: ALF, PLF and TLIF Allocation: unknown related spinal Defined 4 years USA n=1,703 N=15,119 Purpose: compare disease complication and reoperation rates Allo and/or BMP: rhBMP-2 12 to 24 Design: review of 8 years autograft mean Latzman mg of relevant patients 148 N=101 Renal 1.48yrs (2010) with allo/autograft Allocation: BMP given patients Lumbar/lumbosacral Not reported insufficiency n=20 patients after license granted receiving defined Control: USA receiving 23 Purpose: assess 104 mean operations systemic complications operation 4.49yrs s
72 8 Results: Safety
Patient Study ID Intervention Control Surgery Design/Setting AE defined Follow-up Characteristics BMP: rhBMP-2 anterior: mean Maeda (2009) Design: prospective BMP: scoliosis 149 11.6mg/level 2.7yrs ICBG various Allocation: unclear control: spinal posterior: Not defined N=32 spinal deformity surgery Purpose: compare deformity 10mg/level Control: USA efficacy (sacrum/ilium) n=23 mean 4.9yrs Design: retrospective review of Medicare Mines (2011) Procedure 150 rhBMP-2 dose not patient data Pancreatic Unknown code for lumbar Approx 12 reported) Lumbar fusion Allocation: unknown cancer codes N=78194 fusion over months n=15460 Purpose: assess defined USA 67yrs incidence of pancreatic cancer BMP: Design: prospective DDD and mean Vaidya (2007b) rhBMP-2 lumbar: comparison related spinal 151 24.1mths 2mg/level Allograft Allocation: unclear disease, ALIF, TLIF, ACDF Not defined cervical: 1mg/level N = 41 Purpose: compare revision Control: USA n= 36 effectiveness and surgery, adult mean subsidence rates scoliosis 24mths No Design: retrospective rhBMP-2 (dose rhBMP review of all fusion cases Williams (2011) 115 not reported) N=43929 submitted by Scoliosis DDD and Not defined Not n=11933 n= Spinal fusion (excluding cervical) Research Society related spinal but examples reported n=11281 (excl 39397(ex Allocation: unknown disease, given USA cervical) cl Purpose: compare cervical) complication rates DDD – degenerative disc disease
73 8 Results: Safety
The tables above demonstrate that there are important differences between studies in terms of design, comparators used, patient selection criteria surgical approach and the way that measurements were made. The baseline risk of events also differs considerably between studies for some outcomes, as can be seen in the plots that follow. For this reason we do not believe it appropriate to combine the results of these studies in a meta-analysis to give an estimate of the average relative risk across the studies. We present the results of individual studies in single forest plots for each outcome without a summary diamond.
74 8 Results: Safety
General adverse events Heterotopic bone formation Nine studies reported on heterotopic bone formation. Two ALIF studies,130, 133 one ACDF/PCF study,144 five PLF/TLIF studies,118, 119, 124, 125, 127 and one mixed study.121 In four of these there were no cases of heterotopic bone formation in either the rhBMP-2 group or the comparator.127, 130, 133, 144 Two of the remaining five trials compared rhBMP-2 with ICBG, two compared with autologous bone and one with silicated calcium phosphate (SCP). One of the trials was randomised and reported blinded outcome assessment and complete follow-up but other details were not available from this conference abstract.124 The remaining four studies were non-randomised cohorts, none of which matched rhBMP-2 and control patients in their study design or adjusted for confounders in their analyses. Except for one study that did not provide data on baseline comparability,119 these reported included treatment groups to be similar in terms of basic demographic characteristics. However, mean length of follow-up was noticeably longer in rhBMP-2 patients in one of the studies (38.5 versus 24.4 months).125 Figure 28 illustrates the findings for each of these studies. The point estimates among these studies observed a risk of heterotopic bone formation in rhBMP-2 patients of between two and five times that observed in patients with ICBG or similar spinal surgeries. However, only one of these five relative risk estimates had a 95% confidence interval that excluded 1 (was statistically significant). Results are reasonably consistent but the prevalence of heterotopic bone formation among rhBMP-2 patients varied considerably across the studies (from 2% to 52% of patients), perhaps due to lack of a consistent definition of what constitutes heterotopic bone formation. Although these five studies indicate a trend toward increased risk of ‘heterotopic bone formation’ associated with rhBMP-2, the limitations described above must be taken into account.
Figure 28 Relative risk of heterotopic bone formation in five studies.
75 8 Results: Safety
*LIF: Lumbar Interbody Fusion Osteolysis Two PLF/TLIF studies reported on osteolysis. One of these demonstrated an increase in rhBMP-2 relative to autologous bone (relative risk 3.17, 95% CI 1.00 to 10.03), which was borderline statistically significant.118 The other study reported an increase in osteolysis among patients receiving rhBMP-2 relative to ICBG (relative risk 4.26, 95% CI 0.24 to 74.94). However, the rhBMP-2 patients in this latter study were followed for an average of 14 months longer than control participants.125 Figure 29 illustrates the findings for each of these studies. Both studies showed a trend toward increased osteolysis in rhBMP-2 patients but confidence intervals were wide and neither explicitly addressed potential confounding in their design or analysis.
Figure 29 Relative risk of osteolysis in two studies.
76 8 Results: Safety
Infection Five studies reported on infection rates. Two PLF/TLIF studies compared rhBMP-2 with ICBG,120, 125 one cervical study compared rhBMP-2 with no rhBMP-2,144 one ALIF study compared rhBMP-2 with no rhBMP-2129 and the Maverick RCT compared rhBMP-2 in ALIF with a total disc replacement system. Figure 30 illustrates the findings for each of these studies. The Maverick RCT was considered to be at low risk of bias and treatment groups appeared to be similar. An imbalance between groups could be seen for three of the four non-randomised studies (longer follow-up for rhBMP-2 patients in Rihn; more rhBMP-2 patients undergoing revision surgery in Lee; significantly fewer women receiving rhBMP-2 in Xu). The study by Lee et al. was restricted to patients aged over 65 years. All studies had wide confidence intervals and there is no clear indication that infection is any more or less likely with rhBMP-2 than with comparable types of spinal fusion surgery.
Figure 30 Relative risk of infection in four studies.
77 8 Results: Safety
Neurologic events Five studies collected data on neurologic events: one ALIF study,130 one cervical study,139 two PLF/TLIF studies116, 123 and the ALIF Maverick RCT. In two of these studies there were no patients with neurologic events in either the rhBMP-2 group or the comparator.116, 130 Two studies reported small numbers of neurologic events: one comparing rhBMP-2 with ICBG plus interbody cage in TLIF surgery123 and the other comparing rhBMP-2 with local bone and ICBG in PCF surgery.139 The Maverick RCT (comparing with a disc replacement system) reported most of the observed events. The Maverick RCT provides the most methodologically robust evidence for this outcome; neither of the relatively small observational studies accounted for confounding in their design or analysis and one failed to report information on length of follow-up.139 Figure 31 illustrates that the results of these studies are inconsistent, with substantial statistical uncertainty associated with both of the non-randomised studies. There is no convincing evidence from these studies that neurological events are any more or less likely with rhBMP-2 than with comparable types of spinal fusion surgery.
Figure 31 Relative risk of neurologic events in three studies.
78 8 Results: Safety
Hardware failure Ten studies collected data on hardware failure: two ALIF studies including the Maverick RCT,130 two cervical studies,139, 144 five PLF/TLIF studies,116-118, 125, 128 and one mixed approach study.151 In two of these studies there were no instances of hardware failure in either the rhBMP-2 group or the comparator.116, 130 One study provided only numbers of events for the rhBMP-2 group.117 One study (Vaidya el al.151) differed from the others by reporting the number of levels rather than number of patients with hardware failure; this study was excluded from the plot below. The six remaining studies varied in quality. The Maverick RCT had a low risk of bias and among the observational studies there were limitations relating to differential follow-up (Rihn) and incomplete reporting of length of follow-up (Hiremath). None of the observational studies formally accounted for potential confounders in their analyses. Figure 32 illustrates the findings for each of these six studies. Analysis of the IPD from the Maverick trial showed an increased risk of hardware failure in rhBMP-2 compared with the Maverick disc system (RR 2.52, 95% CI 1.25 to 5.11). However, results are inconsistent across studies and no general conclusions about the comparative prevalence of hardware failure in rhBMP-2 surgery can be drawn on the basis of this evidence alone.
Figure 32 Relative risk of hardware failure in six studies.
79 8 Results: Safety
Wound complications Wound complications (including wound infections) were reported by nine papers (one of which reported data separately for ACDF and other surgical approaches115) across all the prespecified surgical groupings. Three studies reported on cervical surgery,115, 138, 139 two studies on PLF/TLIF,120, 125 two studies on ALF/ALIF130, 133 and two studies reported on mixed populations.7, 115 Two studies used medical record databases to compare rhBMP-2 with other treatments.7, 115 Each study reported statistically significant but inconsistent results. Neither study matched or adjusted for confounding factors. One study reported a significant difference between groups in terms of age and diagnosis. One study identified rhBMP-2 and comparator patients from separate institutions with different spinal fusion protocols, matching patients on demographic factors such as age and sex.130 The remaining studies drew participants from the same institution, although (as described above) several showed an imbalance between treatment groups at baseline (Lee, Xu) or had different lengths of follow-up between groups (Rihn). Figure 33 illustrates the findings for each of these studies. There is considerable inconsistency in the results across the studies. There is no convincing evidence from these studies that wound complications are any more or less likely with rhBMP-2 than with comparable types of spinal fusion surgery.
* 95% CIs for Cahill and Williams are entirely within their respective squares Figure 33 Relative risk of wound complications in 10 studies.
80 8 Results: Safety
ALF/ALIF related events Vascular Events One ALIF study130 (which compared rhBMP-2 with bone marrow aspirate) reported on vascular events. No events were observed in either group. Retrograde ejaculation One ALIF study reported on retrograde ejaculation in a consecutive cohort of patients collected over 10 years at one surgical facility where 239 people received rhBMP-2 and 233 people formed a historical control group.129 There was a higher rate of retrograde ejaculation in the rhBMP-2 group compared with controls (RR 7.31, 95% CI 1.69 to 31.62). The findings of this study are presented in Figure 34. Data from this study suggest that there may be an increased rate of retrograde ejaculation with use of rhBMP-2. Although changes over time between patients who received control and rhBMP-2 could in theory introduce a bias, the two sets of patients appeared to be similar in terms of age, diagnosis and number of levels fused. Co-morbidities generally considered to be associated with increased risk of retrograde ejaculation were similarly distributed for each treatment group. Further, it is unlikely that bias due to changes in patient mix over time would be responsible for such a substantial difference between rates of retrograde ejaculation between the two groups.
Figure 34 Relative risk of retrograde ejaculation in one study.
81 8 Results: Safety
Urinary retention One ALIF study reported on urinary retention in a consecutive cohort of patients as described above.129 There was a higher rate of urinary retention following catheter removal in patients who received rhBMP-2 compared with controls (RR 2.14, 95% CI 1.04 to 4.43). As above, although non- randomised studies are generally considered to be at higher risk of bias and confounding, the intervention and control patients appear to be similar. Findings from this study are shown in Figure 35.
Figure 35 Relative risk of urinary retention in one study.
82 8 Results: Safety
Cervical Related Events Dysphagia Eight cervical studies collected data on dysphagia.115, 139-145 In Hiremath et al. there were no instances of dysphagia in either the rhBMP-2 or comparator groups.139 The study by Lu et al. reported only that there was more severe dysphagia in patients receiving rhBMP-2 without any numerical data (abstract only available).141 None of the remaining six studies compared rhBMP-2 directly with ICBG. All compared with a “no rhBMP-2” control group and were of ACDF/PCF surgery. The study by Vaidya reported data at both two and six weeks, with the six week data shown in the forest plot below.143 One study that controlled for potential confounders in its analyses reported the largest relative risk of dysphagia associated with rhBMP-2.142 Two studies provided limited description of matching treatment groups,140, 145 but only one showed these groups to be comparable at baseline.140 The remaining studies either showed differences in baseline characteristics between groups115, 144 or did not report sufficient information to allow judgements about comparability.143 Figure 36 illustrates the findings for each of these studies. Despite some inconsistency across the study findings, four of the six showed a trend towards increased risk of dysphagia among rhBMP-2 patients, of which three were statistically significant. Therefore, there is some evidence to suggest the possibility of an increased risk of dysphagia among rhBMP-2 patients undergoing cervical fusion procedures.
Figure 36 Relative risk of dysphagia in six studies.
Airway obstruction No studies reported on airway obstruction.
83 8 Results: Safety
Neck Pain One PCF study that compared rhBMP-2 to no rhBMP-2 reported on neck pain.144 There was a higher rate of neck pain among patients with rhBMP-2 (19/48 patients compared with 31/156) corresponding to a relative risk of 1.9 (95% CI 1.24 to 3.19). As noted previously, some of the observed difference could be due to confounding or biases inherent in the design of this non-randomised study. Findings from this study are shown in Figure 37.
Figure 37 Relative risk of neck pain in one study. Recurrent laryngeal neck palsy Two PCF studies reported on recurrent laryngeal neck palsy.139, 144 Hiremath compared rhBMP-2 with various comparators including local bone and ICBG but did not provide numbers of observed events in the comparator group.139 Xu144 (described above) observed a higher rate of recurrent laryngeal neck palsy among patients receiving rhBMP-2 versus no rhBMP-2 (3/48 compared with 7/156) corresponding to a relative risk of 1.39. However, the confidence interval is wide (95% CI 0.37 to 5.18), and includes the possibility of a reduced or no risk associated with rhBMP-2. The methodological limitations of this single study add further uncertainty to any conclusion about the relationship between rhBMP-2 and this outcome. The findings of this study are illustrated in Figure 38.
Figure 38 Relative risk of recurrent laryngeal neck palsy in one study.
84 8 Results: Safety
PLF/ TLIF Related Events Leg pain or radiculitis Five PLF/TLIF studies reported on leg pain and/or radiculitis.116, 118, 122, 125, 126 One of these studies did not provide data from both groups for this adverse event so did not allow a comparison.116 Among the other four studies, one compared rhBMP-2 with ICBG in TLIF surgery,125 one compared with autologous bone in PLF,118 one with autograft and allograft in TLIF surgery122 and one with instrumented fusion without rhBMP-2 in TLIF surgery.126 The study by Gray reported data at six weeks and six months; six month data are presented in the forest plot.118 None of these observational studies appeared to account for potential confounding and two reported insufficient information to allow judgements about comparability.118, 126 Figure 39 illustrates the findings for each of these studies. Point estimates of relative risk from these studies are generally consistent, suggesting that patients receiving rhBMP-2 are two to four times as likely to suffer leg pain and/or radiculitis as patients with other surgeries. However, none of the results from these small studies were statistically significant. The prevalence of radiculitis among rhBMP-2 patients ranged from 2% to 17% across the studies.
Figure 39 Relative risk of leg pain or radiculitis in four studies.
Leg weakness No studies collected or reported data on leg weakness.
85 8 Results: Safety
Inflammatory cyst formation Two PLF/TLIF studies reported on inflammatory cyst formation. One compared rhBMP-2 with ICBG, 125 and one large cohort study compared rhBMP-2 with no rhBMP-2 use.115 In the former study only a single observation of inflammatory cyst formation was reported, providing insufficient evidence to draw a conclusion about the relationship between rhBMP-2 and inflammatory cyst formation. The larger study was a retrospective analysis of medical records in which treatment groups differed in terms of age and diagnosis. There was no evidence of a difference in inflammatory cyst formation in this study. Figure 40 illustrates the findings of the two studies.
Figure 40 Relative risk of inflammatory cyst formation in two studies.
86 8 Results: Safety
Cancer A brief summary of the two identified non-randomised comparative studies reporting cancer in rhBMP- 2 is presented here for completeness. A retrospective review148 of 125 patients who had undergone lumbar fusion with rhBMP-2 or auto/allograft alone over an eight year period at a single institution identified patients from hospital records with the objective of assessing the frequency of complications and new diagnoses associated with rhBMP-2. A total of eight cancers were observed among the 101 patients (8%) who were treated without rhBMP-2 and four cancers were identified among the 24 patients (17%) who had received rhBMP-2. This equates to an increase in relative risk of cancer in rhBMP-2 patients of 2.10 which is not statistically significant (95% CI 0.69 to 6.42). Patients who had not received rhBMP-2 had a significantly longer period of follow-up than those who had received rhBMP-2 (4.49 versus 1.48 years, p<0.001) and the two groups of patients differed significantly from one another on several baseline characteristics. Given these limitations, the findings of this small study in relation to cancer risk are likely to be subject to a high risk of bias. A large retrospective review150 sponsored by and undertaken by employees of Wyeth which manufactured rhBMP-2 (subsequently acquired by Pfizer) used Medicare records to identify patients with claim codes for lumbar fusion with or without bone morphogenetic protein (assumed to be mostly rhBMP-2) and pancreatic cancer identified from claim codes for diagnosis and cancer specific therapy (at least two claims). For the period October 2003 to December 2005, 83 pancreatic cancers among 78,194 individuals (114,498 patient-years) were identified in the non-rhBMP-2 group and eight pancreatic cancers were identified among 15,460 individuals (16,018 patient-years) in the rhBMP-2 group. The authors reported no difference in the risk of pancreatic cancer between patients who did not receive BMP and those who did (adjusted hazard ratio 0.70, 95% CI 0.34 to 1.45). Of note was the difference in duration of follow-up between the two groups – 1.47 years (interquartile range 0.73 to 2.21) for the non BMP group and 0.91 (interquartile range 0.41 to 1.54) for the BMP group – together with the large number of patients who were excluded from the original 158,689 Medicare beneficiaries identified as potentially eligible for inclusion. In addition to the reasonable exclusion of those with a prior diagnosis of pancreatic cancer (74), exclusions were made on the grounds of having end stage renal disease or chronic disability (25,048), being members of Medicare HMOs (39,927) and being younger than 67 years (39,927). No details of the numbers of exclusions per group were provided. It is not clear whether these exclusions were specified in advance and the rationale for these particular choices is unclear to us. If the exclusion criteria were decided post hoc, it is possible that they might have been chosen to produce desired results. As no sensitivity analyses are presented of all lumbar fusion patients and no rationale is provided for the choice of exclusion criteria, we consider the study to be at a high risk of bias. These two studies provide little additional useful evidence on the relationship between rhBMP-2 and cancer due to their inconsistent findings and methodological limitations. Summary of data on adverse events from the wider literature When interpreting data from these studies some caution is required because (with the exception of the three RCTs that were included here but ineligible for our earlier consideration of efficacy) the patients were not randomised. Thus, apparent differences between rhBMP-2 and other types of spinal surgery may actually be due to differences between the patients receiving the different types of surgery (selection bias). Assessment of adverse events in rhBMP-2 surgery is made more uncertain in these analyses by the range of different types of spinal surgery used (ALIF, PLF, cervical procedures and so on) and because not all the comparison surgeries used ICBG. The proportions of patients experiencing some adverse events varied considerably across studies, perhaps due to the lack of objective definitions for some adverse events. For all these reasons we did not formally combine data across studies in meta-analyses and instead presented the results separately for each study. Despite these considerable caveats, there is some suggestive evidence of higher rates of certain adverse events among rhBMP-2 patients from the wider literature. Heterotopic bone formation has been more commonly observed among rhBMP-2 patients; whether this resulted in clinical consequences for those patients was unclear.
87 8 Results: Safety
Leg pain and radiculitis have been more commonly observed among rhBMP-2 patients. This corresponds with our findings that the numbers of patients experiencing pain in general was more common after surgery among rhBMP-2 patients in the Medtronic trials (see Section 8.1). Osteolysis was more commonly observed among rhBMP-2 patients but only two studies reported on this event. Dysphagia has been commonly observed for rhBMP-2 patients in cervical spinal surgery but there were some inconsistencies in the results of the studies reporting on this event. Data on retrograde ejaculation and urinary retention were available from one non-randomised study in which the absolute number of observed events was small but demonstrated a statistically significant increase among the rhBMP-2 patients.
88 8 Results: Safety
8.6. Safety results from the MedWatch data A total of 913 MedWatch forms were received in the original data submission from Medtronic which we consider below. We did not consider a second batch of 87 forms was sent by Medtronic in April 2012 and a third batch of 116 forms in October 2012 (after initial submission of our report to YODA). The forms represent voluntary reports from healthcare professionals, patients and other parties on possible adverse events associated with a medical product. As such, these data are solely intended to detect safety hazard signals rather than draw inferences about the incidence or prevalence of adverse events associated with a treatment. We asked Medtronic for any existing collated information or reports using these MedWatch forms and were informed that none were available. We have provided only a descriptive summary of these data. The forms were received electronically as PDF documents. Given the large number of reports, one researcher extracted basic descriptive data (summarised below). The “complaint dates” recorded on PCR forms ran from August 2003 to April 2011 and all related to the INFUSE® preparation of rhBMP-2. Most complaints (46%) were from the published research literature, followed by those reported by Medtronic employees (24%) and health care professionals (16%). Given the relatively high proportion of complaints taken from the literature there is likely to be some overlap in patients reporting adverse events here and in the publications reviewed in section 8.4. Appendix XVII shows the adverse events reported in the MedWatch forms according to type of surgical approach used. More than 80% of all reported adverse events could be grouped into the categories that we specified in our protocol. The most commonly reported events across all types of rhBMP-2 use (including non-spinal indications) were inflammatory cyst formation (16.3%) and heterotopic bone formation (10.4%). There were no reports of retrograde ejaculation among these data. Three hundred and forty reports (37%) were not classified by type of surgery. Most of these related to non-spinal uses of rhBMP-2, including long bone fusion and maxillofacial reconstruction. We have restricted the following analyses to the use of rhBMP-2 in the spine. Type of surgical procedure Only 45 of the 913 MedWatch reports clearly related to FDA approved use of INFUSE® in ALIF surgery (see Table 14). Of these, the most commonly reported adverse event was non-union (16 reports, 36%), followed by inflammatory cyst formation (six reports, 13%). Most reports relating to the use of rhBMP-2 in the spine were in “off-label” indications such as ACDF, PLF and PLIF/TLIF. Of the 171 reports of its use in the cervical spine, the most commonly reported adverse events were dysphagia (40 reports, 23%), inflammatory cyst formation (18 reports, 11%) and “unanticipated inflammation and swelling” (26 reports, 15%). Among “off label” lumbar procedures (PLIF, TLIF, PLF; 357 reports), the most commonly reported events were inflammatory cyst formation (97 reports, 27%), leg pain/radiculitis (38 reports, 11%), and osteolysis (38 reports, 11%). The distribution of events by spinal surgery is shown in Figure 41.
89 8 Results: Safety
Figure 41 Distribution of pre-specified adverse events reported in MedWatch forms in spinal use (08/2003 to 04/2011).
Cause of event Each PCR form contained an Investigation Summary, usually a brief narrative with a judgement about whether the reported event might be related to the device (in this case, INFUSE®). For all reported events, we extracted these judgements into three main categories based on the wording of the investigation summaries: “Device may have caused or contributed”, “Not believed to be related to the product” and “Unable to determine”. We did not attempt to reclassify or independently appraise the accuracy of these judgements in any way. Table 14 summarises these judgements for all reported prespecified events in known spinal applications. Five hundred and seventy-three adverse event reports were clearly related to use of rhBMP-2 in the spine and 472 of these reported one of the events specified in our research protocol. The cause of event was unable to be determined in most (57%) reports; 24% of reports were not believed to be related to the use of INFUSE® and in 19% of reports INFUSE® was considered to have caused or contributed towards the event.
90 8 Results: Safety
The events most commonly judged as being related to the use of rhBMP-2 were heterotopic bone formation (43% of reported events), osteolysis (42% of reported events), and unanticipated inflammation and swelling (31% of reported events). Other adverse events (leg weakness, neck pain) were reported infrequently.
Table 14 Cause of event as described in PCR Investigation Summaries (spinal surgery only). Total Device may Not believed Unable to Not reported reported have caused to be related determine / other or to the contributed product % n/472 % of n % n % n % n % report 472 s report ed events "Unanticipated 32 10 31.3 4 12.5 18 56.3 inflammation and swelling" 6.8 Airway obstruction 18 3.8 1 5.6 2 11.1 15 83.3 Cancer 0 0 0 0 0 0 0 0
Dysphagia 43 9.1 9 20.9 5 11.6 29 67.4 Hardware failure 7 1.5 1 14.3 2 28.6 2 28.6 2 0.4 Heterotopic bone 68 29 42.6 1 1.5 37 54.4 1 0.2 formation 14.4 Inflammatory cyst 123 8 6.5 32 26.0 83 67.5 formation 26.0 Leg pain/radiculitis 39 8.3 2 5.1 23 59.0 14 35.9 Leg weakness 2 0.4 0 0 2 100 0 0 Neck pain 2 0.4 2 100 0 0 0 0 Neurologic events 20 4.2 2 10.0 7 35 11 55.0 Non-union 34 7.2 2 5.8 14 41.2 18 52.9 NR 3 0.6 0 0 1 33.3 2 66.7 Osteolysis 43 9.1 18 41.9 4 9.3 21 48.8 Post-operative pain 5 1.1 0 0 3 60 2 40.0 Recurrent laryngeal 3 0.6 0 0 1 33.3 2 0.67 nerve palsy Retrograde 0 0 0 0 0 0 0 0 ejaculation Vascular events 1 0.2 0 0 0 1 100 Wound complications 29 6.1 5 17.2 11 38.0 13 44.8 TOTAL 472 100 89 18.9 112 23.7 268 56.8 3 0.6% % % % % Summary The most commonly reported events relating to rhBMP-2 in the spine relate to off label use; reports relating to rhBMP-2 use in ALIF were relatively rare. This observation is likely to reflect the wide use of rhBMP-2 outside its FDA approved use in ALIF procedures over the eight-year reporting period covered by these data.8 Ong et al.8 noted that most of the 340,000 uses of BMP between 2003 and 2007 were in the spine, which suggests that MedWatch data pertain to a very small proportion of procedures, although it is not clear how comprehensively MedWatch reports reflect the actual number of adverse events that occur in practice.
91 8 Results: Safety
The most commonly reported events across all spinal uses were inflammatory cyst formation and heterotopic bone formation. This finding was driven by the large number of these events reported in relation to off-label use. All reported incidences of dysphagia were in cervical use of rhBMP-2. There were no reports of retrograde ejaculation and only one non-spinal report specifically relating to cancer. In most of the reports, the cause of the reported event could not be determined. The events most likely to be attributed to rhBMP-2 use were heterotopic bone formation, osteolysis and unexpected inflammation/swelling. Several limitations prevent us from being able to draw strong conclusions from the MedWatch data. Primarily, MedWatch forms provide information on how frequently each event was reported and may not be a fair reflection of the adverse events that have occurred in practice. Not all events that occurred in practice will necessarily have been reported and we cannot determine the motivations behind deciding whether to report or not. We could not exclude the possibility of some direct overlap between these data and the adverse events reported in the wider literature. Due to resource restrictions, only one researcher extracted these data. Therefore potential for some errors or subjective decisions cannot be ruled out.
92 9 Results: Reliability of the Evidence
9. Results: Reliability of the Evidence In this section, we aimed to determine the extent to which systematic review and meta-analysis of the efficacy and safety of rhBMP-2 based on only the published Medtronic literature would reach the same conclusions as analyses based on the IPD received from Medtronic. Also, as Medtronic made available the trial reports produced for regulatory purposes, we examine the extent to which analyses based on the information presented in their confidential reports would accord with the two previous analyses. We therefore compare the quantitative findings of analyses based on three different sources of Medtronic trial data: IPD, CSRs and publicly available published papers and abstracts. It was not our intention to review or systematically critique the CSRs in detail (this would not have been feasible within the strict timescale of the project). However, where there were important differences between the published reports and the IPD supplied we examined the CSRs to see if they contained information that could shed light on the differences. We also sought clarification on important issues from Medtronic mediated by the YODA team. We did not have access to a highly detailed protocol or CSR for the Glassman trial, which precluded it from being included in the following comparisons.
9.1. Comparison of reported outcomes For each Medtronic trial, we recorded which clinical effectiveness outcomes were described as being collected in the study protocol and/or CSR. We then compared these collected outcomes with publicly available abstracts and publications of study findings. Intra-operative outcomes such as operation time and blood loss were measured but not mentioned in the protocols so we have not included these outcomes in the comparison. Results We received protocols and CSRs for all 17 of Medtronic’s relevant rhBMP-2 studies. Three of these suspended recruitment earlier than planned (INFUSE®/Bone Dowel Pivotal RCT, INFUSE®/INTER FIX™ PLIF RCT, and INFUSE®/CORNERSTONE® ACDF Pivotal RCT). Our independent searches identified a total of 32 journal publications and conference abstracts, reporting outcomes from 11 of the 17 existing Medtronic studies (10 RCTs and one single-arm study; see Table 15). Some publications reported on all patients from a single trial, some on a subgroup of trial patients and others reported on combined data across two or more trials (see Table 16). Our searches suggested that five of the remaining six Medtronic trials have never been published or presented (INFUSE®/INTER FIX™ ALIF pilot, INFUSE®/CORNERSTONE® pivotal, INFUSE®/TELAMON PEEK instrumented PLIF pilot, rhBMP-2/CRM two-level pilot, rhBMP-2/BCP Mexico pilot) and limited data from one (rhBMP-2/BCP Canada pivotal RCT) is available only from three short conference abstracts.92-94
Table 15 Publication status of the 17 Medtronic trials. Published in scientific journal Unpublished / abstract only RCTs INFUSE®/LT-CAGE® pilot INFUSE®/INTER FIX™ ALIF pilot INFUSE®/LT-CAGE® open pivotal rhBMP-2/BCP Canada pivotal INFUSE®/bone dowel pilot INFUSE®/CORNERSTONE® ACDF Pivotal INFUSE®/bone dowel pivotal INFUSE®/INTER FIX™ PLIF INFUSE®/MASTERGRAFT® pilot INFUSE®/CORNERSTONE® ACDF pilot rhBMP-2/BCP US pilot AMPLIFY™ pivotal MAVERICK™ Disc pivotal* Single- INFUSE®/LT-CAGE® laparoscopic INFUSE®/TELAMON PEEK instrumented PLIF arm pivotal pilot studies rhBMP-2/CRM two-level pilot rhBMP-2/BCP Mexico pilot *Maverick trial compares rhBMP-2 with the Maverick disc replacement system
93 9 Results: Reliability of the Evidence
Table 16 Outcomes reported in peer-reviewed journal articles compared with Medtronic clinical study reports (CSRs).
Study Design Single arm RCT LT Cage Interfix Mastergr Cornerst Study Name LT Cage Laparascopic Pilot LT Cage Open Bone Dowel Pilot Bone Dowel Pivotal PLIF aft Pilot one Pilot Maverick BCP US Amplify First author Kleeman Burkus Burkus Boden Burkus Burkus Burkus Burkus Burkus Burkus Burkus Burkus Burkus Haid Dawson Baskin Gornet Burkus Boden Glassman Dimar Glassman Dimar Burkus Year 2001 2003a 2009 2000 2002a 2003a 2003b 2009 2011 2002b 2004 2005 2006 2004 2009 2003 2011 2011 2002 2005 2006 2007 2009 2011 Partial Partial Partial Combined Combined Combined Partial Partial Partial Partial Complete study results, partial study results, or Complete Complete Complete Complete Complete Complete Complete Complete Complete Complete Complete Complete Complete Complete study study study study study study study study study study Publication study study study study study study study study study study study study study study combined results from multiple studies reported? results results results results results results results results results results Fusion success/failure Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes No Oswestry Disability Index (low back pain) Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes No Neck Disability Index Yes SF-36 PCS Yes Yes Yes Yes No Yes No Yes No Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes No SF-36 MCS Yes No Yes Yes No No No Yes No Yes No No No No No Yes Yes No No No No Yes No No Back pain Yes No Yes Yes No No Yes No Yes No Yes Yes Yes Yes Yes No Yes No Yes Yes Yes No Neck pain Yes Leg pain Yes No Yes Yes No No Yes No Yes No Yes Yes Yes Yes Yes No Yes No Yes Yes Yes No Arm pain Yes Neurological status No No No Yes Yes No No No No Yes No No No Yes No Yes Yes No No No No No No No Donor site (hip) pain No No No Yes No No No No Yes No Yes No Yes No Yes No Yes No Yes No Return to work/activity Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes Yes Yes Yes No Yes No No No Yes Yes Yes No Disc height measurement No No No No No No No No No Yes No Patient satisfaction Yes No No Yes No No No No Yes No No No Yes No Yes Yes No Yes No No No No No Bovine Collagen Antibody response No No No No Yes No No No Yes No No Yes No Yes No Yes No Yes Yes No No No No Yes Reported outcomes Overall success (composite measure) No No No No No No No No No No No No No Yes No Yes No No No No No No No Additional surgery Yes Yes Yes Yes Yes Yes No Yes No Yes No Yes No Yes Yes Yes Yes No Yes Yes No No Yes No Patient Global Perceived Effect No No No Yes No No No No No No No No No No No No Yes No No No No No No No Doctor's perception of results No No No No No No No No No No No No No No No No No Yes No No No No No Adjacent level motion No No No No Adverse Events mentioned at all? Yes No Yes Yes Yes No No Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Total reported 10 5 9 9 11 5 1 9 3 12 4 10 7 12 9 11 14 2 11 2 8 7 9 2 Total not reported 7 12 8 2 5 11 15 7 13 6 14 7 10 5 7 4 3 15 5 14 8 9 7 14 % of outcomes collected that were reported 58.8 29.4 52.9 81.8 68.8 31.3 6.3 56.3 18.8 66.7 22.2 58.8 41.2 70.6 56.3 73.3 82.4 11.8 68.8 12.5 50 43.8 56.3 12.5 Yes = reported in both the publication and corresponding CSR. No = reported in CSR but not in the publication. Blank cells = outcome did not appear to be measured Three Medtronic RCTs (INFUSE®/INTER FIX™ ALIF pilot, INFUSE®/CORNERSTONE® ACDF pivotal, rhBMP-2/BCP Canada pivotal) and three non-randomised studies (INFUSE®/TELAMON PEEK instrumented PLIF pilot, rhBMP-2/ CRM Two-level pilot, rhBMP-2/BCP Mexico pilot) have not been published to date
94 9 Results: Reliability of the Evidence
CSRs consistently proved to be the most comprehensive source of summary outcome data for all studies. Table 17 shows the proportion of collected outcomes available from individual publications for each Medtronic study. As might be expected, this table indicates that no individual abstract or journal article reported all of the clinical outcomes known to have been collected in a study. For all trials, as would be expected, journal articles provided more complete outcome data than conference abstracts except for the rhBMP-2/BCP Canada pivotal RCT, for which only conference abstracts were found.
Table 17 Proportion of collected outcomes presented in publicly accessible sources.
Proportion of collected Proportion of collected Number of publicly outcomes that were outcomes that were
available sources presented in individual available across all sources sources
All available All available reports reports Journal Journal Study Conference Journal (Journal (Journal Trial Name articles articles design abstracts articles articles and articles and alone alone conference conference abstracts) abstracts) INFUSE®/LT-CAGE® pilot RCT 0 1 82% 82% 82% 82% INFUSE®/LT-CAGE® open RCT 2 5 6 to 69% 6 to 81% 81% pivotal 69% INFUSE®/bone dowel pilot RCT 1 2 24 to 67% 24 to 67% 67% 67% INFUSE®/bone dowel RCT 1 2 29 to 59% 41 to 59% 59% pivotal 59% INFUSE®/INTER FIX™ RCT 1 1 6 to 71% 71% 71% 71% PLIF INFUSE®/CORNERSTONE RCT 0 1 73% 73% 73% 73% ® ACDF pilot INFUSE®/MASTERGRAFT RCT 1 1 40 to 56% 56% 63% 56% ® pilot INFUSE®/INTER FIX™ RCT 0 0 0% 0% 0% 0% ALIF pilot MAVERICK™ Disc pivotal RCT 1 2 11 to 82% 11 to 88% 88% 82% rhBMP-2/BCP US pilot RCT 0 1 69% 69% 69% 69% rhBMP-2/BCP Canada RCT 3 0 6 to 19% 0% 21% 0% pivotal AMPLIFY™ pivotal RCT 1 5 12 to 56% 12 to 69% 69% 56% INFUSE®/CORNERSTONE RCT 0 0 0% 0% 0% 0% ® ACDF Pivotal
INFUSE®/LT-CAGE® Single- 1 3 31 to 59% 31 to 65% 59% arm laparoscopic pivotal 59% INFUSE®/TELAMON PEEK Single- 0 0 0% 0% 0% 0% arm instrumented PLIF pilot rhBMP-2/CRM two-level Single- 0 0 0% 0% 0% 0% arm pilot rhBMP-2/BCP Mexico pilot Single- 0 0 0% 0% 0% 0% arm
95 9 Results: Reliability of the Evidence
Reporting of efficacy outcomes Table 16 and Table 18 summarise the reporting of specific clinical outcomes across peer-reviewed journal publications of Medtronic trials. These show that broadly the same set of efficacy outcomes were collected across the Medtronic trials (with the exception of neck and arm pain, which were restricted to trials of cervical surgery), although some outcomes were more consistently reported in the literature than others. Fusion and pain/disability status were the most common primary outcomes according to the Medtronic trial protocols (see Appendix XIII, Table 28). Around 70% of pain and function measures (ODI; SF-36 PCS, back, hip and leg pain scores) that were collected in Medtronic trials have been presented somewhere in the peer-reviewed literature. The mental component score of the SF-36 was less frequently presented (SF-36 MCS, 44%). Radiographic evidence of fusion was collected in 16 of the 17 Medtronic trials (the control arm of Maverick did not use fusion surgery) and reported in the literature for 11 of these trials. The remaining five trials are unpublished. Table 18 shows that ODI, SF-36 PCS, fusion and additional surgery outcomes were available in some form for all 11 trials that have been published; any missing outcomes were attributable to the six unpublished trials. Other frequently collected (but arguably less important) outcomes are rarely presented in the literature, even when the study itself has been published (e.g. overall success 14%, patient perceived effect 15%, doctor’s perception of results 0%). We did not analyse these measures in our review of effectiveness as we did not identify them a priori as being of clinical importance.
Table 18 Proportion of collected outcomes reported in the peer-reviewed literature. Outcome Number of Number of trials % of trials for Medtronic trials for which outcome which outcome measuring has been has been outcome presented in the presented in peer-reviewed the peer- literature reviewed literature Fusion success/failure 16 11 68.8 ODI (low back pain) 15 11 73.3 NDI (neck pain) 2 1 50 SF-36 PCS 16 11 68.8 SF-36 MCS 16 7 43.8 Back pain 13 9 69.2 Neck pain 2 1 50 Leg pain 13 9 69.2 Arm pain 1 1 100 Neurological status 15 6 40 Donor site pain 10 7 70 Return to work/activity 13 9 69.2 Disc height measurement 10 1 10 Patient satisfaction 14 7 50 Bovine Collagen Antibody 14 7 50 response Overall success (composite 14 2 14.3 measure) Additional surgery 14 11 78.6 Patient Global Perceived Effect 13 2 15.4 Doctor's perception of results 13 0 0 Adjacent level motion 3 1 33.3
96 9 Results: Reliability of the Evidence
Reporting of adverse events The CSRs provided aggregated data on adverse events recorded during each Medtronic study (see
Study and date of first patient
enrolment
RCT
CAGE® Pilot Pilot CAGE® CAGE® CAGE® - - - LT RCT Jan 1997 Bone dowel pilot RCT Apr 1998 LT Open Aug 1998 LT Laparoscopic Nov 1998 Interfix PLIF RCT March 1999 BCP US RCT May 1999 Cornerstone Pilot RCT Sept 1999 Bone dowel RCTPivotal Sept 2000 RCT Amplify March 2002 Mastergraft pilot RCT April 2003 Maverick RCT April 2003 Adverse event as reported Wound dehiscence Yes Back pain Yes Yes Trauma Yes No No No No No No No Yes No Yes Back strain No Endcap event No Urinary retention Yes Fracture No Pseudoarthrosis No Pub‡ Shortness of breath No Numbness/SLA No Drug use No Facet joint pain No Rectal bleeding/haemorrhoids No Anatomical/technical difficulty No Yes Yes No No Yes Yes Back and or leg pain No No No No Yes□ No No Yes No Yes Cancer No No No No No Yes No Yes Cardiovascular No No No No No No No Yes No Yes Death No No No No No No Yes Yes Dural injury No No No Yes Yes□ Yes Yes Gastrointestinal No No No No No No No Yes No Yes * Graft site related No No No Yes No No Yes No Implant displacement/loosening No Yes Yes No No Yes No Yes† Infection No Yes No No No No No Yes No Yes Malpositioned implant No Yes No No Yes Yes Yes † Neurological No No No Yes No No No Yes No Yes Non-union No Yes No No No No Yes No Yes Other No No No No No No No Yes No Yes Other pain No No No No No No No Yes No Yes Respiratory No No No No No No No Yes No Yes Retrograde ejaculation No Yes No No Yes Spinal event No No No No No No No Yes No Yes Subsidence No Yes Yes No No Yes Urogenital No No No No No No No Yes Yes Vascular intra-op No Yes No Pub° No No Yes Vertebral fracture No No No Yes Yes Neck & arm pain No No Yes Dysphagia/Dysphonia No No Headache No Accidental injury/muscle strain No Allergic reaction No Yes Cervical spinal event Yes No Yes Electrolyte imbalance No Endocrine No Incision related No Yes Lower back pain not of back etiology No Spinal event at other lumbar level(s) Yes No Yes Depression Yes Elevated temperature without infection Yes Arthritis/bursitis Yes Carpal tunnel Yes Leg pain Pub□ Yes Lower extremity pain Yes Pain management control Yes Peritoneal tear Yes Rash Yes Spinal event at target level Yes Spinal event at thoracic level Yes No Yes Upper extremity pain No” Yes Vascular event post-op Yes Spinal event at other cervical level Yes Wound infection Pub‡ Total reported in CSRs 13 23 23 22 15 24 15 18 26 27 38 Appendix XIV, Table 31). shows the adverse event categories as reported in the summary tables of CSRs of the 11 published Medtronic trials. After the INFUSE®/LT-CAGE pilot RCT, it appears that a common collection of adverse event categories was specified across all subsequent trials (anatomical/technical difficulty,
97 9 Results: Reliability of the Evidence back and or leg pain, cancer, cardiovascular, death, dural injury, gastrointestinal, graft site related, implant displacement/loosening, infection, malpositioned implant, neurological, non-union, other, other pain, respiratory, retrograde ejaculation, spinal event, subsidence, urogenital vascular intra-op, vertebral fracture). Later trials (AMPLIFY™ pivotal RCT, INFUSE®/MASTERGRAFT® pilot RCT, MAVERICK™ disc pivotal RCT) reported additional adverse event categories. A full evaluation and reclassification of all individual adverse events using the patient narratives provided for almost all patients in the CSR report forms from the Medtronic trials lies outside the scope and resources of this project. However, Appendix XVIII summarises the data structure of individual adverse event forms. Since the categories on these forms do not correlate directly with those in the CSR adverse event summary tables, it appears that these data were reclassified for these summaries.
98 9 Results: Reliability of the Evidence Table 19 Reporting of adverse events in clinical study reports (CSRs) and journal articles for published Medtronic trials.
Study and date of first patient
enrolment
RCT
CAGE® Pilot Pilot CAGE® CAGE® CAGE® - - - LT RCT Jan 1997 Bone dowel pilot RCT Apr 1998 LT Open Aug 1998 LT Laparoscopic Nov 1998 Interfix PLIF RCT March 1999 BCP US RCT May 1999 Cornerstone Pilot RCT Sept 1999 Bone dowel RCTPivotal Sept 2000 RCT Amplify March 2002 Mastergraft pilot RCT April 2003 Maverick RCT April 2003 Adverse event as reported Wound dehiscence Yes Back pain Yes Yes Trauma Yes No No No No No No No Yes No Yes Back strain No Endcap event No Urinary retention Yes Fracture No Pseudoarthrosis No Pub‡ Shortness of breath No Numbness/SLA No Drug use No Facet joint pain No Rectal bleeding/haemorrhoids No Anatomical/technical difficulty No Yes Yes No No Yes Yes Back and or leg pain No No No No Yes□ No No Yes No Yes Cancer No No No No No Yes No Yes Cardiovascular No No No No No No No Yes No Yes Death No No No No No No Yes Yes Dural injury No No No Yes Yes□ Yes Yes Gastrointestinal No No No No No No No Yes No Yes * Graft site related No No No Yes No No Yes No Implant displacement/loosening No Yes Yes No No Yes No Yes† Infection No Yes No No No No No Yes No Yes Malpositioned implant No Yes No No Yes Yes Yes † Neurological No No No Yes No No No Yes No Yes Non-union No Yes No No No No Yes No Yes Other No No No No No No No Yes No Yes Other pain No No No No No No No Yes No Yes Respiratory No No No No No No No Yes No Yes Retrograde ejaculation No Yes No No Yes Spinal event No No No No No No No Yes No Yes Subsidence No Yes Yes No No Yes Urogenital No No No No No No No Yes Yes Vascular intra-op No Yes No Pub° No No Yes Vertebral fracture No No No Yes Yes Neck & arm pain No No Yes Dysphagia/Dysphonia No No Headache No Accidental injury/muscle strain No Allergic reaction No Yes Cervical spinal event Yes No Yes Electrolyte imbalance No Endocrine No Incision related No Yes Lower back pain not of back etiology No Spinal event at other lumbar level(s) Yes No Yes Depression Yes Elevated temperature without infection Yes Arthritis/bursitis Yes Carpal tunnel Yes Leg pain Pub□ Yes Lower extremity pain Yes Pain management control Yes Peritoneal tear Yes Rash Yes Spinal event at target level Yes Spinal event at thoracic level Yes No Yes Upper extremity pain No” Yes Vascular event post-op Yes Spinal event at other cervical level Yes Wound infection Pub‡ Total reported in CSRs 13 23 23 22 15 24 15 18 26 27 38
Yes = reported in CSR and at least one full academic journal publication. No = only reported in clinical study report (CSR) table of adverse events. Pub = only reported in a publication. *Split ilius and other; †grouped together; ^ not of back etiology; “ not of neck etiology ‡Dawson et al72 report pseudarthrosis in two patients, and three wound infections (two surgical site, one graft donor site) °Haid et al63 report one deep vein thrombosis as a vascular complication □Boden et al 87 report one transient leg pain after surgery, one epidural haematoma, and one persistent low back/leg pain among rhBMP-2 patients LT-CAGE® pilot reports fracture, other studies report vertebral fracture Three Medtronic RCTs (Interfix ALIF Pilot, Cornerstone Pivotal, BCP Canada) and three non-randomised studies (Telamon , CRM Two-level, BCP Mexico) have not been published to date
99 9 Results: Reliability of the Evidence
Study and date of first patient
enrolment
RCT
CAGE® Pilot Pilot CAGE® CAGE® CAGE® - - - LT RCT Jan 1997 Bone dowel pilot RCT Apr 1998 LT Open Aug 1998 LT Laparoscopic Nov 1998 Interfix PLIF RCT March 1999 BCP US RCT May 1999 Cornerstone Pilot RCT Sept 1999 Bone dowel RCTPivotal Sept 2000 RCT Amplify March 2002 Mastergraft pilot RCT April 2003 Maverick RCT April 2003 Adverse event as reported Wound dehiscence Yes Back pain Yes Yes Trauma Yes No No No No No No No Yes No Yes Back strain No Endcap event No Urinary retention Yes Fracture No Pseudoarthrosis No Pub‡ Shortness of breath No Numbness/SLA No Drug use No Facet joint pain No Rectal bleeding/haemorrhoids No Anatomical/technical difficulty No Yes Yes No No Yes Yes Back and or leg pain No No No No Yes□ No No Yes No Yes Cancer No No No No No Yes No Yes Cardiovascular No No No No No No No Yes No Yes Death No No No No No No Yes Yes Dural injury No No No Yes Yes□ Yes Yes Gastrointestinal No No No No No No No Yes No Yes * Graft site related No No No Yes No No Yes No Implant displacement/loosening No Yes Yes No No Yes No Yes† Infection No Yes No No No No No Yes No Yes Malpositioned implant No Yes No No Yes Yes Yes † Neurological No No No Yes No No No Yes No Yes Non-union No Yes No No No No Yes No Yes Other No No No No No No No Yes No Yes Other pain No No No No No No No Yes No Yes Respiratory No No No No No No No Yes No Yes Retrograde ejaculation No Yes No No Yes Spinal event No No No No No No No Yes No Yes Subsidence No Yes Yes No No Yes Urogenital No No No No No No No Yes Yes Vascular intra-op No Yes No Pub° No No Yes Vertebral fracture No No No Yes Yes Neck & arm pain No No Yes Dysphagia/Dysphonia No No Headache No Accidental injury/muscle strain No Allergic reaction No Yes Cervical spinal event Yes No Yes Electrolyte imbalance No Endocrine No Incision related No Yes Lower back pain not of back etiology No Spinal event at other lumbar level(s) Yes No Yes Depression Yes Elevated temperature without infection Yes Arthritis/bursitis Yes Carpal tunnel Yes Leg pain Pub□ Yes Lower extremity pain Yes Pain management control Yes Peritoneal tear Yes Rash Yes Spinal event at target level Yes Spinal event at thoracic level Yes No Yes Upper extremity pain No” Yes Vascular event post-op Yes Spinal event at other cervical level Yes Wound infection Pub‡ Total reported in CSRs 13 23 23 22 15 24 15 18 26 27 38
clearly shows that only a minority of adverse events that were recorded and reported in the CSRs (and subsequently made available in the IPD) were reported in the trial publications. Comparing the CSR categories against the adverse events described in published journal articles suggests that adverse event reporting across the Medtronic publications is relatively sparse and inconsistent. Only two publications (one for AMPLIFY™ pivotal RCT102 and one for MAVERICK™
100 9 Results: Reliability of the Evidence
Disc pivotal RCT80) reported a comprehensive table of all adverse events that occurred during the study, similar to that provided in the CSR. The INFUSE®/bone dowel pilot RCT, INFUSE®/CORNERSTONE® ACDF pilot RCT and INFUSE®/bone dowel pivotal RCT reported none. The remaining publications limited their reporting to a small proportion of the adverse events categories noted in the CSRs and there is no apparent consistency in which events were reported. Retrograde ejaculation events appeared in both a publication10 and CSR45 relating to the INFUSE®/LT-CAGE Open Pivotal RCT; both reported a total of six events. However, the journal publication did not report that five of these six events occurred in the rhBMP-2 group. The publication reported this important adverse event as a proportion of all participants (6/146; 4.1%) rather than as a proportion of male participants (6/78; 7.7%). Appendix XIX briefly summarises how adverse events were reported in the available published journal articles. This further emphasises that the extent of adverse event reporting was not consistent across publications. Several publications explicitly restricted reporting to “relevant”, “serious”, “related” or “unanticipated” adverse events; how these categories were defined was rarely apparent from the publications alone.42, 53, 87 Several adverse events specified as being of interest in our protocol (e.g. heterotopic bone formation, osteolysis, radiculitis) were not included as adverse event categories in the main IPD datasets or in any of the CSRs supplied by Medtronic. Yet some journal publications described occurrences that might independently be interpreted as osteolysis54, 55 or heterotopic bone formation.63, 67 For example, Haid et al.63 (INFUSE®/INTER FIX® PLIF RCT) reported a statistically significantly higher rate of “new bone formation extending outside the disc space and into the spinal canal or neuroforamina” in rhBMP-2 than in controls, which we interpreted as being heterotopic bone formation. We therefore examined the CSR for this trial in more detail. The study progress section of the CSR states that "No unanticipated adverse device events have been reported… The reports of posterior bone formation are not considered unanticipated adverse device events since this was a possible adverse event listed in the risk analysis and informed consent form" (page 4). The risk analysis section of the same CSR states: "The issue of posterior bone growth has previously been brought to the attention of the FDA. Please refer to the Interim Report in IDE G020053 dated March 13 2002 for further information. Briefly, in December 1999, Medtronic Sofamor Danek suspended enrolment in this clinical trial. This stoppage occurred mainly because of the results of two patients who developed posterior bone over the cages. The exact cause of the bone formation in these cases is still unknown. Possible contributing factors have been speculated to be the less than ideal placement of the cages, the lack of countersinking performed, the use of ADCON-L material during the surgery of one patient, the use of hemostatic agents such as gel foam sponges, and patient selection. In April 2001, a group of spinal surgeons, including several investigators, evaluated the radiographs of the investigational patients in the clinical trial. They considered 6 of these patients to potentially have a clinically significant amount of posterior bone formation. The reviewers judged 17 patients to have no or slight bone formation and 9 patients to have posterior bone formation but not to a possibly clinically significant amount. A few months later, a group of spinal surgeons reviewed the investigational patients’ radiographs again. At this time they noted similar results with 10 patients having posterior bone formation with some degree of foraminal or canal stenosis, 9 patients having posterior bone formation without foraminal stenosis, and 14 patients having no or greater than the expected posterior bone formation." The CSR did not record these as adverse events and the IPD provided by Medtronic did not include them as an adverse event category. However, IPD for this outcome were provided as data outside of the protocol as part of the Medtronic data submission. Medtronic clarified that these IPD were the result of a further investigation that looked at this outcome in both arms of the trial and these comparative data were reported in the paper by Haid et al.63 Medtronic also confirmed that heterotopic bone formation was systematically reviewed on CT scans by an investigator surgeon and based on protocol definitions at the time the studies were run these radiographic observations of heterotopic bone formation may not have been reported as an adverse event. Similarly the description of “14 patients with transient localised areas of bone remodelling in the vertebral body adjacent to the allograft dowel” in a publication relating to the INFUSE®/Bone Dowel
101 9 Results: Reliability of the Evidence
Pivotal RCT 54 could be interpreted as possible evidence of osteolysis. Corresponding IPD were provided for the INFUSE®/Bone Dowel Pilot and Pivotal trials as data beyond the protocol (although in this case only rhBMP-2 patients were explored). As IPD on possible osteolysis were available only for the rhBMP-2 patients we were unable to analyse these data Medtronic also confirmed that potential adverse events such as heterotopic bone formation, osteolysis and radiculitis that were not originally listed as adverse event categories do not appear as adverse event categories in any of the adverse event summary tables in the CSRs (which were based on clinical signs and symptoms). Therefore, an independent investigator looking specifically for these events is unlikely to be able to identify them in summary data from the Medtronic trials, whether published or not. A more detailed investigation across all CSRs was outside the scope of this review, but superficial examination of the other Medtronic randomised trials do not indicate other post hoc analyses of this nature.
102 9 Results: Reliability of the Evidence
9.2. Efficacy comparison We conducted separate meta-analyses of aggregate outcome data reported in CSRs and in publications and compared these against the results of the IPD meta-analyses presented previously to establish the extent to which conclusions about efficacy might depend upon the source of study data analysed. Pain and function Two completed Medtronic RCTs that were eligible for inclusion in our analyses were not published in scientific journals (INFUSE®/INTER FIX® ALIF Pilot RCT, rhBMP-2/BCP Canada pivotal RCT). Ten Medtronic RCTs had at least one publication; one (MAVERICK™ Disc pivotal RCT) did not include bone graft as a comparator. In the nine relevant Medtronic RCTs for which published data were available, presentation of the main pain and function outcomes (ODI or NDI, SF-36 PCS, back pain and leg pain) was generally complete. Seven trials presented either actual pain scores or change from baseline in these scores at most time periods. In some cases this information had to be estimated from figures. Variability in pain scores (e.g. as standard deviations) or uncertainty in mean pain scores, (e.g. as standard errors) was not commonly reported. Only one trial (INFUSE®/Bone Dowel Pilot RCT) presented these in full and only two other trials (INFUSE®/LT-CAGE Pilot RCT, INFUSE®/LT-CAGE Open pivotal RCT) recorded any measures of variability or uncertainty. Appendix XX gives more details on reporting of pain and function outcomes. A meta-analysis of pain outcomes using only published data was not possible because of the limited reporting of variability or uncertainly. We therefore chose to impute standard deviations for trials where these data were not provided (as is commonly done in systematic review practice). This was achieved by assuming that the standard deviation of pain scores in the rhBMP-2 or ICBG arm was the average of the standard deviations from those trials where this information was reported. The creation of imputed standard deviations provided sufficient information to perform a meta-analysis for all the pain outcomes at all time points, combining the actual and imputed data for the nine published Medtronic RCTs and published results from the Glassman trial. Data on pain and function scores in the CSRs for the Medtronic RCTs were, by contrast, complete for all trials. All trials presented data on all the pain outcomes, with standard deviations, at all time points. We performed meta-analyses of the four pain outcomes at all time points, combining the CSR data from the 11 Medtronic RCTs. We compared the results of the meta-analyses of published data and CSR data with the meta- analyses of IPD presented previously. Figure 42 shows the results of these meta-analyses. It can be seen in Figure 42 that the results of meta-analyses of aggregate data extracted (and imputed where necessary) from published papers, aggregate data from CSR data and trial IPD are generally almost identical for all four outcomes at all times. This is despite the fact that the analyses based on CSRs and on the IPD include data from unpublished trials. There is no evidence of any bias in the reporting of pain outcomes in the published RCT literature.
103 9 Results: Reliability of the Evidence
Figure 42 Meta-analyses of pain and function outcomes comparing complete IPD and clinical study report (CSR) data with publications
Fusion Seven of the nine Medtronic RCTs that were the subject of at least one publication presented data on the number of successful fusions at six, 12 and 24 months after surgery. The other two RCTs published data on successful fusion only at 24 months after surgery. The CSRs for the 11 Medtronic RCTs all presented complete data on successful fusion at six, 12 and 24 months after surgery. We performed meta-analyses for fusion success using the available published data and using the CSR data and compared these with the results from the IPD meta-analyses presented previously. Figure 43 shows the results of these meta-analyses. There are some differences in results across the three data sources. Including unpublished trials with the CSR and IPD leads to increased relative risks, increased heterogeneity and wider confidence
104 9 Results: Reliability of the Evidence
intervals at six months. This is mainly because of one unpublished trial (rhBMP-2/BCP Canada pivotal) with a very beneficial result at this time point. The results across studies become more consistent by 24 months. In any case, the results are generally consistent, have overlapping confidence intervals and there is no clear evidence of bias in the reporting of fusion in the published literature.
Figure 43 Meta-analyses of successful fusion comparing complete IPD and clinical study report (CSR) data with publications
105 9 Results: Reliability of the Evidence
9.3. Safety comparison As described previously reporting of adverse events was both limited and highly inconsistent among publicly available reports of Medtronic trials, precluding a meaningful direct quantitative comparison of adverse event results between data sources. However, for illustration, Table 20 shows the total number of adverse events that could be obtained from Medtronic publications against those observed in the Medtronic IPD. Since the same event may have been reported in more than one publication, our summation of events across publications is likely to overestimate the true number of events observed in the studies. Despite this likely over-counting, far fewer total adverse events were reported across publications than were collected in the IPD. Around 18.5% of all adverse events recorded in the IPD across all RCTs have been reported somewhere in the published literature (19% of rhBMP-2 events and 18% of ICBG events). Among RCTs of INFUSE®, around 10.5% of adverse events collected in the IPD have been reported in the published literature (12% INFUSE, 9% ICBG). Looking at only published RCTs did not substantially increase the proportion of reported adverse events (23% across all published studies; 11% across published INFUSE® studies). This finding suggests any independent systematic review of published data without access to IPD would not be able to fully evaluate the safety of rhBMP-2 in spinal fusion. The vast majority of observed events that were reported across the RCT-related publications come from a single publication related to the AMPLIFY™ pivotal RCT.102 For three categories (arthritis/bursitis, dural injury, vertebral fracture), slightly more events were reported in this publication than were observed in the IPD; the reason for these discrepancies is not clear but might be as a result of additional follow-up. The AMPLIFY™ pivotal RCT publication reported eight cancers in rhBMP-2 treated patients and two cancers in patients undergoing ICBG surgery.102 One publication of the results of the MAVERICK™ Disc pivotal RCT reported two cancers in rhBMP-2 treated patients and three cancers in patients receiving the MAVERICK™ disc replacement.80 There were no other published reports of cancer in the Medtronic publications.
106 9 Results: Reliability of the Evidence Table 20 Comparison of number of adverse events reported in Medtronic IPD versus Medtronic publications.
All studies (n=17) All published studies (n=11) INFUSE® studies (n=12) Published INFUSE® studies (n=9) IPD events Published IPD events Published IPD events Published IPD events Published events* events* events* events* (% of IPD) (% of IPD) (% of IPD) (% of IPD) Single-arm† rhBMP-2 733 169 (23.1) 588 169 (28.7) 661 169 (25.5) 588 169 (28.7) RCT rhBMP-2 1517 288 (19.0) 1222 288 (23.6) 621 75 (12.1) 592 75 (12.7) ICBG 1367 245 (17.9) 1080 245 (22.7) 541 47 (8.7) 516 47 (9.1) †Includes rhBMP-2 (INFUSE) comparison arm from the MAVERICK study. *”Published” values reflect maximum possible number of reported events summed across publications; some events may have been double counted. When calculating the number of adverse events (AEs) reported across publications: 1. Total number of AEs were counted, not just “device-related” or other subsets 2. Potential AEs (e.g. heterotopic bone formation, osteolysis) were counted if they were reported, regardless of whether they were explicitly referred to as “adverse events” or not 3. Number of events or the number of participants experiencing events were counted, whichever number was larger 4. Where two publications reported on exactly the same study population but the reporting of AEs differed, these values from each publication were summed 5. Publications reporting on subgroups of the same study population were not counted if there was a publication also available that reported AEs for the entire study population 6. Where the number of events was not reported according to treatment arm, they were assumed to be distributed evenly between arms
107 9 Results: Reliability of the Evidence
9.4. Risk of bias comparison It is widely recommended that systematic reviews and meta-analyses take into account the potential impact of various study biases on outcomes.152, 153 We investigated the extent to which the perceived risk of bias might be influenced by access to confidential study data. Results Figure 44 to 48 show risk-of-bias assessments based on the most complete data source for each study (its original protocol and CSR received from Medtronic) against the best assessment based on all the publicly available information (full papers, abstracts and the protocol as listed on ClinicalTrials.gov). Assessments were made separately for three types of outcome: patient-reported outcomes, successful fusion and adverse events. Judgements for different outcomes from the same study typically differed in terms of blinding. For example, Figure 44 shows fusion was generally the only outcome for which participants, personnel and outcome assessors were blinded. In cases where the only publicly available information on a trial was its ClinicalTrials.gov protocol, risk of bias due to incomplete outcome data and selective reporting was judged to be high. These tables show that any researcher with access to publicly-available information only (i.e. from trial registries and the research literature) would not be able to make an accurate assessment of the risk of bias associated with the Medtronic randomised trials. For example, access to Medtronic’s original RCT protocols allowed us to establish that the risk of bias arising from inadequate randomisation or concealment of allocation was consistently low but the limited reporting of trial characteristics in publicly available sources precluded a clear judgement from being reached. Consequently, any review of the published research literature would likely fail to acknowledge important methodological strengths of these trials and introduce unnecessary uncertainty into any subsequent conclusions about the effects of rhBMP-2.
108 9 Results: Reliability of the Evidence
Figure 44 Risk of bias judgment of fusion outcomes in trials included in reliability analyses comparing full vs. public access.
109 9 Results: Reliability of the Evidence
Figure 45 Risk of bias judgment of patient reported outcomes in trials included in reliability analyses comparing full vs. public access.
110 9 Results: Reliability of the Evidence
Figure 46 Risk of bias judgment of adverse events in trials included in reliability analyses comparing full vs. public access.
111 9 Results: Reliability of the Evidence
9.5. Summary of the reporting and reliability of Medtronic trials The published scientific literature only partly represents the total outcome data known to be collected on the effects of rhBMP-2. This is partly because some trials were never published or presented in any form or have been presented only in brief in conference abstracts. Even the published trials did not report all the collected outcomes. Outcome data relating to pain, disability, function and fusion are generally reported in the published literature but some data have to be imputed to permit meta-analysis. Poor reporting of methodological characteristics typically means that judgements about risk of bias based solely on publications and publicly available protocols would likely lead to an unnecessary level of uncertainty about the effects of rhBMP-2 as reported in these trials. Despite these observations, meta-analysis of the publicly-available evidence on these efficacy outcomes does not lead to substantially different results from meta-analysis of the IPD or the confidential aggregate data reported in CSRs. Therefore, reporting of these results does not appear to be substantially biased. Adverse event data are presented incompletely and inconsistently in the published literature compared with the CSRs and the rationale for presenting certain adverse events over others is rarely clear. The absolute number of adverse events reported in the literature is substantially smaller than reported in the IPD or CSRs (this is true of both rhBMP-2 and ICBG). Potential adverse events that caused concern in the post-marketing period (heterotopic bone formation, radiculitis, osteolysis) were not routinely collected or reported in the original Medtronic studies. The reasons given for this were that such events were not unanticipated and that adverse data collection was based on signs and symptoms. Therefore, it is not possible to independently establish the rates of these events in the Medtronic trials except where they were specifically the subject of additional investigations outside the protocol. In conclusion, when compared with the IPD analysis, a systematic review of the published literature alone would produce similar findings in relation to clinical and radiological efficacy outcomes but with greater uncertainty (such as wider confidence intervals around effects) because not all data were published. Any conclusions about adverse effects would be based on extremely limited and inconsistent data (dominated by a single AMPLIFY publication), likely resulting in very uncertain conclusions. To achieve even this, a large amount of time and effort needs to be expended in identifying, extracting and assessing all the relevant publications. Even with careful checking there is potential for error. Access to and analysis of IPD afforded the greatest opportunity for interrogating data and enabled analyses that would not otherwise be possible. Our investigation also suggests that even if access to IPD were not available, access to full original study protocols and CSRs would produce more accurate, reliable and robust findings with less time and effort than relying on publicly available data.
112 10 Discussion and Conclusions
10. Discussion and Conclusions
Strengths and limitations of the systematic review and IPD meta-analyses This project aimed to provide a fair and robust assessment of the safety and effectiveness of rhBMP-2 in spinal fusion surgery. Throughout, we followed a pre-specified protocol and have reported any additions to or deviations from our planned approach. We focused specifically on the use of rhBMP-2 in spinal fusion surgery as treatment for degenerative disc disease and related spinal conditions. We did not consider rhBMP-7 and did not consider usage outside of the spine. We included all formulations of rhBMP-2 including the licensed product INFUSE® and the unlicensed product AMPLIFY™. Our aim was to be comprehensive and we did not limit our scope to Medtronic trials but sought to identify and include all relevant research studies. We restricted assessment of efficacy to RCTs. For adverse events we also considered non randomised comparative studies in supplementary analyses as these may identify unexpected events that RCTs did not plan to assess, or identify rare adverse events that RCTs are not typically powered to detect.154 Though such studies can be prone to confounding and bias, this may be less serious in the evaluation of unintended adverse effects than for intended beneficial effects (if the reason for allocation to rhBMP-2 or not rhBMP-2 is associated more with beneficial than with adverse outcomes). 154 Given differences in intervention, study design, and comparators, we did not perform meta- analyses of the non-randomised studies, and throughout this report emphasise the uncertainties brought about by these limitations. The data and documentation provided by Medtronic were extensive and although we used data from summary tables of adverse events in CSRs, and used these reports to try to clarify specific issues that arose, it was never our intention to carry out a detailed investigation of every document provided by Medtronic. Medtronic RCTs used acceptable and consistent methods for randomisation and concealment of allocation across trials. Some participants were excluded from the trials after randomisation, which, if linked to treatment and outcome, can introduce bias. However, with the exception of the AMPLIFY™ (rhBMP-2/CRM) Pivotal RCT and the INFUSE® Bone Dowel Pivotal RCT, the proportion of excluded patients was generally low and not obviously imbalanced across the treatment groups. Up to 24 months the rate of loss to follow up was acceptable and balanced across treatment arms. Beyond 24 months outcome data were too limited to permit reliable analysis. The data provided were otherwise generally complete and consistent. IPD were provided for all the outcomes specified in the trial protocols, although some of the adverse outcomes that we had hoped to see evaluated were absent from the trial protocols. Our analyses were done at a series of time points up to 24 months after surgery, beyond which the IPD provided were limited and analysis would have been unreliable. We are, therefore, unable to comment on efficacy and safety beyond 24 months and results should be interpreted in light of this. In all trials, assessments of fusion status were based on blinded evaluations made by at least two independent assessors. We were concerned, on the basis of information contained in the CSR, that blinding of outcomes may have been broken in the INFUSE®/INTER FIX™ PLIF RCT. However, Medtronic clarified that the reported request to the FDA to amend the protocol to permit unblinded radiologic assessment pertained only to making assessors aware of the timing of radiographs and not of assigned treatment. We did not check that fusion status was accurately derived from component scores supplied. However, this would have checked only the internal consistency of the data supplied according to the Medtronic definition of fusion. Without being able to assess the original radiographs independently we were unable to tackle the more important issue of whether assessments were made accurately. We did, however, note that a high proportion of the assessments made by two independent assessors did not agree and required adjudication by a third person. This highlights that the process of assessing fusion, even when using objective criteria, involves a degree of subjectivity based on the interpretation of a radiograph. We checked the internal consistency of the ODI and SF-36 scores by recalculating them from the component data scores provided, and found no errors. As patients and physicians were not blinded to treatment and most efficacy outcomes were patient-reported, these may have been subject to bias, whereby a patient’s outcome score is partly influenced by their knowledge of which treatment they received. The observed effects on pain and function should be interpreted in light of this potential bias.
113 10 Discussion and Conclusions
The potential influence of knowledge of treatment received similarly introduces concern about risk of bias in a number of the patient-reported adverse pain events. Although the more objectively defined adverse outcomes may have been at lesser risk of bias, there remains potential for bias in assignment of the severity and relatedness of adverse events to treatment. Design of the Medtronic trials has previously been criticised on the grounds that because control arms were not optimal, findings were intrinsically biased in favour of rhBMP-2.11 In usual clinical practice for posterior spinal fusion, ICBG would be supplemented by local bone from the facet, lamina, or spinous process. Although it has been reported that trial protocols dictated that local bone should be discarded and only small quantities of ICBG were allowed to be used as the sole grafting source,155 the protocols did not require this. Nonetheless, only small volumes of ICBG and no local bone were used in the trials. Medtronic have previously responded that using the local bone in the ICBG group would have introduced a confounding factor. However, we believe these criticisms are reasonable and that it would have been preferable if trials had addressed whether rhBMP-2 was better than best practice, which could have included use of local bone. These are design issues that cannot be addressed retrospectively and no systematic review or meta-analysis would be able to do other than acknowledge them and the fact that they may have impacted on findings. The Medtronic trials have also been criticised on grounds of their design as non-inferiority trials11 Whilst we agree that this is a criticism of individual trials, in the context of our meta-analysis it is less important. Our re-analysis of the data did not seek to establish non-inferiority since our intention was to obtain the best estimate of effect from the totality of reliable research evidence. The extent and nature of the data available, from 1,411 individual participants in 12 randomised controlled trials, coupled with the robust and transparent methods used, make this (together with the mirror independent assessment being carried out by the OHSU team, which we assume will be of similar depth and quality) the most reliable analysis of the effectiveness and safety of rhBMP-2 compared to ICBG in spinal fusion surgery completed to date, and consequently the best evidence to inform treatment and policy decisions. Findings Efficacy All our meta-analyses of efficacy used individual participant data from randomised controlled trials. These showed reduction in pain scores to be generally slightly better among the rhBMP-2 patients from about six months onwards. At 24 months the change from baseline in ODI was approximately three and a half percentage points lower (better) for rhBMP-2 patients than for those treated with ICBG. Although statistically significant, this difference is small and falls below the lower bound of the minimum change in score that is deemed to be clinically important, which has been estimated by different observers as between 4 and 17 points.156 It is also less than the 4.1 point benefit observed in the MRC spine stabilisation trial,157 which concluded that there was no evidence of spinal fusion surgery being any more beneficial than intensive rehabilitation. Furthermore, ODI improved significantly over time in both rhBMP-2 and ICBG patients, such that the extra benefit of rhBMP-2 is relatively small in comparison. There was no evidence of a difference in leg pain between the treatment groups. Although data were provided for graft site pain, this was measured only for those allocated to ICBG, and, therefore, no comparison was possible. We would, however, expect that graft site pain, if significant, would be expressed through increased analgesic demand in the post-operative period, and to contribute to general measures of health and wellbeing. We found that surgery using rhBMP-2 was not associated with a reduction in analgesic demand over that of surgery with ICBG. Improvement in pain and physical function, as measured by the SF-36 Physical Component Score, was about 2 percentage points better for rhBMP-2 than ICBG at 24 months. Again, although statistically significant, this between-groups difference is set against a much greater improvement over time for all patients, irrespective of which intervention they received, and is below the 5.41 point threshold considered in a key AMPLIFY™ trial publication as being a clinically important improvement.102 Pain, Physical functioning and Physical role component scores of the SF-36 favoured rhBMP-2 by about five percentage points from three months onwards. Our analyses showed no notable effect on the SF-36 mental health component scores, nor was there evidence of a difference for overall general health. We found no clear evidence of a difference in time to return to work between rhBMP-2 and ICBG. Operation times were reduced by around 20 minutes, presumably due to a time saving from not having to harvest bone from the iliac crest.
114 10 Discussion and Conclusions
There is clear evidence that use of rhBMP-2 results in higher rates of fusion at all measured timepoints, with an observed 12% statistically significant increase in the rate of successful fusion at 24 months based on Medtronic criteria of fusion. Those in the rhBMP-2 group achieved a fusion rate of 81% compared with 69% of ICBG-treated patients at 24 months. Of note is the substantial statistical heterogeneity in the relative risk of successful fusion across trials, particularly at six and twelve months. We found no evidence of a difference in the rates of fusion across different surgical approaches. Successful fusion is likely to benefit patients if it translates to improvements in pain, functioning, or in a reduction for the need for further intervention. However, successful fusion and pain reduction do not seem to be strongly correlated. Trials with higher fusion rates for rhBMP-2 did not also achieve greater pain reduction. In the trials with lower fusion rates among rhBMP-2 recipients than ICBG recipients, pain reduction was still greater among the rhBMP-2 patients. This would suggest that successful fusion (as determined by Medtronic criteria) does not in itself appear to improve pain at 24 months post-surgery. This finding may have been influenced by the relatively stringent definition of successful fusion used in the Medtronic trials, whereby some cases may have 'fused' to some degree but not sufficiently to meet the trial criteria. Therefore, any subsequent improvements in pain in such cases would not be formally attributed to successful fusion. An explanation that cannot be ruled out is that, because both patients and physicians knew the type of treatment received and fusion outcome, this may have introduced bias into assessments of pain, favouring the "new" treatment over ICBG. One of the key benefits of spinal fusion is that it will reduce hardware strain, which should ultimately result in a reduction in late hardware failure. However, the 24 months endpoint, for which the majority of data were available in the present analysis, may be too early to assess this outcome. We also considered whether a range of patient factors including age, sex, smoking, alcohol consumption, body mass index, diabetic status, and history of spinal surgery for back pain could alter the effectiveness of rhBMP-2 in surgery. With the possible exception of previous spinal surgery, we found no clear evidence of interactions between patient characteristics and the effectiveness of rhBMP-2. In other words, improvement in the likelihood of successful fusion and the modest improvements observed for pain and function scores from six months onward were consistent across these different types of individuals. Thus, these data suggest that no particular type of individual (e.g. older or younger, male or female, smoker or non-smoker) benefits any more or any less than average from rhBMP-2 surgery. One possible exception is that in people with a history of spinal surgery, rhBMP-2 is no more effective than ICBG surgery in reducing ODI or in improving fusion rates. Given the number of analyses performed, however, this result may be a chance finding. Our finding that smokers do not seem to benefit any more from rhBMP-2 than non-smokers in terms of ODI, pain, or rates of fusion contrasts with a previous analysis of 148 participants from three centres in the AMPLIFY™ pivotal RCT which suggested that rhBMP-2 might enhance fusion rate in smokers 101 Safety Adverse event information provided in the IPD and summarised in the Medtronic CSRs did not record some of the events that we set out to evaluate a priori. In particular, heterotopic bone formation, osteolysis, and radiculitis were not mentioned in the study adverse event report forms, were not captured in the CSR summary tables of adverse events, and were not provided as standard trial IPD. The reasons why these potential complications were not recorded in the Medtronic trials was not obvious, though correspondence with Medtronic through YODA clarified that the classification of adverse events was based on clinical signs and symptoms, and that observations such as heterotopic bone formation were not considered to be adverse events in the original trial protocols and so were not recorded or reported as such. In the INFUSE®/INTER FIX™ PLIF RCT, following suspension of the trial, a separate investigation of heterotopic bone formation by retrospective review of radiographs of the rhBMP-2 patients’ records was carried out. This was reported in the risk analysis section of the CSR (but not included in adverse event tables) and reported in the trial publication, although not explicitly as an adverse event.63 Relevant IPD from these trials were provided by Medtronic as additional data beyond the protocol which we were able to analyse. Similarly, in the INFUSE® Bone Dowel Pilot and Pivotal RCTs, localised bone remodelling or resorption around the allograft bone dowels was observed in some patients who received rhBMP-2, leading to a retrospective review of the patients in the investigational group. Relevant IPD from these trials were provided by Medtronic as additional data beyond the protocol but as these were available only for the rhBMP-2 patients we were unable to analyse these data. No specific information or data on these outcomes were available from the other Medtronic trials. We note that the currently ongoing Medtronic trial of INFUSE® in TLIF
115 10 Discussion and Conclusions
which began in 2011 does not include either heterotopic bone formation or osteolysis as adverse event outcomes. Analyses of the adverse event IPD that were recorded at all time periods in the Medtronic sponsored trials showed some potential complications to be more common among rhBMP-2 patients, including: back and leg pain, other pain, implant-related events, infections, wound complications, and neurological, urogenital and vascular events. However, absolute numbers of events were small and most of these differences were not statistically significant. At or just after surgery back and leg pain was almost twice as common among rhBMP-2 patients. The greatest increased risks were for retrograde ejaculation and arthritis or bursitis; both were approximately three times higher in the rhBMP-2 group, but the numbers are small and results were not statistically significant. Spinal events (e.g. stenosis) were less common among rhBMP-2 patients, but results were not statistically significant. A previous review11 has criticised publications of the Medtronic trials for failing to focus on the early active period where rhBMP-2 related events are most likely to occur, arguing that, by reporting events over the longer term, statistical noise of random events masked potentially serious early complications. Here, we were able to look at the early post-operative period and found back and leg pain to be more commonly reported as an adverse event in the rhBMP-2 group. This observed increase in early post-operative pain was not detected by the formal pain scales used in the efficacy analysis, possibly because the first pain measurement was not taken until six weeks after surgery by which time some problems could have resolved, or because of potential bias associated with unblinded patient reported outcomes. Adverse events overall were more common in the rhBMP-2 group in the early post-operative period, but there was no difference in the incidence of “severe” events. There was no overall difference in the need for second surgical procedures between rhBMP-2 and ICBG patients. In common with other trials undertaken for regulatory purposes, the numbers of adverse events recorded in this dataset are considerably higher than would be expected in routine clinical practice. Our inspection of the patient narratives included in the INFUSE®/INTER FIX™ PLIF RCT CSR showed that adverse events included many of what might be regarded as trivial events. This observation is likely to reflect the high degree of vigilance in the collection of adverse event data within the Medtronic studies. However, it also creates the problem that rare, and potentially important, events may be masked by larger numbers of irrelevant events, depending on how these are collected, categorised, and reported. It should be noted that this issue is a general one that relates to beyond the Medtronic trials. Based on comparative IPD (i.e. including only data where both treatment groups were followed up over the same period), our analysis indicated that the overall incidence of cancer is higher in the rhBMP-2 group. A one-stage meta-analysis indicated a relative risk of 1.98 suggesting that patients receiving rhBMP-2 were almost twice as likely as those receiving ICBG to develop cancer. This finding was not, however, statistically significant with a confidence interval consistent with a 14% reduction in risk to a 454% increase in risk. The relative risk of cancer was similar across trials. In particular the relative risk of cancer in the AMPLIFY™ (rhBMP-2/CRM) and rhBMP-2/BCP Canada Pivotal RCT, which used different preparations of rhBMP-2 at higher dose, is not greater than in the trials which used INFUSE®. This is contrary to the findings of a recent review which suggested that cancer risk with rhBMP-2 may be dose dependent.158 Medtronic also provided individual post-marketing adverse event forms from the FDA MedWatch reporting system. These are voluntary reports intended to detect safety hazard signals rather than quantify the incidence of adverse events associated with a particular treatment. Not all events that have occurred in practice will have been reported, and we cannot determine the motivations behind deciding whether to report or not. We cannot exclude the possibility of some direct overlap between these data and the adverse events reported in the wider literature. Nonetheless, the most commonly reported events across all spinal uses were inflammatory cyst formation and heterotopic bone formation, most of which were in off-label use (fewer than 8% of known spinal-related reports were related to ALIF procedures). As expected, all reported incidences of dysphagia were in cervical procedures. There were no reports of retrograde ejaculation. We also reviewed the wider literature for data on adverse events. After initial mapping of all the available published studies reporting adverse events in patients receiving rhBMP-2 in spinal fusion surgery, we decided to focus on studies that might provide more accurate or reliable data and therefore excluded all non-comparative and very small (n<10) studies. While estimates from the
116 10 Discussion and Conclusions
remaining studies may also be vulnerable to confounding and bias, they do provide some insight on the use of rhBMP-2 in wider clinical practice. Potential morbidity associated with the use of rhBMP-2 needs to be separated from morbidity associated with the procedure itself, which will vary according to the surgical approach. For example, in an anterior cervical discectomy tissue swelling could result in airway difficulties or dysphagia; in anterior lumbar interbody fusion injury to or inflammation around the nerves of the hypogastric plexus could contribute to retrograde ejaculation; in posterior lumbar fusion with pedicle screws local rhBMP-2 in the disc space or posterolateral gutter could cause nerve root swelling/radiculitis and pain radiating down the leg. All of these complications can occur from these surgeries in the absence of the use of rhBMP-2. Therefore, wherever possible we separated data arising from different surgical procedures (ALIF, PLF, ACDF etc.) in order to identify potential site/surgery specific adverse events. Owing to possible confounding in most of these non-randomised studies, and variation in surgical approaches, comparators and observation timing, we did not combine the results of individual studies to obtain average estimates of the risk of adverse events. Risk of bias and confounding varied across studies, and although some were judged to be at low risk of bias, as a body of evidence these data from the wider literature need to be interpreted cautiously. They did, however, suggest that a number of potential complications were more common in patients treated with rhBMP-2. In particular, higher rates were observed among studies reporting heterotopic bone formation, osteolysis, and leg pain/radiculitis in PLF/PLIF/TLIF surgery, and retrograde ejaculation in ALIF surgery. There are clearly limitations of using the wider literature to investigate adverse events. It may not be a fair reflection of the adverse events that occur in practice, since the publications only represent a fraction of the total number of procedures using rhBMP-2, and the observed rates of adverse events could be influenced by a number of clinical (e.g. patient population, clinician experience, surgical approach, rhBMP-2 dose) and methodological (e.g. selection of patients for treatment groups, approach used to identify and quantify events) factors. There may be a susceptibility to reporting bias whereby, for example, studies examining the case notes of past patients may be motivated to publish if they observe an excess of events on an investigational arm and perhaps less so if the excess is observed for an established treatment or procedure where results may be dismissed as chance findings. Reliability of the evidence Had IPD and CSR data not been available we and others wishing to carry out a systematic review of the effects of rhBMP-2, would have been restricted to the information that was available in the published literature. Given previous concerns expressed about the Medtronic sponsored trials,11 we considered it important to embed our systematic review within a wider exploration of reporting practice. Our intention in doing so was to explore whether published results are an accurate reflection of the underlying trial data. More specifically, we intended to establish whether systematic review and meta- analysis of all trials and all outcomes using patient-level data would give results that differed from those obtained using just the published data. The intention was also to place our IPD analyses in context, and to explore whether there were general lessons about trial design, conduct and disclosure that would be of value in future. Our systematic review of published information found the literature to be complex and fragmented. Three single-arm studies (rhBMP-2/BCP Mexico Pilot, INFUSE®/TELAMON PEEK Instrumented PLIF Pilot, rhBMP-2/CRM 2-level Pilot) and three RCTs (INFUSE®/INTER FIX™ ALIF Pilot RCT, rhBMP- 2/BCP Canada Pivotal RCT, and INFUSE®/CORNERSTONE® ACDF Pivotal RCT) have never been fully published, though the INFUSE®/CORNERSTONE® ACDF Pivotal trial was abandoned after recruiting only three patients. Examination of the trial protocols and CSRs suggest that even among the studies that have been published, a substantial proportion of collected outcomes have not been reported in trial publications. Key efficacy outcome data relating to pain, function, and fusion were consistently reported in the published literature (although for seven out of 10 published trials, estimates of variance had to be imputed to permit meta-analysis) and for efficacy outcomes we found no substantial differences between meta-analysis based on the published aggregate data and those based on the IPD. Therefore, reporting of efficacy in these trials, although incomplete, does not appear to be substantially biased.
117 10 Discussion and Conclusions
In contrast, we found adverse events to be incompletely and inadequately described in the trial publications. Published papers provided far less information than was available in the confidential CSRs (or in the supplied IPD). The way in which the adverse event data were presented in the literature was highly inconsistent and the rationale for presenting some adverse events and not others was rarely clear. Brief, vague statements in some publications that simply noted “no unanticipated device-related adverse events” were inadequate and unhelpful. In our view, such statements, without supporting evidence, should not be considered acceptable for publication. Limited and inconsistent reporting of adverse data suggests that the published academic literature alone is an inadequate resource for evaluating the safety of rhBMP-2. This is not unique to the Medtronic trials, and several authors have noted poor reporting of adverse events from trials in the academic literature more widely.159-162 As discussed previously, some potential adverse events that have caused concern in the post- marketing period (e.g. heterotopic bone formation, radiculitis, osteolysis) were not routinely collected or reported in the original Medtronic studies. In one CSR, even though heterotopic bone formation led to trial suspension and a review found substantially more cases in the rhBMP-2 arm, this was noted but not formally labelled as an adverse event in the report or subsequent academic publication.63 In conclusion, when compared with the IPD analysis, a systematic review of the published literature alone would produce similar findings in relation to efficacy outcomes, but would likely have more uncertainty around these findings due in part to the limited information available to make judgements about risk of bias. Incomplete reporting of methodological characteristics typically means that judgements about risk of bias based on academic publications alone will be unnecessarily negative. Any conclusions about adverse effects would be based on extremely limited and inconsistently reported data, likely to result in very uncertain and potentially misleading results and conclusions. To achieve even this, a huge amount of time and effort needs to be expended in identifying, extracting and assessing all the relevant publications. Even with careful checking, there is the potential for error in these laborious procedures. Wider implications In assessing the evidence on efficacy and safety we were in the fortunate position of having available the raw data from the trials for re-analysis, and Medtronic are to be commended for making these data available. In our IPD analyses we have been able to address a number of issues that previous reviews of publicly available data could not. We have been able to look across trials in a consistent way and across time. We have had access to adverse event data that are not available in the published literature and have been able to analyse data in the period immediately after operation, which previous commentators have identified as being vital.11 Importantly, we have been able to examine relationships between variables in a way that is seldom possible with aggregate data. Our exploration of patient subgroups highlighted a lack of interaction between patient covariates contrary to previous more limited analysis that suggested a differential benefit of rhBMP-2 in smokers.101 Our exploration of the relationship between fusion and pain outcomes would not have been possible without access to the IPD. Furthermore, it took considerable time and effort to identify the trial publications and to establish which papers reported what trials, outcomes, and timepoints before extracting and checking data. IPD meta-analyses are often dismissed as being impractical for resource reasons, taking too long and costing too much. Although we acknowledge that here IPD were provided from the outset (whereas most IPD meta-analyses involve an initial stage of establishing a collaboration and collecting and collating data from multiple sources), we did require time to collate, understand, and check data provided across more than 400 separate data files before beginning the analyses. Overall, our IPD analyses took no longer than the parallel systematic review of the published literature. This lends weight to arguments that IPD systematic reviews and syntheses are a gold standard that should be adopted when addressing important clinical questions.163 Furthermore, this supports the YODA initiative to promote deposition of trial IPD for independent re-analysis (or alternative models to make trial IPD available for research purposes) as the best way of investigating controversial areas. Original study protocols and CSRs are generally treated as confidential and are not usually made available to researchers. Access to these documents could potentially allow a more complete and robust analysis than is possible with the typical approach of extracting and analysing summary data from the published literature. In addition, the time and effort required could be substantially reduced
118 10 Discussion and Conclusions
for some systematic reviews. Importantly, we found the CSRs to be a rich source of data and information. Although the data from these would not support more complex analyses that require linkage between covariates or outcomes, they did permit far more detailed investigation and analysis than was possible using journal articles alone. With suitable safeguards to protect patient privacy and transparency around vested and potential conflict of interest, access to participant-level data for research purposes, should be the ultimate goal of initiatives to increase access to trial data. However, our findings suggest that in the absence of IPD, access to complete study protocols and the clinical study reports that are produced by manufacturers for regulatory or other purposes would produce more accurate, reliable, and robust findings with less time and effort than relying on publicly available data alone. Release of clinical study reports could therefore be an important and potentially almost immediately achievable first step toward increasing transparency of trials conducted for regulatory purposes. We note however, that not all trials are conducted with regulatory intent and consequently do not produce clinical study reports. Conclusions on safety and effectiveness Individual participant data meta-analyses of this generally robust body of evidence indicate that rhBMP-2 improved rates of fusion (according to the Medtronic definition used in the trials) compared with ICBG, at all measured time points, though the majority of patients receiving either treatment achieved successful fusion at 24 months. Use of rhBMP-2 also resulted in statistically significant improvements in disability (ODI) and quality of life (SF-36) up to 24 months but these differences are small and fall below the thresholds regarded to be clinically significant. The difference between rhBMP-2 and ICBG is set against much larger improvements over time for both groups. Furthermore, owing to the lack of blinding on these measures, we cannot rule out that knowledge of treatment received may have influenced outcomes in favour of rhBMP-2. These data suggest that the increase in successful fusion at 24 months does not seem to translate into clinically significant benefits in pain reduction, function, or quality of life. Meta-analyses of IPD suggest there is an increased risk of adverse pain events associated with rhBMP-2 in the immediate post-surgical period. Analyses of the Medtronic data also indicate a possible increase in the risk of cancer associated with rhBMP-2. However, it should be borne in mind that the overall absolute risk of cancer is low in both groups. Studies published in the wider literature and post-marketing data raise concerns about other adverse events not captured or easily apparent in the IPD provided, including heterotopic bone formation, osteolysis, retrograde ejaculation, urinary retention, and dysphagia. Owing to the non-randomised nature of the studies and differences between them, the strength of this as a body of evidence is weak and findings should be interpreted cautiously. In summary, rhBMP-2 seems to increase the chance of successful fusion, according to Medtronic definitions, but this does not translate to clinically meaningful benefits in pain reduction, function or quality of life. The small benefits in these outcomes observed after six months also come at the expense of more pain in the immediately post-operative period, and a possible increased but small absolute risk of cancer. Chances of experiencing other potential complications may also be increased. We suggest it is very important that these findings are expressed clearly to patients in order that they, with their doctors, can make informed choices about the type of surgery that they would prefer.
119
References 1. Glassman SD, Carreon LY, Djurasovic M, Campbell MJ, Puno RM, Johnson JR, et al. RhBMP-2 versus iliac crest bone graft for lumbar spine fusion: a randomized, controlled trial in patients over sixty years of age. Spine (Phila Pa 1976) 2008;33:2843-9. 2. Shapiro S, Rodgers RB, Sloan R, Altstadt T, Miller J. A randomized trial of instrumented posterior lumbar interbody fusion using machined cortical wedges/local bone with or without rhBMP2 in the treatment of degenerative lumbar spondylolisthesis with stenosis. Neurosurgery 2005;57:398. 3. Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. Growth Factors 2004;22:233-41. 4. Ebara S, Nakayama K. Mechanism for the action of bone morphogenetic proteins and regulation of their activity. Spine 2002;27:S10-5. 5. Loeffler BJ, Kellam JF, Sims SH, Bosse MJ. Prospective observational study of donor-site morbidity following anterior iliac crest bone-grafting in orthopaedic trauma reconstruction patients. J Bone Joint Surg Am 2012;JBJS Express Article. 6. Sasso RC, LeHuec JC, Shaffrey C, Spine Interbody Research G. Iliac crest bone graft donor site pain after anterior lumbar interbody fusion: a prospective patient satisfaction outcome assessment. J Spinal Disord Tech 2005;18 Suppl:S77-81. 7. Cahill KS, Chi JH, Day A, Claus EB. Prevalence, complications, and hospital charges associated with use of bone-morphogenetic proteins in spinal fusion procedures. JAMA 2009;302:58-66. 8. Ong KL, Villarraga ML, Lau E, Carreon LY, Kurtz SM, Glassman SD. Off-label use of bone morphogenetic proteins in the United States using administrative data. Spine (Phila Pa 1976) 2010;35:1794-800. 9. Boden SD, Zdeblick T, Sandhu H, Heim S. The use of rhBMP-2 in interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine (Phila Pa 1976) 2000;25:376-81. 10. Burkus J, Gornet M, Dickman C, Zdeblick T. Anterior lumbar interbody fusion using rhBMP-2 with tapered interbody cages. J Spinal Disord Tech 2002;15:337-49. 11. Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. Spine J 2011;11:471-91. 12. Golder S, Loke Y. Search strategies to identify information on adverse effects: a systematic review. J Med Libr Assoc 2009;97:84-92. 13. Systematic review of BMP-2 for spinal fusion: call for evidence BackLetter 2012;27:47. 14. Fairbank J, Pynsent P. The Oswestry Disability Index. Spine (Phila Pa 1976) 2000;25:2940-52. 15. Ware J, Kosinski M, Keller S. SF-36® physical and mental health summary scales: a user's manual. Boston, MA: The Health Institute; 1994. 16. Higgins JPT, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343. 17. Stewart LA. Practical methodology of meta-analyses (overviews) using updated individual patient data. Stat Med 1995;14:2057-79. 18. Riley RD, Lambert PC, Abo-Zaid G. Meta-analysis of individual participant data: rationale, conduct, and reporting. BMJ 2010;340:221. 19. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-88. 20. Higgins J, Whitehead A, Turner R, Omar R, Thompson S. Meta-analysis of continuous outcome data from individual patients. Stat Med 2001;20:2219-41. 21. Turner RM, Omar RZ, Yang M, Goldstein H, Thompson SG. A multilevel model framework for meta-analysis of clinical trials with binary outcomes. Stat Med 2000;19:3417-32. 22. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539-58. 23. SAS 9.2 ed. [program] Cary, NC, USA.: SAS Institute Inc. 24. R Development Core Team. R: A language and environment for statistical computing. [program] Vienna, Austria: R Foundation for Statistical Computing; 2008. 25. Wells G, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. [cited 2012 9th October]. Available from: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. 26. Loke Y, Price D, Herxheimer A. Chapter 14: Adverse effects. In: Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011): The Cochrane Collaboration; 2011. Available from: www.cochrane-handbook.org
120
27. Sanderson S, Tatt I, Higgins J. Tools for assessing quality and susceptibility to bias in observational studies in epidemiology: a systematic review and annotated bibliography. Int J Epidemiol 2007;36:6666-76. 28. INFUSE®Bone graft in transforaminal lumbar interbody fusion for degenerative disease of lumbosacral spine (TLIF). [cited 2012 9th August]. Available from: http://clinicaltrials.gov/ct2/show/NCT01415908?term=rhBMP-2&rank=24. 29. Medtronic Inc. [LT cage pilot document] [Protocol] [Received from Medtronic 1st December 2011]. 30. Medtronic Inc. A prospective randomized feasibility investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the tapered interbody fusion device for anterior lumbar interbody fusion in patients with degenerative disc disease. Vol. 1 [Final report received from Medtronic 1st December 2011]: Medtronic Sofamor Danek; 1999. 31. Medtronic Inc. [LT cage pilot. Vol. 2] [Final Report. Received from Medtronic 1st December 2011]. 32. Pilot study of rhBMP-2/ACS/LT-CAGE® for anterior lumbar interbody fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491373?term=medtronic+fusion&rank=7. 33. Dickman C. A prospective randomised lumbar fusion study comparing rhBMP-2 with autograft for cage ALIF. Neurosurgery 2001;49:521. 34. Gornet M, Burkus J, Dickman C. rhBMP-2 with tapered cages: a prospective, randomised lumbar fusion study. In: North American Society of Spine Surgeons Meeting; October 30-November 3, 2001. Seattle, Washington; 2001. 35. Burkus J, Heim S, Gornet M, Zdeblick T. Is INFUSE bone graft superior to autograft bone? An integrated analysis of clinical trials using the LT-CAGE lumbar tapered fusion device. J Spinal Disord Tech 2003;16:113-22. 36. Burkus JK, Dorchak JD, Sanders DL. Radiographic assessment of interbody fusion using recombinant human bone morphogenetic protein type 2. Spine (Phila Pa 1976) 2003;28:372-7. 37. Burkus JK. Bone morphogenetic proteins in anterior lumbar interbody fusion: old techniques and new technologies. Invited submission from the Joint Section Meeting on Disorders of the Spine and Peripheral Nerves, March 2004. J Neurosurg Spine 2004;1:254-60. 38. Chhabra A, Bentley A, Donnell D, Greenberg D, Oliver E, Alt V. A health-economic evaluation of rhBMP-2 in spine fusion surgery in Germany and UK. Value Health 2006;9:A201. 39. Chhabra A, Nicklasson L, Greenberg D, Alt V. A cost-effectiveness analysis of rhBMP-2 in spine fusion surgery in Sweden and Denmark. Value Health 2007;10:A97. 40. Van Genugten M, Chhabra A, Alt V. A cost-effectiveness analysis of RhBMP-2 in spine fusion surgery in the Netherlands. Value Health 2008;11:A532-A33. 41. Alt V, Chhabra A, Franke J, Cuche M, Schnettler R, Le Huec J-C. An economic analysis of using rhBMP-2 for lumbar fusion in Germany, France and UK from a societal perspective. Eur Spine J 2009;18:800-6. 42. Burkus JK, Gornet MF, Schuler TC, Kleeman TJ, Zdeblick TA. Six-year outcomes of anterior lumbar interbody arthrodesis with use of interbody fusion cages and recombinant human bone morphogenetic protein-2. J Bone Joint Surg Am 2009;91:1181-9. 43. Burkus JK, Gornet MF, Glassman SD, Slosar PJ, Rosner MK, Deckey JE, et al. Blood serum antibody analysis and long -term follow-up of patients treated with recombinant human bone morphogenetic protein-2 in the lumbar spine. Spine (Phila Pa 1976) 2011. 44. Medtronic Inc. [LT cage open pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2000. 45. Medtronic Inc. A clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the NOVUSTM L T device for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease [Final report] [received from Medtronic 1st December 2011]; 2002. 46. Pivotal study of rhBMP-2/ACS/LT-CAGE® device for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491386?term=medtronic+fusion&rank=20. 47. Kleeman TJ, Ahn UM, Talbot-Kleeman A. Laparoscopic anterior lumbar interbody fusion with rhBMP-2: a prospective study of clinical and radiographic outcomes. Spine (Phila Pa 1976) 2001;26:2751-6. 48. Zdeblick T, Heim S, Kleeman T. Laparoscopic approach with tapered metal cages: rhBMP-2 versus autogenous bone graft. In: Annual Meeting of the North American Spine Society. Seattle, WA, USA; 2001. 49. Medtronic Inc. [LT cage lap pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2000.
121
50. Medtronic Inc. A clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the NOVUS™ L T device for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2002. 51. A pivotal study of rhBMP-2/ACS/LT-CAGE® device for anterior lumbar interbody fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491425?term=medtronic+fusion&rank=18. 52. Burkus J, Transfeldt E, Ktchel S. rhBMP-2 and cortical dowels in the lumbar spine: long-term outcomes. In: North American Society of Spine Surgeons. Seattle, Washington; 2001. 53. Burkus KJ, Transfeldt E, Kitchel S, Watkins R, Balderston R. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine (Phila Pa 1976) 2002;27:2396-408. 54. Burkus J, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery. J Bone Joint Surg Am 2005;87:1205-12. 55. Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine (Phila Pa 1976) 2006;31:775-81. 56. Medtronic Inc. [Bone dowel pilot] [Protocol] [Received from Medtronic 1st December 2011]. 57. Medtronic Inc. A clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with allograft bone dowel for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2004. 58. A pilot study of rhBMP-2/ACS/allograft bone dowel for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease. [cited Available from: http://clinicaltrials.gov/ct2/show?term=medtronic+fusion&rank=11. 59. Burkus J, Gornet M, Longley M. A large randomized clinical evaluation of rhBMP-2 versus iliac crest bone graft combined with cortical allograft in lumbar spine fusion. Spine J 2004;4:S35-6. 60. Medtronic Inc. [Bone dowel pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2000. 61. Pivotal study of rhBMP-2/ACS/allograft bone dowel for anterior lumbar interbody fusion in patients with symptomatic degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01494493?term=medtronic+fusion&rank=16. 62. Alexander JT, Branch CL. Recombinant human bone morphogenetic protein 2 in a posterior lumbar interbody fusion construct: 2-year clinical and radiological outcomes. Neurosurgery 2002;51:584. 63. Haid R, Branch C, Alexander J, Burkus J. Posterior lumbar interbody fusion using recombinant human bone morphogenetic protein type 2 with cylindrical interbody cages. Spine J 2004;4:527-38. 64. Medtronic Inc. [Interfix PLIF] [Protocol] [Received from Medtronic 1st December 2011]. 65. Medtronic Inc. A prospective, randomized clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the INTER FIX™ device for posterior lumbar interbody fusion in patients with degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2002. 66. A pivotal study of rhBMP-2/ACS/INTER FIX™ device for posterior lumbar interbody fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491464?term=medtronic+fusion&rank=13. 67. Baskin DS, Ryan P, Sonntag V, Westmark R, Widmayer M. A prospective, randomized, controlled cervical fusion study using recombinant human bone morphogenetic protein-2 with the CORNERSTONE-SR allograft ring and the ATLANTIS anterior cervical plate. Spine (Phila Pa 1976) 2003;28:1219-24. 68. Medtronic Inc. [Cornerstone ACDF Pilot] [Protocol] [Received from Medtronic 1st December 2011]. 69. Medtronic Inc. A pilot prospective, randomized, clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the CORNERSTONE-SR™aIIograft ring and the ATLANTISTM anterior cervical plate system for the treatment of patients with degenerative cervical disc disease [Final report] [Received from Medtronic 1st December 2011]; 2003. 70. INFUSE™ BONE GRAFT/CORNERSTONE-SR™ Allograft Ring/ATLANTIS™ anterior cervical plate- pilot study. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491399?term=medtronic+fusion&rank=26. 71. Bae HW, Stambough J, Glassman SD, Burkus JK. Level-1 data comparing rhBMP-2/ACS combined with an osteoconductive bulking agent with iliac crest bone graft in posterolateral lumbar fusion. Spine J 2007;7:9S.
122
72. Dawson E, Bae H, Burkus J, Stambough J, Glassman S. Recombinant human bone morphogenetic protein-2 on an absorbable collagen sponge with an osteoconductive bulking agent in posterolateral arthrodesis with instrumentation. A prospective randomized trial. J Bone Joint Surg Am 2009;91:1604-13. 73. Medtronic Inc. [Mastergraft pilot] [Protocol] [Received from Medtronic 1st December 2011]; 2002. 74. Medtronic Inc. A pilot prospective, randomized clinical investigation of INFUSE® bone graft and MasterGraft™ granules with the CD HORIZON® spinal system for posterolateral lumbar fusion in patients with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2007. 75. INFUSE® bone graft and MASTERGRAFT® granules with CD HORIZON® for posterolateral lumbar fusion in patients with degenerative disc disease - pilot study. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491542?term=medtronic+fusion&rank=17. 76. Medtronic Inc. [Interfix ALIF pilot] [Protocol] [Received from Medtronic 1st December 2011]. 77. Medtronic Inc. A pilot prospective, randomized, clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the INTERFIX™ device for the anterior lumbar interbody fusion in patients with degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2003. 78. A pilot study of rhBMP-2/ACS with the INTERFIX™ device for the anterior lumbar interbody fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491451?term=medtronic+fusion&rank=10. 79. Gornet M, Burkus J, Matthews H. Maverick total disk replacement versus anterior interbody fusion with the INFUSE Bone Graft/LT-CAGE device: a prospective, randomised, controlled, multicenter IDE trial. [Abstract]. Spine J 2007;7:1S. 80. Gornet MF, Burkus JK, Dryer RF, Peloza JH. Lumbar disc arthroplasty with MAVERICK disc versus stand-alone interbody fusion: a prospective, randomized, controlled, multicenter investigational device exemption trial. Spine (Phila Pa 1976) 2011;36:E1600-E11. 81. Medtronic Inc. [Maverick disc pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2009. 82. Pilot study of rhBMP-2/BCP in patients with spinal degeneration with instability requiring surgical fusion. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01495234?term=medtronic+fusion&rank=6. 83. MAVERICK™ total disc replacement - pivotal study. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT00635843?term=medtronic+fusion&rank=12. 84. Medtronic Inc. [Telamon] [Protocol] [Received from Medtronic 1st December 2011]; 2005. 85. Medtronic Inc. A pilot, prospective, non-randomized clinical investigation of TELAMON PTM implant with INFUSE® bone graft and the CD HORIZON® spinal system for posterior lumbar interbody fusion in patients with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2008. 86. TELAMON P™ Implant/INFUSE® Bone Graft/CD HORIZON® spinal system pilot study. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491516?term=medtronic+fusion&rank=22. 87. Boden SD, Kang J, Sandhu H, Heller J. Use of recombinant human bone morphogenetic protein-2 to achieve posterolateral lumbar spine fusion in humans: a prospective, randomized clinical pilot trial: 2002 Volvo Award in clinical studies. Spine (Phila Pa 1976) 2002;27:2662-73. 88. Medtronic Inc. [BMP BCP US] [Protocol] [Received from Medtronic 1st December 2011]. 89. Medtronic Inc. A pilot prospective, randomized clinical investigation of recombinant human bone morphogenetic protein-2 and biphasic calcium phosphate with or without the TSRH® spinal system for posterolateral lumbar fusion in patients with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2003. 90. Pilot study of rhBMP/BCP with or without the TSRH® spinal system for posterolateral lumbar fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01494441?term=medtronic+fusion&rank=8. 91. Assiri I, du PS, Hurlbert J, Hu R, Salo P, Whittaker T. A prospective randomized clinical study comparing instrumented lumbar fusion rates of Recombinant Human Bone Morphogenic Protein-2 (rhBMP-2) with autogenous iliac crest bone graft in patients with symptomatic degenerative disc disease. Can J Surg 2004;47:7-8. 92. Alexander D, Bailey S, Hurlbert RJ, Abraham E, McBroom R, Mahood J, et al. A long-term radiographic and clinical evaluation of a new rhBMP-2 formulation in a prospective randomized lumbar posterolateral spine fusion study. Neurosurgery 2007;61:200-01.
123
93. Abraham E, Alexander D, Bailey S, Hurlbert J, McBroom R, Mahood J, et al. A long-term radiographic and clinical evaluation of a new rhBMP-2 formulation in a prospective randomized lumbar posterolateral spine fusion study. Can J Surg 2008;51:S6-S7. 94. Abraham EP, Hurlbert J, Alexander D, Bailey S, Fisher C. Evaluation of an rhBMP-2 formulation in 2-level posterolateral lumbar spine arthrodesis. Spine J 2010;10:105S. 95. Medtronic Inc. [BMP BCP Canada] [Protocol] [Received from Medtronic 1st December 2011]; 2002. 96. Medtronic Inc. A prospective, randomized clinical investigation of recombinant human bone morphogenetic protein-2 and biphasic calcium phosphate with either the TSRH® spinal system or the CD horizon® spinal system for posterolateral lumbar fusion in patients with degenerative disc disease. Canada [Final report] [Received from Medtronic 1st December 2011]; 2009. 97. Clinical study of rhBMP-2/BCP with the TSRH® spinal system for posterolateral lumbar fusion in patients with degenerative disc disease. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01494454?term=medtronic+fusion&rank=14. 98. Glassman SD, Dimar JR, Carreon LY, Campbell MJ, Puno RM, Johnson JR. Initial fusion rates with recombinant human bone morphogenetic protein-2/compression resistant matrix and a hydroxyapatite and tricalcium phosphate/collagen carrier in posterolateral spinal fusion. Spine (Phila Pa 1976) 2005;30:1694-8. 99. Dimar JR, Glassman SD, Burkus KJ, Carreon LY. Clinical outcomes and fusion success at 2 years of single-level instrumented posterolateral fusions with recombinant human bone morphogenetic protein-2/compression resistant matrix versus iliac crest bone graft. Spine (Phila Pa 1976) 2006;31:2534-9. 100. Dimar JR, Glassman SD, Burkus JK, Pryor PW, Hardacker JW, Boden SD. Evaluation of an recombinant human bone morphogenetic protein-2/ceramic matrix as an iliac crest bone graft replacement in posterolateral fusion: a multicenter, level I clinical study. Neurosurgery 2006;59:459. 101. Glassman SD, Dimar J, Burkus K, Hardacker J, Pryor P, Boden S, et al. The efficacy of rhBMP-2 for posterolateral lumbar fusion in smokers. Spine (Phila Pa 1976) 2007;32:1693-8. 102. Dimar J, Glassman S, Burkus J, Pryor P, Hardacker J, Carreon L. Clinical and radiographic analysis of an optimized rhBMP-2 formulation as an autograft replacement in posterolateral lumbar spine arthrodesis. J Bone Joint Surg Am 2009;91:1377-86. 103. McInnis MM, Olchanski N, Kemner JE, Goss T. Budget impact of new rhbmp-2 formulation in patients undergoing posterolateral spinal fusion procedures for degenerative disc disease in randomized controlled trial (RCT). Value Health 2010;13:A305. 104. Medtronic Inc. [BMP-CRM pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2006. 105. Medtronic Inc. A prospective randomized clinical investigation of recombinant human bone morphogenetic protein-2 and compression resistant matrix with the CD HORIZON® spinal system for posterolateral lumbar fusion in pateints with symptomatic degenerative disc disease [Final report] [Received from Medtronic 1st December 2011]; 2011. 106. rhBMP-2/CRM/CD HORIZON® spinal system pivotal study. [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT00707265?term=medtronic+fusion&rank=21 107. Medtronic Inc. [BMP-CRM 2-level] [Protocol] [Received from Medtronic 1st December 2011]; 2005. 108. Medtronic Inc. A non-randomized pilot study of recombinant human bone morphogenetic protein- 2 and compression resistant matrix with the CD HORIZON® spinal system for posterolateral lumbar fusion in patients with symptomatic degenerative disc disease at two adjacent vertebral levels [Final report] [Received from Medtronic 1st December 2011]; 2010. 109. rhBMP-2/CRM/CD HORIZON® spinal system pilot study (2-level). [cited 2012 10th July]. Available from: http://clinicaltrials.gov/ct2/show/NCT01491568?term=medtronic+fusion&rank=25. 110. Medtronic Inc. [BCP Mexico pilot] [Protocol] [Received from Medtronic 1st December 2011]. 111. Medtronic Inc. A pilot investigation of recombinant human bone morphgenetic protein-2/biphasic calcium phosphate (rhBMP-2/BCP) in patients with spinal degeneration with instability requiring surgical fusion [Final report] [Received from Medtronic 1st December 2011]; 2001. 112. Medtronic Inc. [Cornerstone pivotal] [Protocol] [Received from Medtronic 1st December 2011]; 2002. 113. Medtronic Inc. A pivotal prospective, randomized, clinical investigation of recombinant human bone morphogenetic protein-2 and absorbable collagen sponge with the Cornerstone-SR™ allograft ring and the ATLANTISTM anterior cervical plate system for the treatment of patients with degenerative cervical disc disease [Final report] [Received from Medtronic 1st December 2011]; 2005.
124
114. INFUSE ® Bone Graft/CORNERSTONE-SR® Allograft Ring/ATLANTIS® Anterior Cervical Plate System Pivotal Trial. [cited 2012 12th July]. Available from: http://clinicaltrials.gov/ct2/show?term=medtronic+fusion&rank=33 115. Williams BJ, Smith JS, Fu K-MG, Hamilton DK, Polly DW, Jr., Ames CP, et al. Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? A comparison of 55,862 cases of spinal fusion with and without bone morphogenetic protein. Spine (Phila Pa 1976) 2011;36:1685-91. 116. Anand N, Hamilton JF, Perri B, Miraliakbar H, Goldstein T. Cantilever TLIF with structural allograft and RhBMP2 for correction and maintenance of segmental sagittal lordosis: long-term clinical, radiographic, and functional outcome. Spine (Phila Pa 1976) 2006;31:E748-53. 117. Glassman SD, Carreon L, Djurasovic M, Campbell MJ, Puno RM, Johnson JR, et al. Posterolateral lumbar spine fusion with INFUSE bone graft. Spine J 2007;7:44-9. 118. Gray RJ, Rampersaud YR. Comparison of the incidence of radiculitis and radiographic adverse event following minimally invasive lumbar transforaminal interbody fusions (TLIF) with and without the use of bone morphogenetic protein (BMP). Spine J 2010;10:8S-9S. 119. Joseph V, Rampersaud YR. Heterotopic bone formation with the use of rhBMP2 in posterior minimal access interbody fusion: a CT analysis. Spine (Phila Pa 1976) 2007;32:2885-90. 120. Lee K-B, Taghavi CE, Hsu MS, Song K-J, Yoo JH, Keorochana G, et al. The efficacy of rhBMP-2 versus autograft for posterolateral lumbar spine fusion in elderly patients. Eur Spine J 2010;19:924-30. 121. Mannion RJ, Nowitzke AM, Wood MJ. Improved fusion in minimally invasive lumbar interbody stabilisation with low dose BMP-2, but at what cost? Br J Neurosurg 2010;24:120. 122. Mindea SA, Shih P, Song JK. Recombinant human bone morphogenetic protein-2-induced radiculitis in elective minimally invasive transforaminal lumbar interbody fusions: a series review. Spine (Phila Pa 1976) 2009;34:1480-4. 123. Mummaneni PV, Pan J, Haid RW, Rodts GE. Contribution of recombinant human bone morphogenetic protein-2 to the rapid creation of interbody fusion when used in transforaminal lumbar interbody fusion: a preliminary report. Invited submission from the Joint Section Meeting on Disorders of the Spine and Peripheral Nerves, March 2004. J Neurosurg Spine 2004;1:19-23. 124. Pimenta L, Marchi L, Oliveira L, Coutinho E. A prospective, randomized, controlled clinical and radiological study to evaluate and compare the use of silicated calcium phosphate and rh-BMP2 in interbody lumbar spine fusion: 36 month follow-up. Spine J 2011;11:130S. 125. Rihn JA, Patel R, Makda J, Hong J, Anderson DG, Vaccaro AR, et al. Complications associated with single-level transforaminal lumbar interbody fusion. Spine J 2009;9:623-9. 126. Rowan FE, O'Malley N, Poynton A. Recombinant human bone morphogenic protein is safe to use in postolateral spinal fusion surgery and is associated with better leg pain scores at late follow up. J Bone Joint Surg Br 2011;93 B:490-91. 127. Singh K, Smucker JD, Gill S, Boden SD. Use of recombinant human bone morphogenetic protein-2 as an adjunct in posterolateral lumbar spine fusion: a prospective CT-scan analysis at one and two years. J Spinal Disord Tech 2006;19:416-23. 128. Taghavi CE, Lee K-B, Keorochana G, Tzeng S-T, Yoo JH, Wang JC. Bone morphogenetic protein-2 and bone marrow aspirate with allograft as alternatives to autograft in instrumented revision posterolateral lumbar spinal fusion: a minimum two-year follow-up study. Spine (Phila Pa 1976) 2010;35:1144-50. 129. Comer GC, Smith MW, Hurwitz E, Mitsunaga KA, Kessler R, Carragee EJ. Retrograde ejaculation following anterior lumbar inter-body fusion with and without Bone Morphogenetic Protein-2 augmentation: a ten-year cohort controlled study. Spine J 2012;10:881-90. 130. Gerszten PC, Tobler WD, Nasca RJ. Retrospective analysis of L5-S1 axial lumbar interbody fusion (AxiaLIF): a comparison with and without the use of recombinant human bone morphogenetic protein-2. Spine J 2011;11:1027-32. 131. McConnell J. A comparison of b-tcpdbmaversus rhBMP-2 in anterior lumbar interbody fusion: a prospective, randomized trial with two-year clinical and radiographic outcomes. Spine J 2011;11:64S- 65S. 132. Pradhan BB, Bae HW, Dawson EG, Patel VV, Delamarter RB. Graft resorption with the use of bone morphogenetic protein: lessons from anterior lumbar interbody fusion using femoral ring allografts and recombinant human bone morphogenetic protein-2. Spine (Phila Pa 1976) 2006;31:E277-84. 133. Slosar PJ, Josey R, Reynolds J. Accelerating lumbar fusions by combining rhBMP-2 with allograft bone: a prospective analysis of interbody fusion rates and clinical outcomes. Spine J 2007;7:301-7.
125
134. Carragee EJ, Mitsunaga KA, Hurwitz EL, Scuderi GJ. Retrograde ejaculation after anterior lumbar interbody fusion using rhBMP-2: a cohort controlled study. Spine J 2011;11:511-16. 135. Carragee E, Wildstein M. A controlled trial of BMP and inilateral transpedicular instrumentation in circumferential single or double level lumbar fusion. Spine J 2007;7:8S-9S. 136. Carragee E, Wildstein M. Subsidence and osteolysis in patients undergoing alif with and without rhBMP-2 graft aumentation. Spine J 2011;11:114S. 137. Buttermann GR. Prospective nonrandomized comparison of an allograft with bone morphogenic protein versus an iliac-crest autograft in anterior cervical discectomy and fusion. Spine J 2008;8:426- 35. 138. Crawford CH, 3rd, Carreon LY, McGinnis MD, Campbell MJ, Glassman SD. Perioperative complications of recombinant human bone morphogenetic protein-2 on an absorbable collagen sponge versus iliac crest bone graft for posterior cervical arthrodesis. Spine (Phila Pa 1976) 2009;34:1390-4. 139. Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use recombinant human bone morphogenetic protein in posterior cervical fusion? Spine (Phila Pa 1976) 2009;34:885-9. 140. Longley MC, Woodward HR, Phillips ED, Fuller JE, McClellan JW, Burd TA, et al. Does the use of bone morphogenetic protein (rhBMP-2) in patients undergoing Anterior Cervical Decompression and Fusion (ACDF) increase their risk of adverse events?: a retrospective case-control study. In: 37th Annual Meeting of the Cervical Spine Research Society, CSRS. Salt Lake City, Utah, United States: Spine (Phila Pa 1976); 2009. 141. Lu DC, Chou D, Rodts G, Mummaneni PV. Multilevel ACDF with and without BMP: a comparison of outcomes and dysphagia rates in 150 patients. J Neurosurg 2010;113:A406. 142. Smucker JD, Rhee JM, Singh K, Yoon ST, Heller JG. Increased swelling complications associated with off-label usage of rhBMP-2 in the anterior cervical spine. Spine (Phila Pa 1976) 2006;31:2813-9. 143. Vaidya R, Carp J, Sethi A, Bartol S, Craig J, Les CM. Complications of anterior cervical discectomy and fusion using recombinant human bone morphogenetic protein-2. Eur Spine J 2007;16:1257-65. 144. Xu R, Bydon M, Sciubba DM, Witham TF, Wolinsky J-P, Gokaslan ZL, et al. Safety and efficacy of rhBMP2 in posterior cervical spinal fusion for subaxial degenerative spine disease: analysis of outcomes in 204 patients. Surg Neurol Int 2011;2:109. 145. Yaremchuk KL, Toma MS, Somers ML, Peterson E. Acute airway obstruction in cervical spinal procedures with bone morphogenetic proteins. Laryngoscope 2010;120:1954-7. 146. Cahill KS, Chi JH, Groff MW, McGuire K, Afendulis CC, Claus EB. Outcomes for single-level lumbar fusion: the role of bone morphogenetic protein. Spine (Phila Pa 1976) 2011. 147. Deyo RA, Ching A, Matsen L, Martin BI, Kreuter W, Jarvik JG, et al. Use of bone morphogenetic proteins in spinal fusion surgery for older adults with lumbar stenosis: trends, complications, repeat surgery, and charges. Spine (Phila Pa 1976) 2012;37:222-30. 148. Latzman JM, Kong L, Liu C, Samadani U. Administration of human recombinant bone morphogenetic protein-2 for spine fusion may be associated with transient postoperative renal insufficiency. Spine (Phila Pa 1976) 2010;35:E231-7. 149. Maeda T, Buchowski JM, Kim YJ, Mishiro T, Bridwell KH. Long adult spinal deformity fusion to the sacrum using rhBMP-2 versus autogenous iliac crest bone graft. Spine (Phila Pa 1976) 2009;34:2205-12. 150. Mines D, Gu Y, Kou TD, Cooper GS. Recombinant human bone morphogenetic protein-2 and pancreatic cancer: a retrospective cohort study. Pharmacoepidemiol Drug Saf 2011;20:111-8. 151. Vaidya R, Weir R, Sethi A, Meisterling S, Hakeos W, Wybo CD. Interbody fusion with allograft and rhBMP-2 leads to consistent fusion but early subsidence. J Bone Joint Surg Br 2007;89:342-5. 152. Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]: The Cochrane Collaboration, 2011. Available from: www.cochrane-handbook.org 153. Centre for Reviews and Dissemination. Systematic reviews: CRD's guidance for undertaking reviews in health care. In. York: University of York, Centre for Reviews & Dissemination; 2009. Available from: http://www.york.ac.uk/inst/crd/SysRev/!SSL!/WebHelp/SysRev3.htm 154. Vandenbroucke J, Psaty B. Benefits and risks of drug treatments:how to combine the best evidence on benefits with the best data about adverse effects. JAMA 2008;300:2417-9. 155. Glassman SD, Gum JL, Crawford CH, 3rd, Shields CB, Carreon LY. Complications with recombinant human bone morphogenetic protein-2 in posterolateral spine fusion associated with a dural tear. Spine J 2011;11:522-6.
126
156. Taylor S, Taylor A, Foy M, Fogg A. Responsiveness of common outcome measures for patients with low back pain. Spine (Phila Pa 1976) 1999;24:1805-12. 157. Fairbank J, Frost H, Wilson-MacDonald J, Yu L-M, Barker K, Collins R, et al. Randomised controlled trial to compare surgical stabilisation of the lumbar spine with an intensive rehabilitation programme for patients with chronic low back pain: the MRC spine stabilisation trial. BMJ 2005;330:1233. 158. DeVine JG, Dettori JR, France JC, Brodt E, McGuire RA. The use of rhBMP in spine surgery: is there a cancer risk? Evidence-Based Spine-Care Journal 2012;3:35-41. 159. de Vries T, van Roon E. Low quality of reporting adverse drug reactions in paediatric randomised controlled trials. Arch Dis Child 2010;95:1023-6. 160. Pitrou I, Boutron I, Ahmad N, Ravaud P. Reporting of safety results in published reports of randomized controlled trials. Arch Intern Med 2009;169:1756-61. 161. Sinha S, Sinha S, Ashby E, Jayaram R, Grocott M. Quality of reporting in randomized trials published in high-quality surgical journals. J Am Coll Surg 2009;209:565-71. 162. Breau R, Gaboury I, Scales CJ, Fesperman S, Watterson J, Dahm P. Reporting of harm in randomized controlled trials published in the urological literature. J Urol 2010;183:1693-7. 163. Stewart LA, Tierney JF. To IPD or not to IPD? Advantages and disadvantages of systematic reviews using individual patient data. Eval Health Prof 2002;25:76-97. 164. Ackerman SJ, Mafilios MS, Polly DW, Jr. Economic evaluation of bone morphogenetic protein versus autogenous iliac crest bone graft in single-level anterior lumbar fusion: an evidence-based modeling approach. Spine (Phila Pa 1976) 2002;27:S94-9. 165. Carreon LY, Glassman SD, Djurasovic M, Campbell MJ, Puno RM, Johnson JR, et al. RhBMP-2 versus iliac crest bone graft for lumbar spine fusion in patients over 60 years of age: a cost-utility study. Spine (Phila Pa 1976) 2009;34:238-43. 166. Glassman SD, Carreon LY, Campbell MJ, Johnson JR, Puno RM, Djurasovic M, et al. The perioperative cost of Infuse bone graft in posterolateral lumbar spine fusion. Spine J 2008;8:443-8. 167. Polly DW, Jr., Ackerman SJ, Shaffrey CI, Ogilvie JW, Wang JC, Stralka SW, et al. A cost analysis of bone morphogenetic protein versus autogenous iliac crest bone graft in single-level anterior lumbar fusion. Orthopedics 2003;26:1027-37. 168. Abdullah KG, Steinmetz MP, Benzel EC, Mroz TE. The state of lumbar fusion extenders. Spine (Phila Pa 1976) 2011;36:E1328-34. 169. Agarwal R, Williams K, Umscheid CA, Welch WC. Osteoinductive bone graft substitutes for lumbar fusion: a systematic review. J Neurosurg Spine 2009;11:729-40. 170. Burkus JK, Heim SE, Gornet MF, Zdeblick TA. The effectiveness of rhBMP-2 in replacing autograft: an integrated analysis of three human spine studies. Orthopedics 2004;27:723-8. 171. Carragee E, Weiner B, Hurwitz E. Comparison of adverse events and disclosures in the original rhBMP-2 trials with fda data and subsequent publications. Spine J 2011;11:21S-22S. 172. ECRI. Interbody cage with bone morphogenetic protein (InFUSE(TM)/LT-CAGE(TM) for degenerative disc disease. Plymouth Meeting, PA: ECRI; 2004. 173. Fenton JJ, Mirza SK, Lahad A, Stern BD, Deyo RA. Variation in reported safety of lumbar interbody fusion - Influence of industrial sponsorship and other study characteristics. Spine (Phila Pa 1976) 2007;32:471-80. 174. Garrison KR, Donell S, Ryder J, Shemilt I, Mugford M, Harvey I, et al. Clinical effectiveness and cost-effectiveness of bone morphogenetic proteins in the non-healing of fractures and spinal fusion: a systematic review. Health Technol Assess 2007;11:1-150. 175. Guerado E, Fuerstenberg CH. What bone graft substitutes should we use in post-traumatic spinal fusion? Injury 2011;42 Suppl 2:S64-71. 176. Hayes Inc. Recombinant human bone morphogenetic protein for use in spinal fusion. Lansdale, PA: HAYES Inc; 2006. 177. Hsu WK, Wang JC. The use of bone morphogenetic protein in spine fusion. Spine J 2008;8:419- 25. 178. Kanakaris NK, Giannoudis PV. Clinical applications of bone morphogenetic proteins: current evidence. J Surg Orthop Adv 2008;17:133-46. 179. McKay WF, Peckham SM, Badura JM. A comprehensive clinical review of recombinant human bone morphogenetic protein-2 (INFUSE Bone Graft). Int Orthop 2007;31:729-34. 180. Menzin J, Nichols C, Mwamburi M, Federico V, Kemner JE. A systematic review of the use of patient-centered outcome measures in studies of spinal fusion surgery for degenerative disc disease: implications for comparative effectiveness research. Value Health 2012;15:A79. 181. Miyazaki M, Tsumura H, Wang JC, Alanay A. An update on bone substitutes for spinal fusion. Eur Spine J 2009;18:783-99.
127
182. Mroz TE, Wang JC, Hashimoto R, Norvell DC. Complications related to osteobiologics use in spine surgery: a systematic review. Spine (Phila Pa 1976) 2010;35:S86-104. 183. Mussano F, Ciccone G, Ceccarelli M, Baldi I, Bassi F. Bone morphogenetic proteins and bone defects: a systematic review. Spine (Phila Pa 1976) 2007;32:824-30. 184. Papakostidis C, Kontakis G, Bhandari M, Giannoudis PV. Efficacy of autologous iliac crest bone graft and bone morphogenetic proteins for posterolateral fusion of lumbar spine: a meta-analysis of the results. Spine (Phila Pa 1976) 2008;33:E680-92. 185. Poynton AR, Lane JM. Safety profile for the clinical use of bone morphogenetic proteins in the spine. Spine (Phila Pa 1976) 2002;27:S40-8. 186. Smoljanovic T, Grgurevic L, Jelic M, Kreszinger M, Haspl M, Maticic D, et al. Regeneration of the skeleton by recombinant human bone morphogenetic proteins. Coll Antropol 2007;31:923-32. 187. Than KD, Wang AC, Rahman SU, Wilson TJ, Valdivia JM, Park P, et al. Complication avoidance and management in anterior lumbar interbody fusion. Neurosurg Focus 2011;31:E6. 188. Thawani JP, Wang AC, Than KD, Lin C-Y, La Marca F, Park P. Bone morphogenetic proteins and cancer: review of the literature. Neurosurgery 2010;66:233-46. 189. Vaz K, Verma K, Protopsaltis T, Schwab F, Lonner B, Errico T. Bone grafting options for lumbar spine surgery: a review examining clinical efficacy and complications. SAS Journal 2010;4:75-86. 190. Wong E. Is there a role for BMPs in spinal fusion surgery? Bone 2009;44:S159. 191. Wu RH, Fraser JF, Hartl R. Minimal access versus open transforaminal lumbar interbody fusion: meta-analysis of fusion rates. Spine (Phila Pa 1976) 2010;35:2273-81. 192. Adogwa O, Parker SL, Shau D, Mendelhall SK, Cheng J, Aaronson O, et al. Long-term outcomes of revision fusion for lumbar pseudarthrosis: clinical article. J Neurosurg Spine 2011;15:393-8. 193. Anderson DG, Andersson GBJ, Boden SD, Damien C, Ebara S, Helm G, et al. Summary statement: Clinical BMP programs. Spine (Phila Pa 1976) 2002;27:S49. 194. A critical eye on infuse use studies. OR Manager 2011;27:5. 195. BMP use up sharply, but outcomes not improved. OR Manager 2012;28:5. 196. Aspenberg P. Under-reported complications related to BMP use in spine surgery. Acta Orthop 2011;82:511-2. 197. Auerbach JD, Ong KL, Lau E, Ochoa J, Schmier JK, Zigler JD. Perioperative outcomes, complications, and costs associated with lumbar spinal fusion in older patients with spinal stenosis and spondylolisthesis: analysis of the United States Medicare claims database. Value Health 2012;15:A3. 198. Axelrad TW, Steen B, Lowenberg DW, Creevy WR, Einhorn TA. Heterotopic ossification after the use of commercially available recombinant human bone morphogenetic proteins in four patients. J Bone Joint Surg Br 2008;90B:1617-22. 199. Axelrad TW, Einhorn TA. Bone morphogenetic proteins in orthopaedic surgery. Cytokine Growth Factor Rev 2009;20:481-8. 200. Benglis D, Wang MY, Levi AD. A comprehensive review of the safety profile of bone morphogenetic protein in spine surgery. Neurosurgery 2008;62:ONS423-31. 201. Beringer W, Mobasser J. Unilateral pedicle screw instrumentation for minimally invasive transforaminal lumbar interbody fusion. Neurosurgery Focus 2006;20:E4. 202. Boden SD, Andersson GBJ, Anderson DG, Damien C, Ebara S, Helm G, et al. Summary statement: overview of bone morphogenetic proteins for spine fusion. Spine (Phila Pa 1976) 2002;27:S1. 203. Burks MV, Nair L. Long-term effects of bone morphogenetic protein-based treatments in humans. J Long Term Eff Med Implants 2010;20:277-93. 204. Burkus JK, Schuler TC, Gornet MF, Zdeblick TA. Anterior lumbar interbody fusion for the management of chronic lower back pain: current strategies and concepts. Orthop Clin North Am 2004;35:25-32. 205. Burkus JK. Clinical outcomes using rhBMP-2 in spinal fusion applications. In: Progress in Inflammation Research. Basel, Switzerland: Birkhauser Verlag AG; 2008. p. 71-80. 206. Cabraja M, Kroppenstedt S. Bone grafting and substitutes in spine surgery. J Neurosurg Sci 2012;56:87-95. 207. Carragee EJ, Bono CM, Scuderi GJ. Pseudomorbidity in iliac crest bone graft harvesting: the rise of rhBMP-2 in short-segment posterior lumbar fusion. Spine J 2009;9:873-9. 208. Chi JH. Exposing conflicts of interest and complications of rhBMP-2. Neurosurgery 2011;69:N21- N22.
128
209. Cho SK, Stoker GE, Bridwell KH. Spinal reconstruction with pedicle screw-based instrumentation and rhBMP-2 in patients with neurofibromatosis and severe dural ectasia and spinal deformity: report of two cases and a review of the literature. J Bone Joint Surg Am 2011;93:e86. 210. Chrastil J, Patel AA. Complications associated with posterior and transforaminal lumbar interbody fusion. J Am Acad Orthop Surg 2012;20:283-91. 211. Crawford CH, Buchowski JM. Bone morphogenetic protein use in anterior cervical spine surgery: a review of current literature concerning indications, safety, and efficacy. Neurosurg Q 2009;19:283- 87. 212. Daentzer D. [The efficacy of rhBMP-2 for posterolateral lumbar fusion in smokers: Comment.]. Z Orthop Unfall 2007;145:552. 213. Daniels AH, Riew KD, Yoo JU, Ching A, Birchard KR, Kranenburg AJ, et al. Adverse events associated with anterior cervical spine surgery. J Am Acad Orthop Surg 2008;16:729-38. 214. David S, Murakami T, Tabor O. Lumbar spinal fusion using rhBMP2: a randomised, blinded and controlled study. Trans Int Soc Study Lumbar Spine 1995;22:14. 215. Deutsch H, Musacchio M. Minimally invasive transforaminal lumbar interbody fusion with unilateral pedicle screw fixation. Neurosurg Focus 2006;20:E10. 216. Epstein NE. Pros, cons, and costs of INFUSE in spinal surgery. Surg Neurol Int 2011;2:10. 217. Epstein NE, Schwall GS. Costs and frequency of "off-label" use of INFUSE for spinal fusions at one institution in 2010. Surg Neurol Int 2011;2:115. 218. Glassman SD, Howard JM, Sweet A, Carreon LY. Complications and concerns with osteobiologics for spine fusion in clinical practice. Spine (Phila Pa 1976) 2010;35:1621-8. 219. Hart RA. Acknowledging the elephant in the room: conflict of interest in industry-sponsored clinical research. Spine J 2011;11:703-04. 220. Harwood PJ, Giannoudis PV. Application of bone morphogenetic proteins in orthopaedic practice: their efficacy and side effects. Expert Opin Drug Saf 2005;4:75-89. 221. Jarrett C, Heller JG, Tsai L. Anterior exposure of the lumbar spine with and without an "access surgeon": morbidity analysis of 265 consecutive cases. J Spinal Disord Tech 2009;22:559-64. 222. Jeong GK, Sandhu HS, Farmer J. Bone morphogenic proteins: applications in spinal surgery. HSS J 2005;1:110-7. 223. Kang B, Choi WC, Lee SH, Jeon SH, Park JD, Maeng DH, et al. An analysis of general surgery- related complications in a series of 412 minilaparotomic anterior lumbosacral procedures. J Neurosurg Spine 2009;10:60-5. 224. Kim M, Choe S. BMPs and their clinical potentials. BMB reports 2011;44:619-34. 225. Kraiwattanapong C, Boden S, Louis-Ugbo J, Attallha E, Barnes B, Hutton W. Comparison of Healos/bone marrow to INFUSE (rhBMP-2/ACS) with a collagen-ceramic sponge bulking agent as graft substitutes for lumbar spine fusion. Spine (Phila Pa 1976) 2005;30:1001-7. 226. Lad SP, Nathan JK, Boakye M. Trends in the use of bone morphogenetic protein as a substitute to autologous iliac crest bone grafting for spinal fusion procedures in the United States. Spine (Phila Pa 1976) 2011;36:E274-81. 227. Lissenberg-Thunnissen SN, de Gorter DJJ, Sier CFM, Schipper IB. Use and efficacy of bone morphogenetic proteins in fracture healing. Int Orthop 2011;35:1271-80. 228. McKay WF, Peckham SM, Badura JM. Development of a novel compression-resistant carrier for recombinant human bone morphogenetic protein-2 (rhBMP-2) and preliminary clinical results. In: Progress in Inflammation Research. Basel, Switzerland: Birkhauser Verlag AG; 2008. p. 7-23. 229. McKay B. Local sustained delivery of recombinant human bone morphogenetic protein-2 (rhBMP-2). Conference Proceedings: Annual International Conference of the IEEE Engineering in Medicine & Biology Society 2009:236-7. 230. Mehta S, Goss B, Williams R. Retrospective comparative analysis of degenerative lumbar spondylolisthesis treated with posterolateral fusion (PLF) with bone morphogenic protein (BMP) 2 with or without additional transforaminal lumbar interbody fusion (TLIF). In: 38th Annual Meeting of the Cervical Spine Research Society, CSRS. Charlotte, NC United States: Spine (Phila Pa 1976); 2010. 231. Mirza SK. Commentary: folly of FDA-approval studies for bone morphogenetic protein. Spine J 2011;11:495-99. 232. Niu CC, Liao JC, Chen WJ, Chen LH. Outcomes of interbody fusion cages used in 1 and 2-levels anterior cervical discectomy and fusion titanium cages versus polyetheretherketone (PEEK) cages. J Spinal Disord Tech 2010;23:310-16. 233. Potter B, Freedman B, Verwiebe E. Transforaminal lumbar interbody fusion: clinical and radiographic results and complications in 100 consecutive patients. J Spinal Disord Tech 2005;18:337-46.
129
234. Rihn JA, Gates C, Glassman SD, Phillips FM, Schwender JD, Albert TJ. The use of bone morphogenetic protein in lumbar spine surgery. Instr Course Lect 2009;58:677-88. 235. Robinson Y, Heyde CE, Tschoke SK, Mont MA, Seyler TM, Ulrich SD. Evidence supporting the use of bone morphogenetic proteins for spinal fusion surgery. Expert Rev Med Devices 2008;5:75-84. 236. Rowan FE, O'Malley N, Poynton A. RhBMP-2 use in lumbar fusion surgery is associated with better leg pain scores at late follow up. Ir J Med Sci 2010;179:S11. 237. Salisbury EA, Olmsted-Davis EA, Davis AR. Adverse events and bone morphogenetic protein-2. Spine J 2011;11:802. 238. Sandhu HS, Anderson DG, Andersson GBJ, Boden SD, Damien C, Ebara S, et al. Summary statement: Safety of bone morphogenetic proteins for spine fusion. Spine (Phila Pa 1976) 2002;27:S39. 239. Sandhu HS, Khan SN. Recombinant human bone morphogenetic protein-2: use in spinal fusion applications. J Bone Joint Surg Am 2003;85-A Suppl 3:89-95. 240. Sass M. [Autologous bone or bone substitutes in spinal fusion]. Z Orthop Unfall 2009;147:411. 241. Schultz DG. FDA Public Health Notification: Life-threatening complications associated with recombinant human bone morphogenetic protein in cervical spine fusion. U.S. Food and Drug Administration; 2008. [cited 2012 23rd January]. Available from: http://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/PublicHealthNotifications/ucm062000.ht m. 242. Solanki G, Vijay S, Chakrapani A, Hendriksz C. Progressive spinal instability in MPS. Use of recombinant human bone morphogenetic protein-2 (RhBMP-2) to augment posterior spine fusion. J Inherit Metab Dis 2008;31:110. 243. Steinbrook R. Industry payments to physicians: lessons from orthopedic surgery. Arch Intern Med 2011;171:1765-66. 244. Szpalski M, Gunzburg R. Recombinant human bone morphogenetic protein-2: a novel osteoinductive alternative to autogenous bone graft? Acta Orthop Belg 2005;71:133-48. 245. Tanaka H, Nagai E, Murata H, Tsubone T, Shirakura Y, Sugiyama T, et al. Involvement of bone morphogenic protein-2 (BMP-2) in the pathological ossification process of the spinal ligament. Rheumatology 2001;40:1163-8. 246. U.S. Food and Drug Administration. InFuseTM Bone Graft/LT-CAGETM Lumbar Tapered Fusion Device - P000058. [cited 2012 23rd January]. Available from: http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/DeviceApprovalsandClearances/ Recently-ApprovedDevices/ucm083423.htm. 247. U.S. Food and Drug Administration. Summary of safety and effectiveness data. Available from: http://www.accessdata.fda.gov/cdrh_docs/pdf/P000058b.pdf 248. Weiner BK, Walker M. Efficacy of autologous growth factors in lumbar intertransverse fusions. Spine (Phila Pa 1976) 2003;28:1968-70. 249. Clinical decision making in spinal fusion for chronic low back pain. Results of a nationwide survey among spine surgeons. Willems P, de Bie R, Oner C, Castelein R, de Kleuver M. BMJ Open 2011 Dec 21;1(2): e000391. Print 2011. Spine J 2012;12:276. 250. World Health Organization. Recombinant human bone morphogenetic protein: life-threatening complications. WHO Drug Information 2008;22:195. 251. Yu B, Tian J, Jin A-M. [Clinical evaluation of bone morphogenetic protein in spinal fusion]. Chin J Clin Rehab 2003;7:2856-57. 252. Albert TJ. Clinical outcomes and fusion success at 2 years of single-level instrumented posterolateral fusions with recombinant human bone morphogenetic protein-2/compression resistant matrix versus iliac crest bone graft - Point of view. Spine (Phila Pa 1976) 2006;31:2540. 253. Benglis D, Wang MY, Levi AD. Update of comprehensive review of the safety profile of bone morphogenetic protein in spine surgery. Reply. Neurosurgery 2010;66:E1030. 254. Smoljanovic T, Bicanic G, Bojanic I. Update of comprehensive review of the safety profile of bone morphogenetic protein in spine surgery. Neurosurgery 2010;66:E1030. 255. Bennett M, Reynolds AS, Dickerman RD. Recent article by Shields et al titled "adverse effects associated with high-dose recombinant human bone morphogenetic protein-2 use in anterior cervical spine fusion". Spine (Phila Pa 1976) 2006;31:2029-30. 256. Shields LBE, Raque GH, Glassman SD, Campbell M, Vitaz T, Harpring J, et al. Adverse effects associated with high-dose recombinant human bone morphogenetic protein-2 use in anterior cervical spine fusion. Spine (Phila Pa 1976) 2006;31:542-7. 257. Benzel EC. Formation of painful seroma and edema after the use of recombinant human bone morphogenetic protein-2 in posterolateral lumbar spine fusions. Comments. Neurosurgery 2010;66:1049.
130
258. Garrett MP, Kakarla UK, Porter RW, Sonntag VKH. Formation of painful seroma and edema after the use of recombinant human bone morphogenetic protein-2 in posterolateral lumbar spine fusions. Neurosurgery 2010;66:1044-9. 259. Boden SD, Sandhu H, Zdeblick TA. Re: Boden SD, Zdeblick TA, Sandhu HS, et al. The use of rhBMP-2 in interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine 2000;25:376-81 Response. Spine (Phila Pa 1976) 2010;35:E1011. 260. Brower RS, Vickroy NM. Brower RS, Vickrov NM. A case of psoas ossification from the use of BMP-2 for posterolateral fusion at L4-L5. Spine 2008;33:E653-55. Response. Spine (Phila Pa 1976) 2009;34:749. 261. Dickerman RD, Reynolds AS, Bennett M. Brower RS, Vickrov NM. A case of psoas ossification from the use of BMP-2 for posterolateral fusion at L4-L5. Spine 2008;33:E653-55. Spine (Phila Pa 1976) 2009;34:749. 262. Buchowski JM, Riew KD, Nussenbaum B. In reference to acute airway obstruction in cervical spinal procedures with bone morphogenetic proteins. Laryngoscope 2011;121:2501. 263. Cahill KS, Claus EB. Complications associated with use of bone-morphogenetic proteins in spinal fusion procedures: reply. JAMA 2009;302:2091. 264. Smoljanovic T, Siric F, Bojanic I. Complications associated with use of bone-morphogenetic proteins in spinal fusion procedures. JAMA 2009;302:2090-1. 265. Woo EJ. Re: A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. Spine J 2011;11:804. 266. Cardoso MJ, Sciubba DM. Is use of bone-morphogenetic proteins for spine fusion surgery cost- effective? Arch Surg 2009;144:996-7. 267. Cimic M, Smoljanovic T, Bojanic I. Re: Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use recombinant human bone morphogenetic protein in posterior cervical fusion? Spine 2009;34:885- 9. Spine (Phila Pa 1976) 2010;35:361. 268. Corbin TP, Peul WC, Schoene ML, Kovacs FM. Re: Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? Spine 2011;36:1658-91. Spine (Phila Pa 1976) 2012;37:258. 269. Dickerman RD, Reynolds AS, Morgan BC, Tompkins J, Cattorini J, Bennett M. rh-BMP-2 can be used safely in the cervical spine: dose and containment are the keys! Spine J 2007;7:508-9. 270. Perri B, Cooper M, Lauryssen C, Anand N. Adverse swelling associated with use of rh-BMP-2 in anterior cervical discectomy and fusion: a case study. Spine J 2007;7:235-9. 271. Dickerman RD, Reynolds A, Morgan B. Calcium phosphate silicate for spinal fusion: a good alternative to bone morphogenetic protein-2! Spine J 2008;8:1046-7. 272. Wong DA, Kumar A, Jatana S, Ghiselli G, Wong K. Neurologic impairment from ectopic bone in the lumbar canal: a potential complication of off-label PLIF/TLIF use of bone morphogenetic protein-2 (BMP-2). Spine J 2008;8:1011-8. 273. Brower RS, Vickroy NM. A case of psoas ossification from the use of BMP-2 for posterolateral fusion at L4-L5. Spine (Phila Pa 1976) 2008;33:E653-5. 274. Dickman CA. A prospective, randomized, controlled cervical fusion study using recombinant human bone morphogenetic protein-2 with the CORNERSTONE-SR allograft ring and the ATLANTIS anterior cervical plate: point of view. Spine (Phila Pa 1976) 2003;28:1225. 275. Dimar JR, 2nd, Glassman SD, Kenneth Burkus J, Pryor PW, Hardacker JW, Carreon LY. Reply to "A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned". Spine J 2011;11:1082-3. 276. Dmitriev AE, Lehman RA, Symes AJ. Adverse events and bone morphogenetic protein-2 Response. Spine J 2011;11:802-03. 277. Epstein N. Outcomes of bone morphogenetic protein-2 in mature adults: posterolateral non- instrument-assisted lumbar decompression and fusion - Commentary. Surg Neurol 2008;69:462. 278. Hamilton DK, Jones-Quaidoo SM, Sansur C, Shaffrey CI, Oskouian R, Jane JA, Sr. Outcomes of bone morphogenetic protein-2 in mature adults: posterolateral non-instrument-assisted lumbar decompression and fusion. Surg Neurol 2008;69:457-61. 279. Gill JB. Re: Dimar JR, Glassman SD, Burkus KJ, et al. Clinical outcomes and fusion success at 2 years of single-level instrumented posterolateral fusions with recombinant human bone morphogenetic protein-2/compression resistant matrix versus iliac crest bone graft. Spine 2006; 31:2534-40. Spine (Phila Pa 1976) 2007;32:708. 280. Heggeness MH. Important considerations on bone morphogenetic protein-2 and neuroinflammation. Spine J 2011;11:506. 281. Dmitriev AE, Lehman RA, Jr., Symes AJ. Bone morphogenetic protein-2 and spinal arthrodesis: the basic science perspective on protein interaction with the nervous system. Spine J 2011;11:500-5.
131
282. Hu SS. Iliac crest bone graft: are the complications overrated? Spine J 2011;11:538-9. 283. Jacobson M. Radiographic and CT evaluation of recombinant human bone morphogenetic protein-2-assisted spinal interbody fusion (American Journal of Roentgenology (2011) 197, (W128- W133)). AJR Am J Roentgenol 2011;197:1024. 284. Sethi A, Craig J, Bartol S, Chen W, Jacobson M, Coe C, et al. Radiographic and CT evaluation of recombinant human bone morphogenetic protein-2-assisted spinal interbody fusion. AJR Am J Roentgenol 2011;197:W128-33. 285. Kang JD. Another complication associated with rhBMP-2? Spine J 2011;11:517-19. 286. Krishnaney AA, Hiremath GK, Steinmetz MP. Re: Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use recombinant human bone morphogenetic protein in posterior cervical fusion? Spine 2009; 34: 885-9. Response. Spine (Phila Pa 1976) 2010;35:361-62. 287. Mirza SK. Folly of FDA-approval studies for bone morphogenetic protein. Spine J 2011;11:495-9. 288. Pagni CA. Outcomes of bone morphogenetic protein-2 in mature adults: posterolateral non- instrument-assisted lumbar decompression and fusion - Commentary. Surg Neurol 2008;69:462. 289. Pawl R. Outcomes of bone morphogenetic protein-2 in mature adults: posterolateral non- instrument-assisted lumbar decompression and fusion - Commentary. Surg Neurol 2008;69:462. 290. Rakovac M, Bojanic I, Smoljanovic T. Recombinant human bone morphogenetic protein 2 labeled use in spinal surgery and sexual dysfunction. Surg Neurol Int 2011;2:55. 291. Singh K, Smucker JD, Gill S, Boden SD. Erratum: Use of recombinant human bone morphogenetic protein-2 as an adjunct in posterolateral lumbar spine fusion: a prospective CT-scan analysis at one and two years (Journal of Spinal Disorders and Techniques (2006) 19, (416-423)). J Spinal Disord Tech 2007;20:185. 292. Slosar PJ, Josey R, Reynolds J. RE: complications attributable to the use of rhBMP-2 inside the femoral ring allograft during anterior lumbar interbody fusion - Response. Spine J 2008;8:414. 293. Smoljanovic T, Pecina M. Re: complications attributable to the use of rhBMP-2 inside the femoral ring allograft during anterior lumbar interbody fusion. Spine J 2008;8:413-4. 294. Smoljanovic T, Vukicevic S, Pecina M. Bone morphogenetic protein and fusion. J Neurosurg Spine 2007;6:378-9. 295. Villavicencio AT, Burneikiene S, Nelson EL, Bulsara KR, Favors M, Thramann J. Safety of transforaminal lumbar interbody fusion and intervertebral recombinant human bone morphogenetic protein-2. J Neurosurg Spine 2005;3:436-43. 296. Smoljanovic T, Pecina M. Re: Burkus J K, Sandhu H S, Gornet M F. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine 2006;31:775-81. Spine (Phila Pa 1976) 2008;33:226. 297. Smoljanovic T, Pecina M. Re: Burkus J K, Transfeldt E E, Kitchel S H, et al. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine 2002;27:2396-408. Spine (Phila Pa 1976) 2008;33:224. 298. Smoljanovic T, Bojanic I, Cimic M. Bone morphogenetic protein. J Neurosurg Spine 2009;11:92- 3. 299. Tumialan LM, Pan J, Rodts GE, Mummaneni PV. The safety and efficacy of anterior cervical discectomy and fusion with polyetheretherketone spacer and recombinant human bone morphogenetic protein-2: a review of 200 patients. J Neurosurg Spine 2008;8:529-35. 300. Smoljanovic T, Bojanic I, Pecina M. The use of bone morphogenetic protein in lumbar spine surgery. J Bone Joint Surg Am 2009;91:2045-6. 301. Smoljanovic T, Bojanic I, Pecina M. End plates resorptions after the applications of rhBMP-2 for interbody spinal fusions. J Spinal Disord Tech 2009;22:309. 302. Vaidya R, Sethi A, Bartol S, Jacobson M, Coe C, Craig JG. Complications in the use of rhBMP-2 in PEEK cages for interbody spinal fusions. J Spinal Disord Tech 2008;21:557-62. 303. Smoljanovic T, Janjanin S, Bojanic I. Avoiding unanticipated adverse effects of recombinant human bone morphogenetic protein-2 therapy in craniofacial surgery with experiences from spinal applications. J Craniofac Surg 2009;20:1626. 304. Smoljanovic T, Bojanic I, Delimar D. Adverse effects of posterior lumbar interbody fusion using rhBMP-2. Eur Spine J 2009;18:920-3. 305. Meisel HJ, Schnoring M, Hohaus C, Minkus Y, Beier A, Ganey T, et al. Posterior lumbar interbody fusion using rhBMP-2. Eur Spine J 2008;17:1735-44. 306. Smoljanovic T, Bicanic G, Bojanic I. Re: Kleeman TJ, Ahn UM, Talbot-Kleeman A. Laparoscopic anterior lumbar interbody fusion with rhBMP-2: a prospective study of clinical and radiographic outcomes. Spine 2001;26:2751-6. Spine (Phila Pa 1976) 2010;35:E1013.
132
307. Smoljanovic T, Bojanic I. Re: Mroz TE, Wang JC, Hashimoto R, et al.Complications related to osteobiologics use in spine surgery: a systematic review. Spine (Phila Pa 1976) 2010;35:S86-104. Spine (Phila Pa 1976) 2010;35:E1010. 308. Smoljanovic T, Bojanic I, Bicanic G, Delimar D. Re: Toth JM, Boden SD, Burkus JK, et al. Short- term osteoclastic activity induced by locally high concentrations of recombinant human bone morphogenetic protein-2 in a cancellous bone environment. Spine 2009;34:539-50. Spine (Phila Pa 1976) 2010;35:597. 309. Smoljanovic T, Bojanic I, Cimic M. Re: Boden SD, Zdeblick TA, Sandhu HS, et al. The use of rhBMP-2 in interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine 2000; 25:376-81. Spine (Phila Pa 1976) 2010;35:E1010-11. 310. Smoljanovic T, Cimic M, Bojanic I. Is a barrier really necessary to prevent radiculitis when using recombinant human bone morphogenetic protein-2 in proximity of nerve roots? Spine J 2010;10:279. 311. Smoljanovic T, Bojanic I, Dokuzovic S. Re: Mindea SA, Shih P, Song JK. Recombinant human bone morphogenetic protein-2-induced radiculitis in elective minimally invasive transforaminal lumbar interbody fusions: a series review. Spine 2009;34:1480-5. Spine (Phila Pa 1976) 2010;35:929. 312. Smoljanovic T, Bojanic I, Vlahovic Z. Safety of posterior interbody fusions of the lumbar spine using rhBMP-2. J Spinal Disord Tech 2010;23:78. 313. Geibel PT, Boyd DL, Slabisak V. The use of recombinant human bone morphogenic protein in posterior interbody fusions of the lumbar spine: a clinical series. J Spinal Disord Tech 2009;22:315-20. 314. Smoljanovic T, Caric D, Bojanic I. RE: Clinical applications of bone morphogenetic proteins: current evidence. Kanakaris, NK; Giannoudis, PV; JSOA 17(3):133-46, 2008. J Surg Orthop Adv 2010;19:135-7. 315. Smoljanovic T, Cimic M, Bojanic I. Aggressive end plate decortication as a cause of osteolysis after rhBMP-2 use in cervical spine interbody fusion. Spine J 2010;10:187-8. 316. Klimo P, Jr., Peelle MW. Use of polyetheretherketone spacer and recombinant human bone morphogenetic protein-2 in the cervical spine: a radiographic analysis. Spine J 2009;9:959-66. 317. Smoljanovic T, Dokuzovic S, Bojanic I. Osteoinductive bone graft substitutes for lumbar fusion. J Neurosurg Spine 2010;13:407-9. 318. Smoljanovic T, Siric F, Bojanic I. Six-year outcomes of anterior lumbar interbody arthrodesis with use of interbody fusion cages and recombinant human bone morphogenetic protein-2. J Bone Joint Surg Am 2010;92:2614-5. 319. Smoljanovic T, Stipcic I, Bojanic I. The level of evidence in clinical studies of the use of rhBMP-2 in spinal interbody fusions. Eur Spine J 2010;19:1387-8. 320. Smoljanovic T, Aljinovic A, Bojanic I. Recommendation for use of rhBMP-2 in spinal interbody fusions. Eur Spine J 2010;19:1385-6. 321. Smoljanovic T, Bojanic I. An evolving perception of the risk of rhBMP-2 use for anterior spinal interbody fusions. Spine J 2011;11:520-1. 322. Smoljanovic T, Rakovac M, Bojanic I. Could chronic host inflammatory response be responsible for delayed onset of retrograde ejaculation after the labeled use of recombinant human bone morphogenetic protein-2? Spine J 2011;11:167-8. 323. Muchow RD, Hsu WK, Anderson PA. Histopathologic inflammatory response induced by recombinant bone morphogenetic protein-2 causing radiculopathy after transforaminal lumbar interbody fusion. Spine J 2010;10:e1-6. 324. Tumialan LM, Rodts GE. Adverse swelling associated with use of rh-BMP-2 in anterior cervical discectomy and fusion. Spine J 2007;7:509-10. 325. Tumialan LM, Rodts GE. rh-BMP-2 can be used safely in the cervical spine: dose and containment are the keys! Spine J 2007;7:509-10. 326. Tumialan LM, Rodts GE, Mummaneni PV. Bone morphogenetic protein. Response. J Neurosurg Spine 2009;11:93-94. 327. Vaidya R. Transforaminal interbody fusion and the "off label" use of recombinant human bone morphogenetic protein-2. Spine J 2009;9:667-9. 328. Vaidya R, Sethi A, Bartol S, Jacobsen M, Coe C, Craig JG. End plates resorptions after the applications of rhBMP-2 for interbody spinal fusions response. J Spinal Disord Tech 2009;22:310. 329. Wang JC. Gene therapy for spinal fusion. Spine J 2011;11:557-9. 330. White AP. Lumbar spinal fusion rates as influenced by bone grafts and bone graft alternatives: a critical appraisal of common clinical and radiographic comparative methodologies. Spine J 2009;9:916-18. 331. Wolfla C. A comprehensive review of the safety profile of bone morphogenetic protein in spine surgery - Comments. Neurosurgery 2008;62:431.
133
332. Woo EJ. Expanded indication for recombinant human bone morphogenetic protein 2. Spine (Phila Pa 1976) 2011;36:1817. 333. Woo EJ. Re: Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic protein increase the incidence of perioperative complications in spinal fusion? Spine 2011;36:1658-91. Spine (Phila Pa 1976) 2012;37:259. 334. Wrubel DM, Kaufman BA. Routine use of recombinant human bone morphogenetic protein-2 in posterior fusions of the pediatric spine: safety profile and efficacy in the early postoperative period. Comments. Neurosurgery 2010;67:1204. 335. Fahim DK, Whitehead WE, Curry DJ, Dauser RC, Luerssen TG, Jea A. Routine use of recombinant human bone morphogenetic protein-2 in posterior fusions of the pediatric spine: safety profile and efficacy in the early postoperative period. Neurosurgery 2010;67:1195-204. 336. Wu JC, Mummaneni PV. Safety, efficacy, and dosing of recombinant human bone morphogenetic protein-2 for posterior cervical and cervicothoracic instrumented fusion with a minimum 2-year follow-up: Commentary. Neurosurgery 2011;69:111. 337. Hamilton DK, Smith JS, Reames DL, Williams BJ, Chernavvsky DR, Shaffrey CI. Safety, efficacy, and dosing of recombinant human bone morphogenetic protein-2 for posterior cervical and cervicothoracic instrumented fusion with a minimum 2-year follow-up. Neurosurgery 2011;69:103-11. 338. Yaremchuk K, Peterson E. In response to acute airway obstruction in cervical spinal procedures with bone morphogenetic proteins. Laryngoscope 2011;121:2502-3. 339. A retrospective review of posterior cervical fusions with rhBMP-2/ACS. Spine (Phila Pa 1976) 2009;20091203. 340. Luque E. Latest clinical results using demineralised bone materials and rhBMP-2: the Mexican experience. In: Total Spine: Advanced Concepts and Constructs. Cancun, Mexico; 2002. 341. Sekhon L, Tomlinson A, Allen B, Lynch J, Hurlbert JR. Immediate postoperative complications and radiological features of interbody fusion between infuse bmp andactifuse: A randomized prospective trial. Spine (Phila Pa 1976). Conference: 37th Annual Meeting of the Cervical Spine Research Society, CSRS 2009;20091203. 342. Alexander D, Oxner W, Soroceanu A, Kelly A, Shakespeare D. A prospective randomized clinical trial of posterolateral lumbosacral spinal fusion with BMP-2 and titanium pedicle screw instrumentation versus BMP-2 alone: preliminary 6-month results. Can J Surg 2009;52:S21. 343. Anderson DW, Burton DC, Jackson RS. Postoperative cervical myelopathy and cord compression associated with the use of recombinant bone morphogenetic protein-2 in posterior cervical decompression, instrumentation, and arthrodesis: a report of two cases. Spine (Phila Pa 1976) 2011;36:E682-6. 344. Anderson CL, Whitaker MC. Heterotopic ossification associated with rhBMP-2 (Infuse) in posterolateral lumbar spine fusion: a case report. Spine (Phila Pa 1976) 2011. 345. Antonucci M, Bennett J, Pincus DW. Use of recombinant human bone morphogenetic protein 2 in pediatric spinal fusion. Neurosurgery 2008;62:1416-17. 346. Balseiro S, Nottmeier EW. Vertebral osteolysis originating from subchondral cyst end plate defects in transforaminal lumbar interbody fusion using rhBMP-2. Report of two cases. Spine J 2010;10:e6-e10. 347. Boakye M, Mummanen iP, Garrett M, Rodts G, Haid R. Anterior cervical discectomy and fusion involving a polyetheretherketone spacer and bone morphogenetic protein. J Neurosurg Spine 2005;2:521-5. 348. Carreon LY, Glassman SD, Brock DC, Dimar JR, Puno RM, Campbell MJ. Adverse events in patients re-exposed to bone morphogenetic protein for spine surgery. Spine (Phila Pa 1976) 2008;33:391-3. 349. Chen N-F, Smith ZA, Stiner E, Armin S, Sheikh H, Khoo LT. Symptomatic ectopic bone formation after off-label use of recombinant human bone morphogenetic protein-2 in transforaminal lumbar interbody fusion. J Neurosurg Spine 2010;12:40-6. 350. Crandall D, Patterson J, Huish E, Revella J, Datta J, Chang M, et al. RhBMP-2 in TLIF: dose related complications from a large series. Spine J 2011;11:61S. 351. Deutsch H. High-dose bone morphogenetic protein-induced ectopic abdomen bone growth. Spine J 2010;10:e1-4. 352. Hamilton DK, Smith JS, Reames DL, Williams BJ, Shaffrey CI. Use of recombinant human bone morphogenetic protein-2 as an adjunct for instrumented posterior arthrodesis in the occipital cervical region: an analysis of safety, efficacy, and dosing. J Craniovertebr Junction Spine 2010;1:107-12. 353. Hansen SM, Sasso RC. Resorptive response of rhBMP2 simulating infection in an anterior lumbar interbody fusion with a femoral ring. J Spinal Disord Tech 2006;19:130-4.
134
354. Helgeson MD, Lehman RA, Jr., Patzkowski JC, Dmitriev AE, Rosner MK, Mack AW. Adjacent vertebral body osteolysis with bone morphogenetic protein use in transforaminal lumbar interbody fusion. Spine J 2011;11:507-10. 355. Hodges SD, Eck JC, Newton D. Retrospective study of posterior cervical fusions with rhBMP-2. Orthopedics 2012;35:e895-8. 356. Hoffmann M, Jones CB. Complication rates utilizing rhBMP-2 for lumbar posterolateral fusions. Spine J 2011;Conference: 26th Annual Meeting of the North American Spine Society, NASS 2011 Chicago, IL United States. Conference Start: 20111102 Conference End: 20111105. Conference Publication::164S-65S. 357. Hoffmann M, Jones CB. Recombinant human bone morphogenetic protein-2 (rhBMP-2) in posterolateral spine fusion: what's the correct successful dose/dosage? Spine J 2011;11:113S-14S. 358. Hoffmann M, Jones CB. Recombinant human bone morphogenetic protein-2 (rhBMP-2) in posterolateral lumbar spine fusion: complications in the elderly. Spine J 2011;11:61S-62S. 359. Katayama Y, Matsuyama Y, Yoshihara H, Sakai Y, Nakamura H, Imagama S, et al. Clinical and radiographic outcomes of posterolateral lumbar spine fusion in humans using recombinant human bone morphogenetic protein-2: an average five-year follow-up study. Int Orthop 2009;33:1061-7. 360. Kepler CK, Huang RC, Meredith D, Cunningham M, Boachie-Adjei O. Delayed pleural effusion after anterior thoracic spinal fusion using bone morphogenetic protein-2. Spine (Phila Pa 1976) 2011;36:E365-9. 361. Knox JB, Dai JM, 3rd, Orchowski J. Osteolysis in transforaminal lumbar interbody fusion with bone morphogenetic protein-2. Spine (Phila Pa 1976) 2011;36:672-6. 362. Kuklo TR, Rosner MK, Polly DW, Jr. Computerized tomography evaluation of a resorbable implant after transforaminal lumbar interbody fusion. Neurosurg Focus 2004;16:E10. 363. Lanman TH, Hopkins TJ. Early findings in a pilot study of anterior cervical interbody fusion in which recombinant human bone morphogenetic protein-2 was used with poly(L-lactide-co-D,L-lactide) bioabsorbable implants. Neurosurg Focus 2004;16:E6. 364. Lanman TH, Hopkins TJ. Lumbar interbody fusion after treatment with recombinant human bone morphogenetic protein-2 added to poly(L-lactide-co-D,L-lactide) bioresorbable implants. Neurosurg Focus 2004;16:E9. 365. Lehman RA, Jr. Vertebral body osteolysis after minimal-access transforaminal interbody fusion. Spine J 2011;11:581-2. 366. Lewandrowski K-U, Nanson C, Calderon R. Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: a report of five cases. Spine J 2007;7:609-14. 367. Lindley TE, Dahdaleh NS, Menezes AH, Abode-Iyamah KO. Complications associated with recombinant human bone morphogenetic protein use in pediatric craniocervical arthrodesis. J Neurosurg 2011;7:468-74. 368. McClellan JW, Mulconrey DS, Forbes RJ, Fullmer N. Vertebral bone resorption after transforaminal lumbar interbody fusion with bone morphogenetic protein (rhBMP-2). J Spinal Disord Tech 2006;19:483-6. 369. Mannion RJ, Nowitzke AM, Wood MJ. Promoting fusion in minimally invasive lumbar interbody stabilization with low-dose bone morphogenic protein-2--but what is the cost? Spine J 2011;11:527-33. 370. Moshel YA, Hernandez EI, Kong L, Liu C, Samadani U. Acute renal insufficiency, supraventricular tachycardia, and confusion after recombinant human bone morphogenetic protein-2 implantation for lumbosacral spine fusion. J Neurosurg Spine 2008;8:589-93. 371. Mulconrey DS, Bridwell KH, Flynn J, Cronen GA, Rose PS. Bone morphogenetic protein (RhBMP-2) as a substitute for iliac crest bone graft in multilevel adult spinal deformity surgery: minimum two-year evaluation of fusion. Spine (Phila Pa 1976) 2008;33:2153-9. 372. Neal CJ, Helgeson MD, Lehman RA, Mack AW, Rosner MK. Adjacent vertebral body osteolysis with bone morphogenetic protein use in transforaminal lumbar interbody fusion. J Neurosurg 2008;109:A371-A72. 373. Oetgen ME, Richards BS. Complications associated with the use of bone morphogenetic protein in pediatric patients. J Pediatr Orthop 2010;30:192-8. 374. Owens K, Glassman SD, Howard JM, Djurasovic M, Witten JL, Carreon LY. Perioperative complications with rhBMP-2 in transforaminal lumbar interbody fusion. Eur Spine J 2011;20:612-7. 375. Pargament J, Stambough JL, Clouse EK. Swelling associated with use of rhBMP-2 in posterolateral lumbar fusion: a case study. Curr Orthop Pract 2009;20:698-702. 376. Pradhan B, Bae W, Kropf M, Patel V, Delamarter R. Radiographic staging, characterization, and outcome of anterior cervical discectomy and fusion using allograft, recombinant human bone
135
morphogenetic protein-2 and plate. Poster no. 1304. In: 51st Annual Meeting of the Orthopaedic Research Society. Washington DC; 2005. 377. Reames DL, Hamilton DK, Smith JS, Williams BS, Chernavvsky DR, Shaffrey CI. Safety, efficacy, and dosing of recombinant human bone morphogenetic protein 2 (rhBMP 2) for posterior cervical and cervico-thoracic instrumented fusion with a minimum two year follow-up. J Neurosurg 2011;115:A457- A58. 378. Rihn JA, Makda J, Hong J, Patel R, Hilibrand AS, Anderson DG, et al. The use of RhBMP-2 in single-level transforaminal lumbar interbody fusion: a clinical and radiographic analysis. Eur Spine J 2009;18:1629-36. 379. Robin BN, Chaput CD, Zeitouni S, Rahm MD, Zerris VA, Sampson HW. Cytokine-mediated inflammatory reaction following posterior cervical decompression and fusion associated with recombinant human bone morphogenetic protein-2: a case study. Spine (Phila Pa 1976) 2010;35:E1350-4. 380. Schwender JD, Holly LT, Rouben DP, Foley KT. Minimally invasive transforaminal lumbar interbody fusion (TLIF): technical feasibility and initial results. J Spinal Disord Tech 2005;18 Suppl:S1- 6. 381. Shah RK, Moncayo VM, Smitson RD, Pierre-Jerome C, Terk MR. Recombinant human bone morphogenetic protein 2-induced heterotopic ossification of the retroperitoneum, psoas muscle, pelvis and abdominal wall following lumbar spinal fusion. Skeletal Radiol 2010;39:501-4. 382. Shahlaie K, Kim KD. Occipitocervical fusion using recombinant human bone morphogenetic protein-2: adverse effects due to tissue swelling and seroma. Spine (Phila Pa 1976) 2008;33:2361-6. 383. Shen HX, Buchowski JM, Yeom JS, Liu G, Lin N, Riew KD. Pseudarthrosis in multilevel anterior cervical fusion with rhBMP-2 and allograft: analysis of one hundred twenty-seven cases with minimum two-year follow-up. Spine (Phila Pa 1976) 2010;35:747-53. 384. Stachniak JB, Diebner JD, Brunk ES, Speed SM. Analysis of prevertebral soft-tissue swelling and dysphagia in multilevel anterior cervical discectomy and fusion with recombinant human bone morphogenetic protein-2 in patients at risk for pseudarthrosis. J Neurosurg Spine 2011;14:244-9. 385. Stambough JL, Clouse EK, Stambough JB. Instrumented one and two level posterolateral fusions with recombinant human bone morphogenetic protein-2 and allograft: a computed tomography study. Spine (Phila Pa 1976) 2010;35:124-9. 386. Whang PG, O'Hara BJ, Ratliff J, Sharan A, Brown Z, Vaccaro AR. Pseudarthrosis following lumbar interbody fusion using bone morphogenetic protein-2: intraoperative and histopathologic findings. Orthopedics 2008;31. Available from: http://www.orthosupersite.com/view.aspx?rid=32087 387. Zebala L, Buchowski J, Bridwell K, Cho S, Pahys J, Kang M, et al. RhBMP-2 and modern surgical techniques significantly reduce the pseudarthrosis rate in long fusions to the sacrum for complex adult spinal deformity. Spine J 2011;11:149S-50S. 388. Simmonds M, Tierney J, Bowden J, Higgins J. Meta-analysis of time-to-event data: a comparison of two-stage methods. Res Synth Methods 2011.
136 Appendix I - Protocol
Appendix I
Systematic Review and Meta-Analysis of the Safety and Efficacy of Recombinant Human Bone Morphogenetic Protein-2 (rhBMP-2) - PROTOCOL
Background Human bone morphogenetic protein-2 (rhBMP-2) is used in orthopaedic and spinal surgery to promote fusion. Following 2002 FDA approval for use in anterior lumbar fusion (ALIF) surgery, numbers of spinal procedures using rhBMP-2 grew rapidly, including many in off-label indications. This is despite the fact that the majority of procedures will fuse without use of rhBMP-2 and that in many cases fusion per se is not a pre-requisite for successful surgery. More recently, a number of small studies have raised concern over high rates of adverse events (AE) some of which are potentially life-threatening and which had not been reported in licensing studies. The need for rhBMP- 2, its efficacy and its AE profile is therefore under considerable scrutiny and a robust re-evaluation of the research evidence is vital. In recognition of the importance of this issue, the manufacturers of rhBMP-2 (Medtronic Inc) released all of its clinical research data that are relevant to the use of rhBMP-2 to Yale University for independent scrutiny and review. Yale has contracted two academic groups to carry out independent and unrestricted systematic reviews of the safety and efficacy of rhBMP-2 in spinal fusion, including re-analysis of the individual participant data from Medtronic studies. This protocol describes the systematic review and associated methodological comparisons to be carried out by the Centre for Reviews and Dissemination (CRD), University of York, UK. The systematic review will consider data from Medtronic studies alongside any non-industry funded clinical research data on rhBMP-2. In addition to the reviews of benefits and harms described below, the planned research will inform discussions around the comparability of IPD and aggregate data syntheses, and provide objective evidence around data disclosure, selective reporting, industry sponsorship, the availability of individual participant data, and how these might affect the ability to learn about efficacy and adverse events. In addition, this work will contribute to a wider debate on standards for data disclosure and dealing with potential conflict of interest.
Objectives To evaluate whether rhBMP-2 is more or less effective than standard bone graft therapy (SBGT) in spinal fusion by rigorous systematic review and meta-analysis of relevant studies, including an analysis of individual participant data (IPD), considering • Potential benefits of rhBMP-2, focusing on evidence from randomised controlled trials and on outcomes that are meaningful to patients. • Potential harms of rhBMP-2 by identifying serious adverse events that have been reported from its use in clinical trials and in general medical practice.
Methods for synthesis of evidence of clinical effectiveness and safety Inclusion and exclusion criteria Two reviewers will independently screen all titles and abstracts retrieved from electronic database and other searches. Full paper manuscripts of any publications that may be relevant will be obtained (where possible) and the relevance of each study assessed by two reviewers according to the criteria below. Studies that do not meet all of the criteria will be excluded and their bibliographic details listed with reasons for exclusion. Any discrepancies will be resolved by consensus and, if necessary, a third reviewer will be consulted.
137 Appendix I - Protocol
Participants Studies including patients undergoing spinal fusion surgery for treatment of degenerative disc disease, spondylolisthesis or any other relevant spinal condition will be included. Although the licensed indication is for use with anterior lumbar interbody fusion, inclusion will not be restricted by operative approach as this will allow evaluation of evidence pertaining to off-label use, particularly with respect to adverse events in cervical and lumbosacral spinal surgery. Anterior lumbar interbody fusion (ALIF), posterolateral lumbar fusion (PLF), posterolateral lumbar interbody fusion (PLIF), transforaminal lumbar interbody fusion (TLIF), and anterior cervical discectomy and fusion (ACDF) will all be eligible. Inclusion will not be restricted by type of surgery (open, minimally invasive or laparoscopic). Studies of rhBMP-2 use outside spinal fusion surgery (e.g. in long bone fractures) will be excluded, as will all animal and in vitro studies. This study will focus only on rhBMP2 and not on other recombinant forms of BMP, such as rhBMP7. Interventions Studies evaluating rhBMP-2 in spinal fusion will be included in the review, including both Medtronic’s INFUSE/Inductos and AMPLIFY rhBMP-2 carrier/preparations. Comparators For the evaluation of benefits, studies comparing rhBMP-2 against any standard bone graft techniques (SBGT) will be eligible for inclusion. Studies without a comparator will also be eligible for inclusion in the evaluation of AEs potentially related to the use of rhBMP-2. Outcomes Inclusion will not be limited by which outcomes are reported. Any studies that are eligible but do not report outcomes of interest will be included and reported as such. Study designs Effectiveness: Only RCTs meeting the above criteria will be included in the review of comparative effectiveness. Adverse events: In accordance with best practice,1 the review of AEs will not be restricted to RCTs. A decision about which study designs will be included will be taken once we have completed a mapping exercise to determine the volume and scope of studies reporting adverse events, and made a preliminary assessment of the completeness and quality of the available registry data. Attention will be paid to the potential for overlap among identified studies (i.e. the same patients being included in case reports, cohort studies and registries).
Search strategy A systematic literature search will be performed of the following databases: BIOSIS Previews, Cochrane Central Register of Controlled Trials (Central), the Database of Abstracts of Reviews of Effects (DARE), EMBASE, MEDLINE, MEDLINE in Process and Other Non-Indexed Citations, PubMed, Science Citation Index, TOXLINE, and the FDA website. The search strategy will be designed to retrieve any studies relevant to the effectiveness or adverse events (AEs) of rhBMP-2 in spinal fusion. Synonyms will be searched for in title and abstract and appropriate indexing/ keywords selected. No search filters for specific study designs will be used owing to limitations of searching beyond randomised controlled trials (RCTs) for AEs.2 Searches will not be restricted by publication status or date. References of relevant papers will be checked for further relevant studies. Authors of included trials will be asked to notify us of any unpublished studies of which they are aware. A summary of the search strategies used are presented in Appendix 1. (This appendix is not included in this report, The full search strategy can be found in appendix II).
138 Appendix I - Protocol
Provision of IPD Yale University has secured the release of all clinical trial data (published and unpublished), post- marketing surveillance data and spontaneous adverse event data from Medtronic (the manufacturer of rhBMP-2). Yale will make these data available for inclusion in this systematic review. We will also seek IPD for any additional RCTs identified by the literature searches.
Obtaining Data Data extraction strategy (for published aggregate data) A data extraction form will be developed, piloted and adjusted as necessary. Data extracted will include details of study design, setting, and sponsor as well as outcome, trial and patient characteristics. Data will be extracted into EPPI-Reviewer/Excel. Data extraction will be undertaken independently by two researchers with discrepancies resolved by consensus or recourse to a third researcher if necessary. Provision of individual participant data Investigators of trials for which IPD is not supplied by Medtronic will be contacted and asked to participate in the review by providing individual participant data for inclusion and re-analysis. If they agree to participate, fully anonymised data on all randomised patients relating to the outcomes and trial and patient characteristics described above will be requested. Data will be accepted in any suitable electronic format, but an example format detailing the recommended coding will be created and offered to all collaborators. Simple checks on the data will be made to ensure data are correctly coded, that missing data are correctly identified and to ensure that the data are consistent with published results. Data from all trials (including Medtronic-sponsored trials) will be incorporated into a single, database with fields that are consistent (as far as possible) across both Medtronic and non-Medtronic trials. Data storage and confidentiality All IPD and adverse event data from Medtronic will be transferred to CRD via password-protected memory stick. Data from other investigators will be in a de-identified format and received via secure FTP transfer or encrypted email. All data will be held in a password protected area of the CRD server. Access will be limited to staff working directly on the project. Copying data to laptop computers or memory sticks will be prohibited.
Outcomes We will consider a range of outcomes and place emphasis on clinical or functional over radiological outcomes, particularly those that are directly meaningful to patients. Effectiveness outcomes The outcomes of interest will be: • disease-specific questionnaires (e.g. Oswestry Disability Index, Neck Disability Index) • patient QL / functional status questionnaires (e.g. SF-36) • post-operative pain – surgical site and bone graft donor site • duration of hospital stay • operating time • successful return to work/usual activity • fusion status • time to discharge* Analysis of outcomes marked * are likely to be possible only in the IPD review.
139 Appendix I - Protocol
Adverse event outcomes General • heterotopic bone formation • osteolysis • infection • neurological events (new / worse leg pain, sensory disturbance, reflex changes, bladder disturbance) • cancer • hardware failure (e.g. cage subsidence, implant breakages) Surgery-specific • ALIF – major vascular injury, retrograde ejaculation, urinary retention • ACDF - dysphagia, airway obstruction, neck pain, recurrent laryngeal nerve palsy • PLIF/TLIF/PLF – leg pain/radiculitis, leg weakness, inflammatory cyst formation Data on serious adverse events (AEs) will be extracted from relevant observational sources as well as from RCTs (where reported) Other data In order to investigate how the efficacy and safety of rhBMP-2 might be influenced by trial and patient- level characteristics we will also extract/obtain data on the following, if available: • spinal location of surgery (e.g. cervical or lumbosacral) • type of surgery (e.g. ALIF, PLIF etc.) • rhBMP-2 dose/volume • cage type • nature of spinal condition (e.g. degenerative disc disease, spondylolisthesis) • comparator treatment • previous surgical interventions * • age * • sex * • smoking status * Data marked * may only be available from IPD
Risk of bias (quality) assessment Critical appraisal of RCTs will be based on trial publications, protocols, and where available on IPD. Risk of bias in RCTS will be assed using the Cochrane Risk of Bias tool.3 Other study designs will be assessed based on CRD guidance.4 Assessment will be undertaken independently by two researchers with any discrepancies resolved by consensus or recourse to a third researcher if necessary. All IPD will be subject to detailed checking including examination of patterns of missing data, integrity of randomisation via pattern of randomisation and balance across baseline characteristics, and internal consistency.
Methods of analysis/synthesis A narrative and tabular summary of key study characteristics will be undertaken. Published main results and quality assessment of individual RCTs will also be tabulated. Aggregate data from RCTs Where appropriate (based on clinical similarity of trials and the necessary data being available) aggregate study results (from publications/FDA reports) will be combined in a series of random- effects meta-analyses. Separate analyses will be conducted for each outcome listed above.
140 Appendix I - Protocol
It is anticipated that the measures used to assess continuous outcomes will vary between studies, and in these cases standardized mean differences will be calculated where appropriate, and combined in random-effects meta-analyses. Heterogeneity and inconsistency across trials will be assessed using χ2 tests and quantified using the I2 statistic.5 Exploring clinical heterogeneity (subgroup analyses and meta-regression) We will investigate how trial-level and (where feasible) patient-level covariates influence the effectiveness of rhBMP-2 therapy. Trials will be grouped by type of surgery as follows: ALIF; PLIF; TLIF; PLF; ACDF. Meta-analyses will be conducted within each surgery subgroup and the results across subgroups compared using tests of heterogeneity to identify differences between surgery types. Similar subgroup analyses and tests for differences across them will be undertaken to identify differences in effectiveness of treatment at differing spinal locations. If feasible, meta-regression will be used to investigate dose-response relationships. Individual participant data Two stage modelling IPD will be analysed using standard two stage meta-analytic techniques. IPD from each trial will be analysed separately, using the same methods across trials, for each outcome listed above.6-7 The resultant summary statistics for each trial will then be combined using random-effects models to give overall estimates of the effect for each outcome explored. For time to event outcomes, we will apply Cox proportional hazards models to each trial, providing the proportional hazards assumption is not clearly violated.8 Exploring clinical heterogeneity (subgroup analyses) Characteristics that vary at the trial level (e.g. comparator type) will be investigated by analysing grouped trials or by meta-regression as described above for aggregate data. Patient-level characteristics, including underlying condition (degenerative disc disease, spondylolisthesis) age, smoking history, and previous surgery at the same site will be explored (data permitting) in univariate two-stage subgroup analyses.9-10 Characteristics such as type of surgery that may vary between trials (i.e. trials include only one type of surgery) will be analysed using whichever of the above approaches best suits the data. One stage modelling Time and data permitting, we will also analyse the IPD in a one-stage modelling framework.11 This will enable us to take account of multiple patient characteristics when comparing rhBMP-2 and SBGT (stratified by trial) and also enable simultaneous exploration of multiple potential interactions between treatment and patient or trial-level covariates. For these analyses, we will implement multilevel modelling approaches for binary,12 continuous,13 ordinal,14 and time-to-event15 outcomes. All IPD analyses will use complete case analysis as imputation of missing data is not feasible within the timescale of this project. Combining aggregate and individual participant data Where IPD is not available for relevant RCTs we will seek to combine aggregate data and IPD in meta-analyses using two-stage methods as described above and, time-permitting, using more advanced methods which have been described elsewhere.16 Results from these analyses will be compared to analyses of aggregate data only and of IPD only as a sensitivity analysis. Adverse events Adverse events will be categorised as described above and results tabulated. Where possible, rates of adverse events in rhBMP-2 surgery and in SBGT will be compared statistically. Where possible, meta-analyses of adverse events will be conducted to compare the incidence and nature of events across trials. Such analyses may be limited given the possible rarity and limited reporting of adverse events in RCTs. The underlying principles of meta-analysis will be applied even where statistical combination is not feasible: specifically, this will include evaluation of consistency of findings, direction and magnitude of effect, and strength of evidence, in addition to considerations of the quality of the evidence.
141 Appendix I - Protocol
Methodological exploration Investigations of bias We will investigate possible sources of bias in RCTs and studies of adverse events. In particular we will explore the role of funding source (industry/non industry), publication status (published/unpublished) study design (randomised/ non randomised; controlled/uncontrolled; different observational designs) and components of risk of bias assessments using subgroup meta-analyses and/or meta-regression to compare results from studies that differ by the factor in question. Comparing aggregate and IPD approaches IPD analyses will be compared with aggregate data analyses. Where results differ we will establish which differences are attributable to data availability (unpublished trials, unreported outcomes, excluded patients) and which are attributable to the methods of analyses. The summary statistics generated for each trial from the IPD will also be compared with the corresponding published results. Data disclosure timeline Construction of a timeline of what was reported where and when will provide background information on publication practice and the way that information about efficacy and adverse effects enter the public domain.
Outputs and dissemination A comprehensive report detailing all the analyses described above will be submitted to the coordinating centre by the agreed deadline. This report will meet the requirements of the PRISMA statement for the reporting of systematic reviews and meta-analyses for the appropriate sections of the review.17 We will also disseminate the findings through high-impact journal publications, conference presentations and other relevant channels. We intend that the first publication will coincide with the release of the full report by Yale.
142 Appendix I - Protocol
References
1. Loke YK, Price D, Herxheimer A. Systematic reviews of adverse effects: framework for a structured approach. BMC Med Res Methodol 2007;7:32-32. Available from: http://www.hubmed.org/display.cgi?uids=17615054 2. Golder S, Loke Y. Search strategies to identify information on adverse effects: a systematic review. J Med Libr Assoc 2009;97:84-92. Available from: http://www.hubmed.org/display.cgi?uids=19404498 3. Higgins JPT, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343. Available from: http://www.hubmed.org/display.cgi?uids=22008217 4. Centre for Reviews and Dissemination. Systematic reviews: CRD's guidance for undertaking reviews in health care. York: University of York; 2009. 5. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-60. Available from: http://www.hubmed.org/display.cgi?uids=12958120 6. Riley RD, Lambert PC, Abo-Zaid G. Meta-analysis of individual participant data: rationale, conduct, and reporting. BMJ 2010;340:c221. Available from: http://www.hubmed.org/display.cgi?uids=20139215 7. Stewart LA, Clarke MJ. Practical methodology of meta-analyses (overviews) using updated individual patient data. Cochrane Working Group. Stat Med 1995;14:2057-79. Available from: http://www.hubmed.org/display.cgi?uids=8552887 8. Simmonds M, Tierney J, Bowden J, Higgins J. Meta-analysis of time-to-event data: a comparison of two-stage methods. Res Synth Methods 2011. 9. Fisher DJ, Copas AJ, Tierney JF, Parmar MKB. A critical review of methods for the assessment of patient-level interactions in individual participant data meta-analysis of randomized trials, and guidance for practitioners. J Clin Epidemiol 2011;64:949-67. Available from: http://www.hubmed.org/display.cgi?uids=21411280 10. Thompson SG, Higgins JPT. Treating individuals 4: can meta-analysis help target interventions at individuals most likely to benefit? Lancet 2005;365:341-46. Available from: http://www.hubmed.org/display.cgi?uids=15664231 11. Simmonds MC, Higgins JPT, Stewart LA, Tierney JF, Clarke MJ, Thompson SG. Meta-analysis of individual patient data from randomized trials: a review of methods used in practice. Clin Trials 2005;2:209-17. Available from: http://www.hubmed.org/display.cgi?uids=16279144 12. Turner RM, Omar RZ, Yang M, Goldstein H, Thompson SG. A multilevel model framework for meta-analysis of clinical trials with binary outcomes. Stat Med 2000;19:3417-32. Available from: http://www.hubmed.org/display.cgi?uids=11122505 13. Higgins J, Whitehead A, Turner R, Omar R, Thompson S. Meta-analysis of continuous outcome data from individual patients. Stat Med 2001;20:2219-41. 14. Whitehead A, Omar RZ, Higgins JP, Savaluny E, Turner RM, Thompson SG. Meta-analysis of ordinal outcomes using individual patient data. Stat Med 2001;20:2243-60. Available from: http://www.hubmed.org/display.cgi?uids=11468762 15. Bowden J, Tierney J, Simmonds M, Copas A, JPT H. Individual patient data meta-analysis of time-to-event outcomes: one-stage versus two-stage approaches to estimating the hazard ratio. Res Synth Methods 2011. 16. Riley RD, Simmonds MC, Look MP. Evidence synthesis combining individual patient data and aggregate data: a systematic review identified current practice and possible methods. J Clin Epidemiol 2007;60:431-39. Available from: http://www.hubmed.org/display.cgi?uids=17419953 17. Liberati A, Altman D, Tetzlaff J, Mulrow C, Gøtzsche P, Ioannidis J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009;339:b2700.
143 Appendix II – Search strategies
Appendix II
Search strategies
Database: Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations and Ovid MEDLINE(R) <1948 to Present> Database interface: OvidSP Search date: 21st November 2011 Records identified: 827 Search Strategy: ------1 bone morphogenetic proteins/ or bone morphogenetic protein 2/ (10872) 2 ((bone morphogen$ or osteogen$ or osteoinduct$) adj (protein$ or factor$ or polypeptide$ or poly-peptide$)).af. (16131) 3 (bmp or bmp2 or bmp-2).af. (11384) 4 (rhbmp or rhbmp2 or rhbmp-2).af. (1262) 5 (rh-bmp or rh-bmp2 or rh-bmp-2).af. (72) 6 (infuse or amplify).af. (15590) 7 or/1-6 (33622) 8 Spinal Fusion/ (14504) 9 (spine or spinal).af. (322484) 10 spondylosyndes$.af. (7) 11 spondylodes$.af. (631) 12 lumbar interbody arthrodesis.af. (22) 13 ((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) adj3 fusion$).af. (7931) 14 fusion cage.af. (125) 15 or/8-14 (323015) 16 7 and 15 (1312) 17 exp animals/ not humans.sh. (3715340) 18 16 not 17 (827)
Database: Embase <1974 to 2011 November 18> Database interface: OvidSP Search date: 21st November 2011 Records identified: 1542 Search Strategy: ------1 bone morphogenetic protein/ or bone morphogenetic protein 2/ (12676)
144 Appendix II – Search strategies
2 ((bone morphogen$ or osteogen$ or osteoinduct$) adj (protein$ or factor$ or polypeptide$ or poly-peptide$)).af. (19870) 3 (bmp or bmp2 or bmp-2).af. (11817) 4 (rhbmp or rhbmp2 or rhbmp-2).af. (1435) 5 (rh-bmp or rh-bmp2 or rh-bmp-2).af. (90) 6 (infuse or amplify).af. (17204) 7 or/1-6 (38898) 8 Spine Fusion/ (13270) 9 (spine or spinal).af. (369003) 10 spondylosyndes$.af. (10) 11 spondylodes$.af. (1709) 12 lumbar interbody arthrodesis.af. (23) 13 ((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) adj3 fusion$).af. (9892) 14 fusion cage.af. (196) 15 or/8-14 (369756) 16 7 and 15 (1898) 17 animal experiment/ (1579120) 18 16 not 17 (1542)
Database: Cochrane Central Register of Controlled Trials (CENTRAL) Database interface: John Wiley Search date: 21st November 2011 Records identified: 218 Search Strategy:
ID Search Hits
#1 MeSH descriptor Bone Morphogenetic Proteins, this term only 86
#2 MeSH descriptor Bone Morphogenetic Protein 2, this term only 29
(morphogen* NEXT (protein* or factor* or polypeptide* or poly-peptide*)) or #3 (osteogen* NEXT (protein* or factor* or polypeptide* or poly-peptide*)) or 144
(osteoinduct* NEXT (protein* or factor* or polypeptide* or poly-peptide*))
#4 (bmp or bmp2 or bmp-2) 86
#5 (rhbmp or rhbmp2 or rhbmp-2) 69
#6 (rh-bmp or rh-bmp2 or rh-bmp-2) 2
#7 (infuse or amplify) 4780
#8 (#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7) 4964
145 Appendix II – Search strategies
#9 MeSH descriptor Spinal Fusion, this term only 553
#10 (spine or spinal) 13602
#11 (spondylosyndes*) 0
#12 (spondylodes*) 21
#13 ("lumbar interbody arthrodesis") 5
((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or #14 606 posterolateral) NEAR/3 fusion*)
#15 ("fusion cage") 31
#16 (#9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15) 13647
#17 (#8 AND #16) 218
Database: Science Citation Index Expanded (SCI-EXPANDED) 1899-present Database interface: Web of Science (ISI) on Web of Knowledge Search date: 22nd November 2011 Records identified: 1302 Search strategy:
# 24 1,302 #14 not #23 Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 23 2,703,837 #22 OR #21 OR #20 OR #19 OR #18 OR #17 OR #16 OR #15 Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 22 361,716 Title=(genera or taxonomy or species or fauna or habitat or marine or ecology) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 21 183,260 Title=(cow or cattle or bovine or livestock or swine or poultry) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 20 179,103 Title=(rabbit or rabbits or moss or mosses or fungus or fungi) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
146 Appendix II – Search strategies
# 19 33,263 Title=(fossil or fossils or lichen or lichens or mushroom or mushrooms) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 18 96,171 Title=(bat or bats or bee or bees or grass or grasses or bird or birds or avian) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 17 301,191 Title=(bovine or sheep or fly or flies or fish or fishes or fisheries or horse or horses or equine) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 16 350,781 Title=(animal or animals or dog or dogs or canine or cat or cats or feline) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 15 1,356,848 Title=(rat or rats or mouse or mice or hamster or hamsters) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 14 1,562 #13 AND #6 Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 13 231,771 #12 OR #11 OR #10 OR #9 OR #8 OR #7 Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 12 134 Topic=("fusion cage") Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 11 8,117 Topic=((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) NEAR/3 fusion*) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 10 23 Topic=("lumbar interbody arthrodesis")
147 Appendix II – Search strategies
Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 9 400 Topic=(spondylodes*) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 8 6 Topic=(spondylosyndes*) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 7 229,996 Topic=(spine or spinal) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 6 36,552 #5 OR #4 OR #3 OR #2 OR #1 Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 5 16,891 Topic=(infuse or amplify) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 4 80 Topic=(rh-bmp or rh-bmp2 or rh-bmp-2) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 3 1,534 Topic=(rhbmp or rhbmp2 or rhbmp-2) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 2 12,462 Topic=(bmp or bmp2 or bmp-2) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
# 1 14,240 Topic=(("bone morphogen*" or osteogen* or osteoinduct*) NEAR/1 (protein* or factor* or polypeptide* or "poly-peptide*")) Databases=SCI-EXPANDED Timespan=All Years Lemmatization=Off
148 Appendix II – Search strategies
Database PubMED Database interface: http://www.ncbi.nlm.nih.gov/sites/entrez Search date: 22nd November 2011 Records identified: 1176 Search strategy: #38 Search #30 AND #37 07:04:25 1176 #37 Search #31 OR #32 OR #33 OR #34 OR #35 OR #36 07:04:06 312561 #36 Search fusion cage [tw] 07:03:37 124 #35 Search lumbar fusion [tw] OR cervical fusion [tw] OR posterior fusion [tw] OR anterior fusion [tw] OR lumbosacral fusion [tw] OR transforminal fusion [tw] OR posterolateral fusion [tw] 07:03:19 3193 #34 Search lumbar interbody arthrodesis [tw] 07:02:55 23 #33 Search spondylosyndes* [tw] OR spondylodes* [tw] 07:02:41 604 #32 Search spine [tw] OR spinal [tw] 07:02:27 312301 #31 Search spinal fusion[MeSH Terms] 07:02:07 14187 #30 Search #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 07:01:48 33173 #29 Search infuse [tw] OR amplify [tw] 07:00:31 15366 #28 Search rh-bmp [tw] OR rh-bmp2 [tw] OR rh-bmp-2[tw] 07:00:16 74 #27 Search rhbmp [tw] OR rhbmp2 [tw] OR rhbmp-2 [tw] 07:00:01 1244 #26 Search bmp [tw] OR bmp2 [tw] OR bmp-2[tw] 06:59:48 10745 #25 Search osteoinduct* protein*[tw] 06:59:30 850 #24 Search osteoinduct* factor*[tw] 06:59:09 625 #23 Search osteoinduct* polypeptide*[tw] 06:58:53 11 #22 Search osteoinduct* poly-peptide*[tw] 06:58:38 0 #21 Search osteogen protein*[tw] 06:58:24 1420 #20 Search osteogen factor*[tw] 06:58:11 483 #19 Search osteogen polypeptide*[tw] 06:57:53 121 #18 Search osteogen poly-peptide*[tw] 06:57:35 1 #17 Search bone osteogen protein*[tw] 06:57:18 486 #16 Search bone osteogen factor*[tw] 06:57:00 299 #15 Search bone osteogen polypeptide*[tw] 06:56:40 11 #14 Search bone osteogen poly-peptide*[tw] 06:56:20 0 #13 Search bone morphogen* protein*[tw] 06:56:02 12902 #12 Search bone morphogen* factor*[tw] 06:55:47 8714 #11 Search bone morphogen* polypeptide*[tw] 06:55:32 153 #10 Search bone morphogen* poly-peptide*[tw] 06:55:20 0 #9 Search bone morphogenetic protein 2[MeSH Terms] 06:55:01 3432 #8 Search bone morphogenetic proteins[MeSH Terms] 06:54:39 11074
149 Appendix II – Search strategies
Database: DARE (Database of Abstracts of Reviews of Effects) Database interface: John Wiley Search date: 24th November 2011 Records identified: 6 records Search Strategy: same strategy as used for CENTRAL above
Database: HTA (Health Technology Assessment) Database interface: John Wiley Search date: 24th November 2011 Records identified: 7 records Search Strategy: same strategy as used for CENTRAL above
Database: BIOSIS Previews 1969 to 2008 Database interface: Web of Science Search date: 24th November 2011 Records identified: 1503 Search Strategy:
# 14 1,503 #13 AND #6 Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 13 192,365 #12 OR #11 OR #10 OR #9 OR #8 OR #7 Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 12 101 TS=("fusion cage") Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 11 3,585 TS=((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) NEAR/3 fusion*) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 10 11 TS=("lumbar interbody arthrodesis") Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
150 Appendix II – Search strategies
# 9 224 TS=(spondylodes*) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 8 2 TS=(spondylosyndes*) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 7 191,278 TS=(spine or spinal) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 6 135,050 #5 OR #4 OR #3 OR #2 OR #1 Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 5 122,315 TS=(infuse or amplify) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 4 23 TS=(rh-bmp or rh-bmp2 or rh-bmp-2) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 3 564 TS=(rhbmp or rhbmp2 or rhbmp-2) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 2 9,009 TS=(bmp or bmp2 or bmp-2) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
# 1 9,886 TS=(("bone morphogen*" or osteogen* or osteoinduct*) NEAR/1 (protein* or factor* or polypeptide* or "poly-peptide*")) Databases=BIOSIS Previews Timespan=All Years Lemmatization=On
Database: BIOSIS Previews, 2009 to 2011 Database interface: Dialog Classic Search date: 24th November 2011 Records identified: 226
151 Appendix II – Search strategies
Search Strategy: s bone(w)morphogenetic(w)protein?/ti,ab,de s bone(w)morphogenetic(w)protein-2/ti,ab,de s ((bone(w)morphogen?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de s ((osteogen?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de s ((osteoinduct?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de s (bmp or bmp2 or bmp-2)/ti,ab,de s (rhbmp or rhbmp2 or rhbmp-2)/ti,ab,de s (rh-bmp or rh-bmp2 or rh-bmp-2)/ti,ab,de s (infuse or amplify)/ti,ab,de s s1:s9 s (Spine or spinal)/ti,ab,de s spondylosyndes?/ti,ab,de s spondylodes?/ti,ab,de s lumbar(w)interbody(w)arthrodesis/ti,ab,de s ((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) (3w)(fusion?))/ti,ab,de s (fusion(w)cage)/ti,ab,de s s11:s16 s s10 AND s17
Database: Toxfile, 1964 onwards Database interface: Dialog Classic Search date: 25th November 2011 Records identified: 204 Search Strategy:
File 156:ToxFile 1965-2011/Nov W2 (c) format only 2011 Dialog
? s bone(w)morphogenetic(w)protein?/ti,ab,de 101939 BONE/TI,AB,DE 3069 MORPHOGENETIC/TI,AB,DE 578385 PROTEIN?/TI,AB,DE S1 2202 BONE(W)MORPHOGENETIC(W)PROTEIN?/TI,AB,DE
? s bone(w)morphogenetic(w)protein-2/ti,ab,de 101939 BONE/TI,AB,DE 3069 MORPHOGENETIC/TI,AB,DE
152 Appendix II – Search strategies
0 PROTEIN-2/TI,AB,DE S2 0 BONE(W)MORPHOGENETIC(W)PROTEIN-2/TI,AB,DE
? s ((bone(w)morphogen?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de 101939 BONE/TI,AB,DE 8273 MORPHOGEN?/TI,AB,DE 578385 PROTEIN?/TI,AB,DE 737187 FACTOR?/TI,AB,DE 16000 POLYPEPTIDE?/TI,AB,DE 0 POLY-PEPTIDE?/TI,AB,DE S3 2240 ((BONE(W)MORPHOGEN?) (W) (PROTEIN? OR FACTOR? OR POLYPEPTIDE? OR POLY-PEPTIDE?))/TI,AB,DE
? s ((osteogen?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de 3820 OSTEOGEN?/TI,AB,DE 578385 PROTEIN?/TI,AB,DE 737187 FACTOR?/TI,AB,DE 16000 POLYPEPTIDE?/TI,AB,DE 0 POLY-PEPTIDE?/TI,AB,DE S4 123 ((OSTEOGEN?) (W) (PROTEIN? OR FACTOR? OR POLYPEPTIDE? OR POLY-PEPTIDE?))/TI,AB,DE
? s ((osteoinduct?) (w) (protein? or factor? or polypeptide? or poly-peptide?))/ti,ab,de 215 OSTEOINDUCT?/TI,AB,DE 578385 PROTEIN?/TI,AB,DE 737187 FACTOR?/TI,AB,DE 16000 POLYPEPTIDE?/TI,AB,DE 0 POLY-PEPTIDE?/TI,AB,DE S5 26 ((OSTEOINDUCT?) (W) (PROTEIN? OR FACTOR? OR POLYPEPTIDE? OR POLY-PEPTIDE?))/TI,AB,DE
? s (bmp or bmp2 or bmp-2)/ti,ab,de 1705 BMP/TI,AB,DE 196 BMP2/TI,AB,DE 0 BMP-2/TI,AB,DE S6 1816 (BMP OR BMP2 OR BMP-2)/TI,AB,DE
? s (rhbmp or rhbmp2 or rhbmp-2)/ti,ab,de 212 RHBMP/TI,AB,DE 10 RHBMP2/TI,AB,DE
153 Appendix II – Search strategies
0 RHBMP-2/TI,AB,DE S7 220 (RHBMP OR RHBMP2 OR RHBMP-2)/TI,AB,DE
? s (rh-bmp or rh-bmp2 or rh-bmp-2)/ti,ab,de 0 RH-BMP/TI,AB,DE 0 RH-BMP2/TI,AB,DE 0 RH-BMP-2/TI,AB,DE S8 0 (RH-BMP OR RH-BMP2 OR RH-BMP-2)/TI,AB,DE
? s (infuse or amplify)/ti,ab,de 209 INFUSE/TI,AB,DE 2293 AMPLIFY/TI,AB,DE S9 2501 (INFUSE OR AMPLIFY)/TI,AB,DE
? s s1:s9 S10 5201 S1:S9
? s (Spine or spinal)/ti,ab,de 6037 SPINE/TI,AB,DE 32780 SPINAL/TI,AB,DE S11 36826 (SPINE OR SPINAL)/TI,AB,DE
? s spondylosyndes?/ti,ab,de S12 0 SPONDYLOSYNDES?/TI,AB,DE
? s spondylodes?/ti,ab,de S13 11 SPONDYLODES?/TI,AB,DE
? s lumbar(w)interbody(w)arthrodesis/ti,ab,de 7922 LUMBAR/TI,AB,DE 96 INTERBODY/TI,AB,DE 179 ARTHRODESIS/TI,AB,DE S14 1 LUMBAR(W)INTERBODY(W)ARTHRODESIS/TI,AB,DE
? s ((lumbar or cervical or posterior or anterior or lumbosacral or transforminal or posterolateral) (3w)(fusion?))/ti,ab,de 7922 LUMBAR/TI,AB,DE 20388 CERVICAL/TI,AB,DE 11470 POSTERIOR/TI,AB,DE 18967 ANTERIOR/TI,AB,DE
154 Appendix II – Search strategies
1439 LUMBOSACRAL/TI,AB,DE 0 TRANSFORMINAL/TI,AB,DE 243 POSTEROLATERAL/TI,AB,DE 37465 FUSION?/TI,AB,DE S15 257 ((LUMBAR OR CERVICAL OR POSTERIOR OR ANTERIOR OR LUMBOSACRAL OR TRANSFORMINAL OR POSTEROLATERAL) (3W)(FUSION?))/TI,AB,DE
? s (fusion(w)cage)/ti,ab,de 36237 FUSION/TI,AB,DE 3129 CAGE/TI,AB,DE S16 4 (FUSION(W)CAGE)/TI,AB,DE
? s s11:s16 S17 36854 S11:S16
? s s10 AND s17 5201 S10 36854 S17 S18 204 S10 AND S17
155 Appendix III – Call for evidence
Appendix III
Systematic review of bone morphogenic protein-2 (rhBMP-2) for spinal fusion - call for evidence
The Centre for Reviews and Dissemination (CRD) is undertaking a systematic review and individual participant data (IPD) meta-analysis of the comparative effectiveness of rhBMP-2 (marketed as INFUSE) for spinal fusion. The review has been commissioned by the Yale University Open Data Access (YODA) initiative as part of an overarching project to systematically review the safety and effectiveness of rhBMP-2, including re-analysis of IPD that have been made available to Yale on an unrestricted basis by the manufacturer (Medtronic Inc). YODA aims to improve access to patient-level data from clinical trials, and provide independent, scientifically rigorous, objective and fair analyses of such data.
CRD will undertake a comprehensive and rigorous systematic review and meta-analysis of individual participant data (IPD) of all relevant randomised controlled trials that have compared rhBMP-2 with standard bone graft therapy.
We will include all relevant randomised controlled trials, irrespective of whether conducted by the manufacturer or not, and irrespective of whether published or not.
We are therefore interested in hearing from anyone who has conducted, or is aware of, unpublished or partially published research in this area. For example, trials which have been presented at conferences but not fully reported elsewhere.
We are currently aware of 17 trials funded by the manufacturer and have searched the published literature but welcome any information regarding further unpublished research. If you know of any such trials please contact Mark Rodgers on behalf of the York Spinal Fusion Project by email: crd- [email protected] or phone: +44 (0)1904 321086.
Link to CRD project page http://www.york.ac.uk/inst/crd/projects_in_progress.cfm Link to YODA page http://medicine.yale.edu/core/projects/yodap/index.aspx
156 Appendix IV – Screening criteria
Appendix IV
Full text screening selection criteria
Effectiveness review P: Humans (any age) undergoing spinal fusion (any approach) for degenerative disc disease or spondylolisthesis I: use of rhBMP-2 (INFUSE or AMPLIFY, any method) C: Standard bone graft therapy (SBGT) O: Any outcome S: Randomised controlled trials
Adverse events review P: Humans (any age) undergoing spinal fusion (any approach) for degenerative disc disease or spondylolisthesis I: use of rhBMP-2 (INFUSE or AMPLIFY, any method) C: any comparator or no comparator O: Any adverse events outcome S: Controlled trials (randomised and non-randomised), observational studies, reviews of medical records, case studies, post-marketing event reports
Code studies in Custom Field 6 as: EXCLUDE for studies to be excluded INCLUDE for studies to be included in effectiveness AND adverse events review AE ONLY for studies to be included in AE review only BACKGROUND for background reading SR for any systematic reviews EE for any economic evaluations LETTER for any letters to editors (indicate, if applicable, which paper the letter refers to, using the EndNote library number)
157 Appendix V– RCT data extraction
Appendix V
Items extracted from published reports of Medtronic and other RCTs used to investigate the reliability of the published evidence
Trial Trial identifier Internal trial identifier Medtronic trial name The name used by Medtronic to refer to the trial where relevant (e.g INFUSE/LT-CAGE Open Pivotal RCT) Endnote no. Insert publication number but leave out “#” Author First author only Year Year of publication Country Extracted/checked by Insert your initials once you have completed extracting/checking the extraction for this publication
Participants Inclusion criteria Describe the participant inclusion criteria Exclusion criteria Describe the participant exclusion criteria Diagnosis of included pts: n/N (%) Describe the proportion of different diagnoses among the actual included participants e.g. 123/123 DDD (100%) N randomised Mean (sd) age If measure of variability is not reported, insert empty parentheses Male: n/N (%) Tobacco use: n/N (%) Weight mean in lbs (sd) Most studies appear to report in lbs Workers compensation n/N (%) Litigation n/N (%) Working n/N (%) COMMENTS Any comments on patient selection/demographics not covered above Interventions/comparators Surgical approach Surgical approach used in both groups e.g. PLF, PLIF, ALIF, ACDF. Also state if single-level, multi-level or both. Open/laparoscopic. rhBMP-2 intervention rhBMP-2 concentration in mg/cc (e.g. 1.5mg/cc), carrier (e.g. ACS, BCP, CRM), +/- allograft or autograft fixation/instrumentation (e.g. LT-CAGE, CD HORIZON)
158 Appendix V– RCT data extraction
rhBMP-2 dose Give range for total rhBMP-2 dose in mg (e.g. 3.9-7.8mg). Note: this is different to concentration (e.g. 1.5mg/mL is the standard INFUSE concentration) Comparator intervention Comparator (e.g. ICBG), Any additional intervention (e.g. allograft), fixation/instrumentation (e.g. LT-CAGE, CD HORIZON) Outcomes – for each time point (Post-surgery (PS), 6 weeks, 3, 6, 12, 24 months) n evaluated Number of patients in evaluated at each follow-up point, where stated by the study authors. This may not be clear or vary by outcome measure, hence why extraction of each dichotomous outcome is in the form “n/N”. Oswestry Disability Index (mean, Mean Oswestry Low Back Pain disability score for each SD) group. If measure of variability is not reported, insert empty parentheses “()”. Neck Disability Index Only likely to be reported in ACDF studies. SF-36 Physical Component Mean (SD). If the PCS is not reported, mark “*” and describe Summary score (PCS; mean, SD) the SF-36 domains that were reported. If measure of variability is not reported, insert empty parentheses “()”. Back/neck pain (mean, SD) Mean (SD) score out of 20. If a different scoring system is used, mark “*” and describe in comments box. Back pain in lumbar surgery, neck pain in cervical surgery. If measure of variability is not reported, insert empty parentheses “()”. Leg pain (mean, SD) Mean (SD) score out of 20. If a different scoring system is used, mark with “*” and describe in comments box. If measure of variability is not reported, insert empty parentheses “()”. Neurological success (n/N) Number of patients considered “Neurologic success” / Number of patients assessed for this outcome Donor site pain (mean, SD) Mean (SD) score out of 20. Only applies to ICBG groups. If a different scoring system is used, mark with “*” and describe in comments box. If measure of variability is not reported, insert empty parentheses “()”. Return to work/activity (n/N) Number of patients working / Number of patients assessed for this outcome Fusion (n/N) Number of patients considered fusion “success” / Number of patients evaluated for this outcome 2nd Surgical procedure (n/N) Number of patients undergoing 2nd procedure for any reason / Number of patients evaluated for this outcome Operating time (hours) Mean (SD) in hours Duration of stay (days) Mean (SD) in days Total no. AEs reported (n) If not clear at what point in time AEs occurred, just report this number for the last follow-up point in the trial (e.g. total number of AEs in trial A are reported on the 24 month worksheet) Heterotopic bone formation (n/N) Osteolysis (n/N)
159 Appendix V– RCT data extraction
Infection (n/N) Neurologic events (new/worse leg Mark with * and give specific details in comments box if pain, sensory disturbance, reflex available. changes, bladder disturbance) (n/N) Cancer (n/N) Hardware failure (cage subsidence, implant breakage) (n/N) Vascular events (n/N) Absolute number of vascular events reported / Number of patients assessed for this outcome Retrograde ejaculation (n/N) Number of events / number of male participants for each group (e.g. 6/78) Urinary retention (n/N) Dysphagia (n/N) Airway obstruction (n/N) Neck pain events (n/N) Recurrent laryngeal nerve palsy (n/N) Leg pain/radiculitis (n/N) Leg weakness (n/N) Inflammatory cyst formation (n/N) Other reported AEs Select “probably unrelated to rhBMP-2” or “possibly related to rhBMP-2” from AE categorisation document Other reported outcomes Simply name the other recorded outcomes COMMENTS ON OUTCOMES Risk of bias Risk of bias 1 – fusion assessment Risk of bias 2 – Patient centred subjective measures Risk of bias 3 – adverse events Random sequence generation Info from the comments field of the preceding worksheet may be used to inform this assessment. Allocation concealment
Blinding of participants and personnel Use RoB guidance to help extraction Blinding of outcome assessment Incomplete outcome data Selective reporting Other sources of bias External validity issues Other Funding source Medtronic or other Comments
160 Appendix VI – Wider literature data extraction
Appendix VI
Items extracted from publications of non-randomised studies and RCTs with comparators other than ICBG which reported adverse events Study identifier First author, year of publication Country Funding Funding not reported (i.e., funding not mentioned in paper) Industry link reported No industry link reported rhBMP-2 group Dose of rhBMP-2 received and number of participants Control group Control treatment received and number of participants Surgery Spinal location of surgery (ALIF, TLIF, various etc.) Design/Setting Design as described by study authors Mode of participant allocation to rhBMP-2/control arm Purpose of the study Patient details diagnosis AE defined Did the paper provide the authors’ definition of “adverse events” or “complications” – yes/no Follow-up Duration of follow-up in months Total AEs reported Total number of adverse events, if available Specific AEs n/N (%) for any pre-specified adverse events according to spinal location of surgery
161 Appendix VII – Risk of bias tool
Appendix VII
Risk of Bias tool assessment criteria
Domain Description Review authors’ judgment Selection bias: Selection bias (biased allocation to Describe the method used to generate the allocation sequence in sufficient detail to Random sequence interventions) due to inadequate allow an assessment of whether it should produce comparable groups. generation generation of a randomised sequence. Selection bias (biased allocation to Describe the method used to conceal the allocation sequence in sufficient detail to Selection bias: interventions) due to inadequate determine whether intervention allocations could have been foreseen in advance of, or Allocation concealment concealment of allocations prior to during, enrolment. assignment. Performance bias due to knowledge of Performance bias: Describe all measures used, if any, to blind study participants and personnel from the allocated interventions by Blinding of participants knowledge of which intervention a participant received. Provide any information participants and personnel during the and personnel relating to whether the intended blinding was effective. study. Detection bias: Describe all measures used, if any, to blind outcome assessors from knowledge of Detection bias due to knowledge of the Blinding of outcome which intervention a participant received. Provide any information relating to whether allocated interventions by outcome assessment the intended blinding was effective. assessors. Describe the completeness of outcome data for each main outcome, including attrition and exclusions from the analysis. State whether attrition and exclusions were reported, Attrition bias: Attrition bias due to amount, nature or the numbers in each intervention group (compared with total randomized participants), Incomplete outcome data handling of incomplete outcome data. reasons for attrition/exclusions where reported, and any re-inclusions in analyses performed by the review authors. Reporting bias: State how the possibility of selective outcome reporting was examined by the review Reporting bias due to selective Selective reporting authors, and what was found. outcome reporting. State any important concerns about bias not addressed in the other domains in the Other bias: tool. Bias due to problems not covered Other sources of bias If particular questions/entries were pre-specified in the review’s protocol, responses elsewhere in the table. should be provided for each question/entry
162 Appendix VII – Risk of bias tool
RANDOM SEQUENCE GENERATION Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence Spinal fusion-specific considerations Low: If using computer randomisation Unclear: If method not described N/A: Single-arm study The investigators describe a random component in the sequence generation process such as: Referring to a random number table; Using a computer random number generator; Coin tossing; Criteria for a judgement of ‘Low risk’ of bias. Shuffling cards or envelopes; Throwing dice; Drawing of lots; Minimization - Minimization may be implemented without a random element, and this is considered to be equivalent to being random.
163 Appendix VII – Risk of bias tool
The investigators describe a non-random component in the sequence generation process. Usually, the description would involve some systematic, non-random approach, for example: Sequence generated by odd or even date of birth; Sequence generated by some rule based on date (or day) of admission; Sequence generated by some rule based on hospital or clinic record number. Criteria for the judgement of ‘High Other non-random approaches happen much less frequently than the systematic approaches mentioned above and risk’ of bias. tend to be obvious. They usually involve judgement or some method of non-random categorization of participants, for example: Allocation by judgement of the clinician; Allocation by preference of the participant; Allocation based on the results of a laboratory test or a series of tests; Allocation by availability of the intervention. Criteria for the judgement of Insufficient information about the sequence generation process to permit judgement of ‘Low risk’ or ‘High risk’. ‘Unclear risk’ of bias.
ALLOCATION CONCEALMENT Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment Spinal fusion-specific considerations Low: Envelopes explicitly sealed, opaque and sequential. Unclear: method not described Criteria for a judgement of ‘Low Participants and investigators enrolling participants could not foresee assignment because one of the following, or an risk’ of bias. equivalent method, was used to conceal allocation: Central allocation (including telephone, web-based and pharmacy-controlled randomization); Sequentially numbered drug containers of identical appearance; Sequentially numbered, opaque, sealed envelopes.
164 Appendix VII – Risk of bias tool
Criteria for the judgement of ‘High Participants or investigators enrolling participants could possibly foresee assignments and thus introduce selection bias, risk’ of bias. such as allocation based on: Using an open random allocation schedule (e.g. a list of random numbers); Assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or nonopaque or not sequentially numbered); Alternation or rotation; Date of birth; Case record number; Any other explicitly unconcealed procedure. Criteria for the judgement of Insufficient information to permit judgement of ‘Low risk’ or ‘High risk’. This is usually the case if the method of ‘Unclear risk’ of bias. concealment is not described or not described in sufficient detail to allow a definite judgement – for example if the use of assignment envelopes is described, but it remains unclear whether envelopes were sequentially numbered, opaque and sealed.
BLINDING OF PARTICIPANTS AND PERSONNEL Performance bias due to knowledge of the allocated interventions by participants and personnel during the study Spinal fusion-specific considerations Participants and personnel may be made aware of allocation in studies using ICBG as a comparator due to the presence of an incision/pain at the graft harvesting site. In theory blinding might be at least partially achieved by a sham incision at the same site in rhBMP-2 patients, though this is unlikely to have occurred. They may also be made aware of allocation by differences in the surgical approach (e.g. minimally invasive vs. open). Fusion: Blinding of participants/personnel unlikely to cause performance bias leading to an impact on fusion outcome, so low risk of bias. Patient-reported: High if unblinded. AEs: High if unblinded for adverse events. Unclear if blinding not mentioned, regardless of outcome. Criteria for a judgement of ‘Low Any one of the following: risk’ of bias. No blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding; Blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken.
165 Appendix VII – Risk of bias tool
Criteria for the judgement of ‘High Any one of the following: risk’ of bias. No blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding; Blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome is likely to be influenced by lack of blinding. Criteria for the judgement of Any one of the following: ‘Unclear risk’ of bias. Insufficient information to permit judgement of ‘Low risk’ or ‘High risk’; The study did not address this outcome. BLINDING OF OUTCOME ASSESSMENT Detection bias due to knowledge of the allocated interventions by outcome assessors Spinal fusion-specific considerations Fusion success: Low risk if (1) differences between rhBMP and bone graft could not be discerned on imaging (2) evaluation of radiographs/CTs was independent of any direct patient examination and assessors were explicitly stated to be blinded or masked.
Patient-assessed subjective outcomes (Oswestry, SF-36 etc): In this case, the patient is the outcome assessor. If they have not been blinded to allocation, then this should be marked as ‘high’ risk. Clinician/investigator-assessed subjective outcomes (Global improvement scores, adverse events?): In this case, the clinician/investigator is the outcome assessor. If they have not been blinded to allocation, then this should be marked as ‘high’ risk.
Fusion: Low if blinded assessors. High if unblinded. Unclear if nothing reported. This depends on rhBMP being indistinguishable from ICBG on imaging. Patient: As the patient is the outcome assessor, same as previous question i.e. high if unblinded. AEs: High if both unblinded. Low if both blinded. Though often unclear who is the AE assessor. Criteria for a judgement of ‘Low Any one of the following: risk’ of bias. No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; Blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
166 Appendix VII – Risk of bias tool
Criteria for the judgement of ‘High Any one of the following: risk’ of bias. No blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; Blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding. Criteria for the judgement of Any one of the following: ‘Unclear risk’ of bias. Insufficient information to permit judgement of ‘Low risk’ or ‘High risk’; The study did not address this outcome. INCOMPLETE OUTCOME DATA Attrition bias due to amount, nature or handling of incomplete outcome data Spinal fusion-specific considerations In justification section insert % FU in each arm plus reasons (or that reasons were not reported) n/a: protocols Unclear: if loss to FU not reported, or if was reported but reasons not given. Low risk: Where numbers lost were very small (<5%) and balanced and reasons were reported Criteria for a judgement of ‘Low Any one of the following: risk’ of bias. No missing outcome data; Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; For continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; Missing data have been imputed using appropriate methods.
167 Appendix VII – Risk of bias tool
Criteria for the judgement of ‘High Any one of the following: risk’ of bias. Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; For continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; ‘As-treated’ analysis done with substantial departure of the intervention received from that assigned at randomization; Potentially inappropriate application of simple imputation. Criteria for the judgement of Any one of the following: ‘Unclear risk’ of bias. Insufficient reporting of attrition/exclusions to permit judgement of ‘Low risk’ or ‘High risk’ (e.g. number randomized not stated, no reasons for missing data provided); The study did not address this outcome. SELECTIVE REPORTING Reporting bias due to selective outcome reporting Fusion: If fusion rates reported = Low bias. If they collected data >12 months but did not report >12 month fusion = high risk.
Patient reported: If they report pain and function outcomes, rate as “Low risk”. If any of these are not reported, rate as High risk. If only p values were reported for these outcomes, also rate as high risk. AEs: If collection of AEs is mentioned in the protocol but not at all in the paper (i.e. not even “There were no adverse events”) then = High. Otherwise rate as unclear. Criteria for a judgement of ‘Low Any of the following: risk’ of bias. The study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way; The study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre-specified (convincing text of this nature may be uncommon).
168 Appendix VII – Risk of bias tool
Criteria for the judgement of ‘High Any one of the following: risk’ of bias. Not all of the study’s pre-specified primary outcomes have been reported; One or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. subscales) that were not pre-specified; One or more reported primary outcomes were not pre-specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); One or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta- analysis; The study report fails to include results for a key outcome that would be expected to have been reported for such a study. Criteria for the judgement of Insufficient information to permit judgement of ‘Low risk’ or ‘High risk’. It is likely that the majority of studies will fall into ‘Unclear risk’ of bias. this category.
169 Appendix VIII – Classification of adverse events
Appendix VIII
Classification of adverse events
Category (as defined in Events as described in study publications our protocol) Non-union non-union (failure) non-union (outcome pending) 2nd surgical procedure post-surgical ER visits any mention of revision surgery, explorative second surgery, repeat fusion, evacuation, removal of instrumentation etc. Heterotopic bone Hyperostosis formation Osteolysis wound complications* Infection lumbar hematoma/hematoma other graft site complication perioperative swelling as defined prolonged drainage wound dehiscence Neurologic events no improvement in neurologic deficit numbness/SLE paresis spinal event - at cervical/thoracic/lumbar/target level Cancer Hardware failure fatigue failure of screws implant loosening or displacement screw lucency Vascular events Retrograde ejaculation perioperative swelling as defined Urinary retention Dysphagia oesophageal dilation required Airway obstruction respiratory in ACDF Neck pain Recurrent laryngeal nerve palsy
170 Appendix VIII – Classification of adverse events
Leg pain/radiculitis leg pain neural injury neurological deficit neuropathic pain non-mechanical leg pain perioperative swelling as defined Leg weakness neural injury neurological deficit Inflammatory cyst lumbar seroma formation perioperative swelling as defined Other – potentially adjacent disc herniation rhBMP-2 related allergic reaction back pain back pain/and or leg pain deep vein thrombosis / pulmonary embolism dyspnea elevated temperature hoarseness lower extremity pain neck and/or arm pain (esp if in cervical) neck swelling other pain percutaneous endoscopic gastrostomy/feeding tubes placed (e.g. as a consequence of ileus) peritoneal tear respiratory distress respiratory failure shoulder pain tracheotomies unplanned intubation upper extremity pain
171 Appendix VIII – Classification of adverse events
Other – probably “other” unrelated to rhBMP-2 acute tubular necrosis anatomic/technical difficulty arthritis/bursitis atrial fibrillation bowel perforation cardiac problems cardiac/pulmonary/stroke complications cardiovascular carpal tunnel syndrome death dural tear / CSF leak / intra-operative dural rent dysphasia GI problems headache hyponatremia ileus malpositioned implant medical complications (this means things like DVT, PE, chest infection, cardiac problems) pancreatitis pneumonia rectal bleeding/haemorrhoids respiratory (in a lumbar trial) Retained drain subsidence trauma urogenital urologic injuries vascular intra-op vertebral fracture *The category “wound complications” was added during the data extraction process following consultation with the clinical co-author.
172 Appendix IX – Characteristics of excluded papers
Appendix IX
Characteristics of excluded papers Economic Evaluations Reference Title Ackermann 2002 164 Economic evaluation of bone morphogenetic protein versus autogenous iliac crest bone graft in single-level anterior lumbar fusion: an evidence-based modeling approach Alt 2009 41 An economic analysis of using rhBMP-2 for lumbar fusion in Germany, France and UK from a societal perspective Carreon 2009 165 RhBMP-2 versus iliac crest bone graft for lumbar spine fusion in patients over 60 years of age: a cost-utility study Chhabra 2006 38 A health-economic evaluation of rhBMP-2 in spine fusion surgery in Germany and UK Chhabra 2007 39 A cost-effectiveness analysis of rhBMP-2 in spine fusion surgery in Sweden and Denmark Glassman 2008 166 The perioperative cost of Infuse bone graft in posterolateral lumbar spine fusion McInnis 2010 103 Budget impact of new rhbmp-2 formulation in patients undergoing posterolateral spinal fusion procedures for degenerative disc disease in randomized controlled trial (RCT) Polly 2003 167 A cost analysis of bone morphogenetic protein versus autogenous iliac crest bone graft in single-level anterior lumbar fusion Van Genugten 2008 40 A cost-effectiveness analysis of RhBMP-2 in spine fusion surgery in the Netherlands
Systematic Reviews Reference Title Abdullah 2011 168 The state of lumbar fusion extenders Agarwal 2009 169 Osteoinductive bone graft substitutes for lumbar fusion: a systematic review Burkus 2004 170 The effectiveness of rhBMP-2 in replacing autograft: an integrated analysis of three human spine studies
173 Appendix IX – Characteristics of excluded papers
Carragee 2011 171 Comparison of adverse events and disclosures in the original rhBMP-2 trials with fda data and subsequent publications Carragee 2011 11 A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned ECRI 2004 172 Interbody cage with bone morphogenetic protein (InFUSE(TM)/LT-CAGE(TM) for degenerative disc disease Fenton 2007 173 Variation in reported safety of lumbar interbody fusion - Influence of industrial sponsorship and other study characteristics Garrison 2007 174 Clinical effectiveness and cost-effectiveness of bone morphogenetic proteins in the non- healing of fractures and spinal fusion: a systematic review Guerado 2011 175 What bone graft substitutes should we use in post-traumatic spinal fusion? Hayes Inc. 2006 176 Recombinant human bone morphogenetic protein for use in spinal fusion Hsu 2008 177 The use of bone morphogenetic protein in spine fusion Kanakaris 2008 178 Clinical applications of bone morphogenetic proteins: current evidence McKay 2007 179 A comprehensive clinical review of recombinant human bone morphogenetic protein-2 (INFUSE Bone Graft) Menzin 2012 180 A systematic review of the use of patient-centered outcome measures in studies of spinal fusion surgery for degenerative disc disease: implications for comparative effectiveness research Miyazaki 2009 181 An update on bone substitutes for spinal fusion Mroz 2010 182 Complications related to osteobiologics use in spine surgery: a systematic review Mussano 2007 183 Bone morphogenetic proteins and bone defects: a systematic review Papakostidis 2008 184 Efficacy of autologous iliac crest bone graft and bone morphogenetic proteins for posterolateral fusion of lumbar spine: a meta-analysis of the results Poynton 2002 185 Safety profile for the clinical use of bone morphogenetic proteins in the spine Smoljanovic 2007 186 Regeneration of the skeleton by recombinant human bone morphogenetic proteins Than 2011 187 Complication avoidance and management in anterior lumbar interbody fusion Thawani 2010 188 Bone morphogenetic proteins and cancer: review of the literature
174 Appendix IX – Characteristics of excluded papers
Vaz 2010 189 Bone grafting options for lumbar spine surgery: a review examining clinical efficacy and complications Wong 2009 190 Is there a role for BMPs in spinal fusion surgery Wu 2010 191 Minimal access versus open transforaminal lumbar interbody fusion: meta-analysis of fusion rates
Papers excluded due to unmet inclusion criteria Reference Title Reason for exclusion Adogwa 2011 192 Long-term outcomes of revision fusion for lumbar pseudarthrosis: No use of rhBMP-2 clinical article Anderson 2002 193 Summary statement: Clinical BMP programs Not relevant Anonymous 2011 194 A critical eye on infuse use studies Commentary Anoymous 2012195 BMP use up sharply, but outcomes not improved Commentary Aspenberg 2011 196 Under-reported complications related to BMP use in spine surgery Editorial Auerbach 2012 197 Perioperative outcomes, complications, and costs associated with Not focussed on lumbar spinal fusion in older patients with spinal stenosis and use of BMP-2 spondylolisthesis: analysis of the United States Medicare claims database Axelrad 2008 198 Heterotopic ossification after the use of commercially available Irrelevant recombinant human bone morphogenetic proteins in four patients population Axelrad 2009 199 Bone morphogenetic proteins in orthopaedic surgery Survey Benglis 2008 200 A comprehensive review of the safety profile of bone morphogenetic Review protein in spine surgery Beringer 2006 201 Unilateral pedicle screw instrumentation for minimally invasive No use of rhBMP-2 transforaminal lumbar interbody fusion Boden 2002 202 Summary statement: Overview of bone morphogenetic proteins for Commentary spine fusion
175 Appendix IX – Characteristics of excluded papers
Burks 2010 203 Long-term effects of bone morphogenetic protein-based treatments in Review humans Burkus 2004 204 Anterior lumbar interbody fusion for the management of chronic lower No use of rhBMP-2 back pain: current strategies and concepts Burkus 2008 205 Clinical outcomes using rhBMP-2 in spinal fusion applications Review Cabraja 2012206 Bone grafting and substitutes in spine surgery Not focussed on use of BMP-2 Carragee 2009 207 Pseudomorbidity in iliac crest bone graft harvesting: the rise of Editorial rhBMP-2 in short-segment posterior lumbar fusion Chi 2011 208 Exposing conflicts of interest and complications of rhBMP-2 Commentary Cho 2011 209 Spinal reconstruction with pedicle screw-based instrumentation and Irrelevant rhBMP-2 in patients with neurofibromatosis and severe dural ectasia population and spinal deformity: report of two cases and a review of the literature Chrastil 2012210 Complications associated with posterior and transforaminal lumbar Not focussed on interbody fusion use of BMP-2 Crawford 2009 211 Bone morphogenetic protein use in anterior cervical spine surgery: a Review review of current Literature concerning indications, safety, and efficacy Daentzer 2007 212 The efficacy of rhBMP-2 for posterolateral lumbar fusion in smokers: Not relevant Comment. [German] Daniels 2008 213 Adverse events associated with anterior cervical spine surgery Review David 1995 214 Lumbar spinal fusion using rhBMP2: A randomised, blinded and Animal study controlled study Deutsch 2006 215 Minimally invasive transforaminal lumbar interbody fusion with No use of rhBMP-2 unilateral pedicle screw fixation Epstein 2011 216 Pros, cons, and costs of INFUSE in spinal surgery Review Epstein 2011 217 Costs and frequency of "off-label" use of INFUSE for spinal fusions at Frequency analysis one institution in 2010 Glassman 2010 218 Complications and concerns with osteobiologics for spine fusion in Review clinical practice
176 Appendix IX – Characteristics of excluded papers
Hart 2011 219 Acknowledging the elephant in the room: Conflict of interest in Commentary industry-sponsored clinical research Harwood 2005 220 Application of bone morphogenetic proteins in orthopaedic practice: Review their efficacy and side effects Jarrett 2009 221 Anterior exposure of the lumbar spine with and without an "access No use of rhBMP-2 surgeon": morbidity analysis of 265 consecutive cases Jeong 2005 222 Bone morphogenic proteins: applications in spinal surgery Review Kang 2009 223 An analysis of general surgery-related complications in a series of No use of rhBMP-2 412 minilaparotomic anterior lumbosacral procedures Kim 2011 224 BMPs and their clinical potentials Review Kraiwattanapong 2005 225 Comparison of Healos/bone marrow to INFUSE (rhBMP-2/ACS) with Animal study a collagen-ceramic sponge bulking agent as graft substitutes for lumbar spine fusion Lad 2011 226 Trends in the use of bone morphogenetic protein as a substitute to Frequency analysis autologous iliac crest bone grafting for spinal fusion procedures in the United States Lissenberg-Thunnissen 2011 227 Use and efficacy of bone morphogenetic proteins in fracture healing Review McKay 2008 228 Development of a novel compression-resistant carrier for recombinant Product description human bone morphogenetic protein-2 (rhBMP-2) and preliminary clinical results McKay 2009 229 Local sustained delivery of recombinant human bone morphogenetic Product description protein-2 (rhBMP-2) Mehta 2010 230 Retrospective comparative analysis of degenerative lumbar Use of rhBMP-7 spondylolisthesis treated with posterolateral fusion (PLF) with bone morphogenic protein (BMP) 2 with or without additional transforaminal lumbar interbody fusion (TLIF) Mirza 2011 231 Commentary: Folly of FDA-approval studies for bone morphogenetic Commentary protein Niu 2010 232 Outcomes of interbody fusion cages used in 1 and 2-levels anterior No use of rhBMP-2 cervical discectomy and fusion titanium cages versus polyetheretherketone (PEEK) cages
177 Appendix IX – Characteristics of excluded papers
Ong 2010 8 Off-label use of bone morphogenetic proteins in the United States Epidemiological using administrative data study Potter 2005 233 Transforaminal lumbar interbody fusion: clinical and radiographic No use of rhBMP-2 results and complications in 100 consecutive patients Rihn 2009 234 The use of bone morphogenetic protein in lumbar spine surgery Review Robinson 2008 235 Evidence supporting the use of bone morphogenetic proteins for Review spinal fusion surgery Rowan 2010 236 RhBMP-2 use in lumbar fusion surgery is associated with better leg Conference listing pain scores at late follow up Salisbury 2011 237 Adverse events and bone morphogenetic protein-2 Commentary Sandhu 2002 238 Summary statement: Safety of bone morphogenetic proteins for spine Not relevant fusion Sandhu 2003 239 Recombinant human bone morphogenetic protein-2: use in spinal Review fusion applications Sass 2009 240 Autologous bone or bone substitutes in spinal fusion [German] Not relevant Schultz 2008 241 FDA Public Health Notification: Life-threatening complications Product description associated with recombinant human bone morphogenetic protein in cervical spine fusion Solanki 2008 242 Progressive spinal instability in MPS. Use of recombinant human Irrelevant bone morphogenetic protein-2 (RhBMP-2) to augment posterior spine population fusion Steinbrook 2011 243 Industry payments to physicians: lessons from orthopedic surgery Commentary Szpalski 2005 244 Recombinant human bone morphogenetic protein-2: a novel Review osteoinductive alternative to autogenous bone graft? Tanaka 2001 245 Involvement of bone morphogenic protein-2 (BMP-2) in the Irrelevant pathological ossification process of the spinal ligament population U.S. Food and Drug Administration 246 InFuseTM Bone Graft/LT-CAGETM Lumbar Tapered Fusion Device - Product description P000058 U.S Food and Drug Administration 247 Summary of safety and effectiveness data Safety review
178 Appendix IX – Characteristics of excluded papers
Weiner 2003 248 Efficacy of autologous growth factors in lumbar intertransverse No use of rhBMP-2 fusions Willems 2012 249 Clinical decision making in spinal fusion for chronic low back pain. Survey Results of a nationwide survey among spine surgeons World Health Organization 2008 250 Recombinant human bone morphogenetic protein: life-threatening Commentary complications Yu 2003 251 Clinical evaluation of bone morphogenetic protein in spinal fusion Background information published in Chinese
Responses/letters to editors Reference Title Response to Albert 2006 252 Clinical outcomes and fusion success at 2 years of single-level Dimar 2006 99 instrumented posterolateral fusions with recombinant human bone morphogenetic protein-2/compression resistant matrix versus iliac crest bone graft - Point of view Benglis 2010 253 Update of comprehensive review of the safety profile of bone Smoljanovic 2010 morphogenetic protein in spine surgery. Reply. 254 Bennett 2006 255 Recent article by Shields et al titled "adverse effects associated with Shields 2006 256 high-dose recombinant human bone morphogenetic protein-2 use in anterior cervical spine fusion" Benzel 2010 257 Formation of painful seroma and edema after the use of recombinant Garrett 2010 258 human bone morphogenetic protein-2 in posterolateral lumbar spine fusions. Comments Boden 2010 259 Re: Boden SD, Zdeblick TA, Sandhu HS, et al. The use of rhBMP-2 in Boden 2000 9 interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine 2000;25:376-81 Response Brower 2009 260 Brower RS, Vickrov NM. A case of psoas ossification from the use of Dickerman 2009 261 BMP-2 for posterolateral fusion at L4-L5. Spine 2008;33:E653-55. Response
179 Appendix IX – Characteristics of excluded papers
Buchowski 2011 262 In reference to acute airway obstruction in cervical spinal procedures Yaremchuk 2010 145 with bone morphogenetic proteins Cahill 2009 263 Complications associated with use of bone-morphogenetic proteins in Smoljanovic 2009 spinal fusion procedures: reply 264 Carragee 2011 11 Re: A critical review of recombinant human bone morphogenetic Woo 2011 265 protein-2 trials in spinal surgery: emerging safety concerns and lessons learned Response Cardoso 2009 266 Is use of bone-morphogenetic proteins for spine fusion surgery cost- Cahill 2009 7 effective? Cimic 2010 267 Re: Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use Hiremath 2009 139 recombinant human bone morphogenetic protein in posterior cervical fusion? Spine 2009;34:885-9 Corbin 2012 268 Re: Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic Williams 2011 115 protein increase the incidence of perioperative complications in spinal fusion? Spine 2011;36:1658-91 Dickerman 2007 269 rh-BMP-2 can be used safely in the cervical spine: dose and Perri 2007 270 containment are the keys! Dickerman 2008 271 Calcium phosphate silicate for spinal fusion: a good alternative to Wong 2008 272 bone morphogenetic protein-2! Dickerman 2009 261 Brower RS, Vickrov NM. A case of psoas ossification from the use of Brower 2008 273 BMP-2 for posterolateral fusion at L4-L5. Spine 2008;33:E653-55 Dickman 2003274 A prospective, randomized, controlled cervical fusion study using Baskin 200367 recombinant human bone morphogenetic protein-2 with the CORNERSTONE-SR allograft ring and the ATLANTIS anterior cervical plate: point of view Dimar 2011 275 Reply to "A critical review of recombinant human bone morphogenetic Carragee 2011 11 protein-2 trials in spinal surgery: emerging safety concerns and lessons learned" Dmitriev 2011 276 Adverse events and bone morphogenetic protein-2 Response Salisbury 2011 237
180 Appendix IX – Characteristics of excluded papers
Epstein 2008 277 Outcomes of bone morphogenetic protein-2 in mature adults: Hamilton 2008 278 posterolateral non-instrument-assisted lumbar decompression and fusion - Commentary Gill 2007 279 Re: Dimar JR, Glassman SD, Burkus KJ, et al. Clinical outcomes and Dimar 2006 99 fusion success at 2 years of single-level instrumented posterolateral fusions with recombinant human bone morphogenetic protein- 2/compression resistant matrix versus iliac crest bone graft. Spine 2006; 31:2534-40 Heggeness 2011 280 Important considerations on bone morphogenetic protein-2 and Dmitriev 2011 281 neuroinflammation Hu 2011 282 Iliac crest bone graft: are the complications overrated? Not relevant Jacobson 2011 283 Radiographic and CT evaluation of recombinant human bone Erratum to Sethi morphogenetic protein-2-assisted spinal interbody fusion (American 2011 284 Journal of Roentgenology (2011) 197, (W128-W133)) Kang 2011 285 Commentary: Another complication associated with rhBMP-2? Carragee 2011 134 Krishnaney 2010 286 Re: Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use Cimic 2010 267 recombinant human bone morphogenetic protein in posterior cervical fusion? Spine 2009; 34: 885-9. Response Mirza 2011 287 Folly of FDA-approval studies for bone morphogenetic protein Carragee 2011 11 Pagni 2008 288 Outcomes of bone morphogenetic protein-2 in mature adults: Hamilton 2008 278 posterolateral non-instrument-assisted lumbar decompression and fusion - Commentary Pawl 2008 289 Outcomes of bone morphogenetic protein-2 in mature adults: Hamilton 2008 278 posterolateral non-instrument-assisted lumbar decompression and fusion - Commentary Rakovac 2011 290 Recombinant human bone morphogenetic protein 2 labeled use in Epstein 2011 216 spinal surgery and sexual dysfunction Singh 2007 291 Erratum: Use of recombinant human bone morphogenetic protein-2 Erratum to Singh as an adjunct in posterolateral lumbar spine fusion: a prospective CT- 2006 127 scan analysis at one and two years (Journal of Spinal Disorders and Techniques (2006) 19, (416-423))
181 Appendix IX – Characteristics of excluded papers
Slosar 2008 292 Accelerating lumbar fusions by combining rhBMP-2 with allograft Smoljanovic 2008 bone: a prospective analysis of interbody fusion rates and clinical 293 outcomes Smoljanovic 2007 294 Bone morphogenetic protein and fusion Villavicencio 2005 295 Smoljanovic 2008 293 Re: complications attributable to the use of rhBMP-2 inside the Slosar 2007 133 femoral ring allograft during anterior lumbar interbody fusion Smoljanovic 2008 296 Re: Burkus J K, Sandhu H S, Gornet M F. Influence of rhBMP-2 on Burkus 2006 55 the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine 2006;31:775-81 Smoljanovic 2008 297 Re: Burkus J K, Transfeldt E E, Kitchel S H, et al. Clinical and Burkus 2002 53 radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine 2002;27:2396-408 Smoljanovic 2009 298 Bone morphogenetic protein Tumialan 2008 299 Smoljanovic 2009 300 The use of bone morphogenetic protein in lumbar spine surgery Not relevant Smoljanovic 2009 301 End plates resorptions after the applications of rhBMP-2 for interbody Vaidya 2008 302 spinal fusions Smoljanovic 2009 264 Complications associated with use of bone-morphogenetic proteins in Cahill 2009 7 spinal fusion procedures Smoljanovic 2009 303 Avoiding unanticipated adverse effects of recombinant human bone Not relevant morphogenetic protein-2 therapy in craniofacial surgery with experiences from spinal applications Smoljanovic 2009 304 Adverse effects of posterior lumbar interbody fusion using rhBMP-2 Meisel 2008 305 Smoljanovic 2010 306 Re: Kleeman TJ, Ahn UM, Talbot-Kleeman A. Laparoscopic anterior Kleeman 2001 47 lumbar interbody fusion with rhBMP-2: a prospective study of clinical and radiographic outcomes. Spine 2001;26:2751-6 Smoljanovic 2010 254 Update of comprehensive review of the safety profile of bone Benglis 2008 200 morphogenetic protein in spine surgery
182 Appendix IX – Characteristics of excluded papers
Smoljanovic 2010 307 Re: Mroz TE, Wang JC, Hashimoto R, et al.Complications related to Mroz 2010 182 osteobiologics use in spine surgery: a systematic review. Spine (Phila Pa 1976) 2010;35:S86-104 Smoljanovic 2010 308 Re: Toth JM, Boden SD, Burkus JK, et al. Short-term osteoclastic Not relevant activity induced by locally high concentrations of recombinant human bone morphogenetic protein-2 in a cancellous bone environment. Spine 2009;34:539-50 Smoljanovic 2010 309 Re: Boden SD, Zdeblick TA, Sandhu HS, et al. The use of rhBMP-2 in Boden 2000 9 interbody fusion cages. DeFinitive evidence of osteoinduction in humans: a preliminary report. Spine 2000; 25:376-81 Smoljanovic 2010 310 Is a barrier really necessary to prevent radiculitis when using Rihn 2009 125 recombinant human bone morphogenetic protein-2 in proximity of nerve roots? Smoljanovic 2010 311 Re: Mindea SA, Shih P, Song JK. Recombinant human bone Mindea 2009 122 morphogenetic protein-2-induced radiculitis in elective minimally invasive transforaminal lumbar interbody fusions: a series review. Spine 2009;34:1480-5 Smoljanovic 2010 312 Safety of posterior interbody fusions of the lumbar spine using Geibel 2009 313 rhBMP-2 Smoljanovic 2010 314 RE: Clinical applications of bone morphogenetic proteins: current Kanakaris 2008 178 evidence. Kanakaris, NK; Giannoudis, PV; JSOA 17(3):133-46, 2008 Smoljanovic 2010 315 Aggressive end plate decortication as a cause of osteolysis after Klimo 2009 316 rhBMP-2 use in cervical spine interbody fusion Smoljanovic 2010 317 Osteoinductive bone graft substitutes for lumbar fusion Agarwal 2009 169 Smoljanovic 2010 318 Six-year outcomes of anterior lumbar interbody arthrodesis with use Burkus 2009 42 of interbody fusion cages and recombinant human bone morphogenetic protein-2 Smoljanovic 2010 319 The level of evidence in clinical studies of the use of rhBMP-2 in Miyazaki 2009 181 spinal interbody fusions Smoljanovic 2010 320 Recommendation for use of rhBMP-2 in spinal interbody fusions Miyazaki 2009 181
183 Appendix IX – Characteristics of excluded papers
Smoljanovic 2011 321 Commentary: An evolving perception of the risk of rhBMP-2 use for Carragee 2011 134 anterior spinal interbody fusions Smoljanovic 2011 322 Could chronic host inflammatory response be responsible for delayed Muchow 2010 323 onset of retrograde ejaculation after the labelled use of recombinant human bone morphogenetic protein-2? Tumialan 2007 324 Adverse swelling associated with use of rh-BMP-2 in anterior cervical Perri 2007 270 discectomy and fusion Tumialan 2007 325 rh-BMP-2 can be used safely in the cervical spine: dose and Perri 2007 270 containment are the keys! Tumialan 2009 326 Bone morphogenetic protein RESPONSE Smoljanovic 2009 298 Vaidya 2009 327 Transforaminal interbody fusion and the "off label" use of recombinant Rihn 2009 125 human bone morphogenetic protein-2 Vaidya 2009 328 End plates resorptions after the applications of rhBMP-2 for interbody Smoljanovic 2009 spinal fusions response 301 Wang 2011 329 Gene therapy for spinal fusion Olabisi 2011 White 2009 330 Lumbar spinal fusion rates as influenced by bone grafts and bone Not relevant graft alternatives: a critical appraisal of common clinical and radiographic comparative methodologies Wolfla 2008 331 A comprehensive review of the safety profile of bone morphogenetic Benglis 2008 200 protein in spine surgery - Comments Woo 2011 332 Expanded indication for recombinant human bone morphogenetic Ong 2010 8 protein 2 Woo 2011 265 Re: A critical review of recombinant human bone morphogenetic Carragee 2011 11 protein-2 trials in spinal surgery: emerging safety concerns and lessons learned Woo 2012 333 Re: Williams BJ, Smith JS, Fu KM, et al. Does bone morphogenetic Williams 2011 115 protein increase the incidence of perioperative complications in spinal fusion? Spine 2011;36:1658-91
184 Appendix IX – Characteristics of excluded papers
Wrubel 2010 334 Routine use of recombinant human bone morphogenetic protein-2 in Fahim 2010 335 posterior fusions of the pediatric spine: safety profile and efficacy in the early postoperative period. Comments Wu 2011 336 Safety, efficacy, and dosing of recombinant human bone Hamilton 2011 337 morphogenetic protein-2 for posterior cervical and cervicothoracic instrumented fusion with a minimum 2-year follow-up: Commentary Yaremchuk 2011 338 In response to acute airway obstruction in cervical spinal procedures Buchowski 2011 262 with bone morphogenetic proteins
Full text could not be retrieved – full text screening not possible Reference Title Anonymous 2009 339 A retrospective review of posterior cervical fusions with rhBMP-2/ACS Luque 2002 340 Latest clinical results using demineralised bone materials and rhBMP-2: the Mexican experience Sekhon 2009 341 Immediate postoperative complications and radiological features of interbody fusion between infuse bmp andactifuse: A randomized prospective trial
Excluded adverse events papers from the wider literature Reference Title Reason for exclusion Alexander 2009342 A prospective randomized clinical trial of posterolateral lumbosacral N<100 & spinal fusion with BMP-2 and titanium pedicle screw instrumentation uncontrolled versus BMP-2 alone: preliminary 6-month results Anderson 2011343 Postoperative cervical myelopathy and cord compression associated N≤10 with the use of recombinant bone morphogenetic protein-2 in posterior cervical decompression, instrumentation, and arthrodesis: a report of two cases Anderson 2011344 Heterotopic ossification associated with rhBMP-2 (Infuse) in N≤10 posterolateral lumbar spine fusion: a case report
185 Appendix IX – Characteristics of excluded papers
Antonucci 2008345 Use of recombinant human bone morphogenetic protein 2 in pediatric Paediatric spinal fusion population Balseiro 2010346 ertebral osteolysis originating from subchondral cyst end plate defects N≤10 in transforaminal lumbar interbody fusion using rhBMP-2. Report of two cases Boakye 2005347 Anterior cervical discectomy and fusion involving a uncontrolled polyetheretherketone spacer and bone morphogenetic protein Brower 2008273 A case of psoas ossification from the use of BMP-2 for posterolateral N≤10 fusion at L4-L5 Carreon 2008348 Adverse events in patients re-exposed to bone morphogenetic protein uncontrolled for spine surgery Chen 2010349 Symptomatic ectopic bone formation after off-label use of N≤10 recombinant human bone morphogenetic protein-2 in transforaminal lumbar interbody fusion Crandall 2011350 RhBMP-2 in TLIF: dose related complications from a large series uncontrolled Deutsch 2010351 High-dose bone morphogenetic protein-induced ectopic abdomen N≤10 bone growth Fahim 2010335 Routine use of recombinant human bone morphogenetic protein-2 in Paediatric posterior fusions of the pediatric spine: safety profile and efficacy in population the early postoperative period Garrett 2010258 Formation of painful seroma and edema after the use of recombinant uncontrolled human bone morphogenetic protein-2 in posterolateral lumbar spine fusions Geibel 2009313 The use of recombinant human bone morphogenic protein in uncontrolled posterior interbody fusions of the lumbar spine: a clinical series Glassman 2011a155 Complications with recombinant human bone morphogenetic protein- uncontrolled 2 in posterolateral spine fusion associated with a dural tear Glassman 2011b Complications with recombinant human bone morphogenic protein-2 uncontrolled in posterolateral spine fusion: a consecutive series of 1037 cases
186 Appendix IX – Characteristics of excluded papers
Hamilton 2008278 Outcomes of bone morphogenetic protein-2 in mature adults: uncontrolled posterolateral non-instrument-assisted lumbar decompression and
fusion Hamilton 2010352 Use of recombinant human bone morphogenetic protein-2 as an uncontrolled adjunct for instrumented posterior arthrodesis in the occipital cervical region: an analysis of safety, efficacy, and dosing Hamilton 2011337 Safety, efficacy, and dosing of recombinant human bone uncontrolled morphogenetic protein-2 for posterior cervical and cervicothoracic instrumented fusion with a minimum 2-year follow-up Hansen 2006353 Resorptive response of rhBMP2 simulating infection in an anterior N≤10 lumbar interbody fusion with a femoral ring Helgeson 2011354 Adjacent vertebral body osteolysis with bone morphogenetic protein uncontrolled use in transforaminal lumbar interbody fusion Hodges 2012355 Retrospective study of posterior cervical fusions with rhBMP-2 uncontrolled Hoffman 2011a356 Complication rates utilizing rhBMP-2 for lumbar posterolateral fusions uncontrolled Hoffman 2011b357 Recombinant human bone morphogenetic protein-2 (rhBMP-2) in uncontrolled posterolateral spine fusion: what's the correct successful dose/dosage? Hoffman 2011c358 ecombinant human bone morphogenetic protein-2 (rhBMP-2) in uncontrolled posterolateral lumbar spine fusion: complications in the elderly Katayama 2009359 Clinical and radiographic outcomes of posterolateral lumbar spine All patients fusion in humans using recombinant human bone morphogenetic received rhBMP-2 protein-2: an average five-year follow-up study Kepler 2011360 Delayed pleural effusion after anterior thoracic spinal fusion using N≤10 bone morphogenetic protein-2 Klimo 2009316 Use of polyetheretherketone spacer and recombinant human bone uncontrolled morphogenetic protein-2 in the cervical spine: a radiographic analysis Knox 2011361 Osteolysis in transforaminal lumbar interbody fusion with bone uncontrolled morphogenetic protein-2 Kuklo 2004362 Computerized tomography evaluation of a resorbable implant after uncontrolled transforaminal lumbar interbody fusion
187 Appendix IX – Characteristics of excluded papers
Lanman 2004363 Early findings in a pilot study of anterior cervical interbody fusion in uncontrolled which recombinant human bone morphogenetic protein-2 was used with poly(L-lactide-co-D,L-lactide) bioabsorbable implants Lanman 2004364 Lumbar interbody fusion after treatment with recombinant human uncontrolled bone morphogenetic protein-2 added to poly(L-lactide-co-D,L-lactide) bioresorbable implants Lehman 2011365 Vertebral body osteolysis after minimal-access transforaminal N≤10 interbody fusion Lewandrowski 2007366 Vertebral osteolysis after posterior interbody lumbar fusion with N≤10 recombinant human bone morphogenetic protein 2: a report of five cases Lindley 2011367 Complications associated with recombinant human bone Paediatric morphogenetic protein use in pediatric craniocervical arthrodesis population McClellan 2006368 Vertebral bone resorption after transforaminal lumbar interbody fusion uncontrolled with bone morphogenetic protein (rhBMP-2) Mannion 2011369 Promoting fusion in minimally invasive lumbar interbody stabilization uncontrolled with low-dose bone morphogenic protein-2 – but what is the cost? Meisel 2008305 Posterior lumbar interbody fusion using rhBMP-2 uncontrolled Moshel 2008370 Acute renal insufficiency, supraventricular tachycardia, and confusion N≤10 after recombinant human bone morphogenetic protein-2 implantation for lumbosacral spine fusion Muchow 2010323 Histopathologic inflammatory response induced by recombinant bone N≤10 morphogenetic protein-2 causing radiculopathy after transforaminal lumbar interbody fusion Mulconrey 2008371 Bone morphogenetic protein (RhBMP-2) as a substitute for iliac crest uncontrolled bone graft in multilevel adult spinal deformity surgery: minimum two- year evaluation of fusion Neal 2008372 Adjacent vertebral body osteolysis with bone morphogenetic protein uncontrolled use in transforaminal lumbar interbody fusion Oetgen 2010373 Complications associated with the use of bone morphogenetic protein Paediatric in pediatric patients population
188 Appendix IX – Characteristics of excluded papers
Owens 2011374 Perioperative complications with rhBMP-2 in transforaminal lumbar uncontrolled interbody fusion Pargament 2009375 Swelling associated with use of rhBMP-2 in posterolateral lumbar N≤10 fusion: a case study Perri 2007270 Adverse swelling associated with use of rh-BMP-2 in anterior cervical N≤10 discectomy and fusion: a case study Pradhan 2005376 Radiographic staging, characterization, and outcome of anterior N not reported & cervical discectomy and fusion using allograft, recombinant human uncontrolled bone morphogenetic protein-2 and plate. Poster no. 1304 Reames 2011377 Safety, efficacy, and dosing of recombinant human bone uncontrolled morphogenetic protein 2 (rhBMP 2) for posterior cervical and cervico- thoracic instrumented fusion with a minimum two year follow-up Rihn 2009378 The use of RhBMP-2 in single-level transforaminal lumbar interbody uncontrolled fusion: a clinical and radiographic analysis Robin 2010379 Cytokine-mediated inflammatory reaction following posterior cervical N≤10 decompression and fusion associated with recombinant human bone morphogenetic protein-2: a case study Schwender 2005380 Minimally invasive transforaminal lumbar interbody fusion (TLIF): Unclear how many technical feasibility and initial results participants received rhBMP-2 Sethi 2011284 Radiographic and CT evaluation of recombinant human bone uncontrolled morphogenetic protein-2-assisted spinal interbody fusion Shah 2010381 Recombinant human bone morphogenetic protein 2-induced N≤10 heterotopic ossification of the retroperitoneum, psoas muscle, pelvis and abdominal wall following lumbar spinal fusion Shahlaie 2008382 Occipitocervical fusion using recombinant human bone N≤10 morphogenetic protein-2: adverse effects due to tissue swelling and seroma Shen 2010383 Pseudarthrosis in multilevel anterior cervical fusion with rhBMP-2 and uncontrolled allograft: analysis of one hundred twenty-seven cases with minimum two-year follow-up
189 Appendix IX – Characteristics of excluded papers
Shields 2006256 Adverse effects associated with high-dose recombinant human bone uncontrolled morphogenetic protein-2 use in anterior cervical spine fusion Stachniak 2011384 Analysis of prevertebral soft-tissue swelling and dysphagia in uncontrolled multilevel anterior cervical discectomy and fusion with recombinant human bone morphogenetic protein-2 in patients at risk for pseudarthrosis Stambough 2009385 Instrumented one and two level posterolateral fusions with uncontrolled recombinant human bone morphogenetic protein-2 and allograft: a computed tomography study Tumialan 2008299 The safety and efficacy of anterior cervical discectomy and fusion uncontrolled with polyetheretherketone spacer and recombinant human bone morphogenetic protein-2: a review of 200 patients Vaidya 2008302 Complications in the use of rhBMP-2 in PEEK cages for interbody uncontrolled spinal fusions Villavicencio 2005295 Safety of transforaminal lumbar interbody fusion and intervertebral uncontrolled recombinant human bone morphogenetic protein-2 Whang 2008386 Pseudarthrosis following lumbar interbody fusion using bone N≤10 morphogenetic protein-2: intraoperative and histopathologic findings Wong 2008272 Neurologic impairment from ectopic bone in the lumbar canal: a N≤10 potential complication of off-label PLIF/TLIF use of bone morphogenetic protein-2 (BMP-2) Zebala 2011387 RhBMP-2 and modern surgical techniques significantly reduce the uncontrolled pseudarthrosis rate in long fusions to the sacrum for complex adult spinal deformity
190 Appendix X – IPD provided by Medtronic
Appendix X
IPD provided by Medtronic Data file provided Details
Discharge time Time of discharge Adverse events prior to discharge Brace use after discharge
Demographics Age, sex race, education, etc.
Neurological data * Motor, sensory and leg movement tests used to assess neurological status
Oswestry * Scores for all 10 Oswestry lower back pain questionnaire items
Neck disbility index * Scores for all 10 NDI questionnaire items (Cervical spine trials)
Back, leg and hip pain * Scores for two questions on back pain Scores for two questions on leg pain Scores for three questions on hip pain (ICBG patients only)
Arm and neck pain * As for back and leg pain (Cervical spine trials)
Postoperative data * Brace use Tobacco use Weight Return to work or usual activity Adverse events Physician's opinion of results
Pre -operative data Examination date History of spinal surgery Tobacco use Height and weight Alcohol consumption Diabetic, gout, seizures, hepatitis Worker's compensation or litigation case Waddel tests
Qualification data *** Details of disease
191 Appendix X – IPD provided by Medtronic
Past spinal fusion surgery Data to confirm eligibility
SF -36 * Scores for all 36 component sof the SF-36 questionnaire
Surgical data Date of surgery Spinal level Treatment group Operation time Operational approach Blood loss Adverse events at surgery
Patient survey * Pain -releif medication use Satisfaction with surgery Perception of back pain Return to work
Radiologic data * *** Evidence of bone growth, used to determine fusion Disc space Stability Implant problems Radiolucency Evidence of bridging bone
Adverse events * Category Severity Relatedness to device or surgery
Success data * Succesful fusion Oswestry "success" Neurological success Overall success Disc height success
Neurological summary * Motor, sensory and reflex summary data
Pain summary * Oswestry score Back pain score
192 Appendix X – IPD provided by Medtronic
Leg pain score Change from baseline in the above
SF -36 summary * Physical component score Mental component score Scores on eight sub-categories Change from baseline in the above Success status of above
Second surgery * Details of futrther surgeries
Serious adverse events * Serious adverse events reported
Treatment Treatment allocation
* Data provided at multiple time points (post-operative, 6 weeks, 3 months etc) *** Not used in our analyses
193 Appendix XI – IPD provided by Glassman
Appendix XI
IPD provided by Glassman Data type Details
Demographic Randomisaton date Date of birth, sex height weight etc.
Surgical Surgery date Operating time Prior surgery Prior fusion Surgical level Blood loss Hospital stay
Adverse events Various categories
Pain scores * Oswestry score SF-36 PCS SF-36 MCS Back pain Leg pain Graft site pain
* reported pre-operative, six, 12 and 24 months after surgery
194 Appendix XII – Example SAS output (data checking)
Appendix XII
Example SAS output created during the data checking process. Checking of individual participant data for the LT Cage Open trial supplied by Medtronic.
Table 21 Distribution of demographic data (age, previous surgery, smoking, alcohol, BMI), looking for outliers.
Variable N Mean Std Dev Minimum Maximum
DATEPE 287 14268.68 81.5720572 14115.00 14425.00 AGE 281 42.3167260 9.7476758 19.0000000 78.0000000 NUMPREV 119 1.3277311 0.7377157 0 4.0000000 CIG 88 15.2272727 10.3173521 1.0000000 50.0000000 DRINKS 83 4.6419637 5.0709526 0 21.0000000 BMI 284 28.1592217 4.8009222 18.9985795 48.8281250
Table 22 Pain scores (Oswestry. SF-36 components, back and leg pain), checking for outliers and coding errors.
Variable N Mean Std Dev Minimum Maximum
oswtot 1766 34.6194790 20.6759195 0 96.0000000 pfscore 1766 53.1434931 28.4067763 0 100.0000000 rpscore 1765 28.2436261 41.0208815 0 100.0000000 rescore 1763 61.6562677 44.7042052 0 100.0000000 vtscore 1766 50.4869762 23.9162067 0 100.0000000 mhscore 1766 68.4014723 21.2743892 0 100.0000000 sfscore 1766 60.7446206 30.6087512 0 100.0000000 bpscore 1766 47.1843148 27.9841391 0 100.0000000 ghscore 1766 71.9740940 22.1076189 5.0000000 100.0000000 bptot 1760 9.8318182 6.1382693 0 20.0000000 legtot 1760 7.7596591 6.5621330 0 20.0000000 hptot 788 4.9403553 5.9361487 0 20.0000000
195 Appendix XII – Example SAS output (data checking)
Table 23 Treatment coding, checking for consistency across treatment coding in different data files
Table of trt2 by TRTGRP
TRTGRP
trt2 Investigational Control Total
Investigational 145 0 145 51.42 0.00 51.42 100.00 0.00 100.00 0.00
Control 0 137 137 0.00 48.58 48.58 0.00 100.00 0.00 100.00
Total 145 137 282 51.42 48.58 100.00
Frequency Missing = 8
196 Appendix XII – Example SAS output (data checking)
Table 24 Previous surgery, checking for balanced randomisation across treatment groups
Table of TRTGRP by PREVSURG Fisher's Exact Test PREVSURG Cell (1,1) Frequency (F) 54 TRTGRP Yes No Total Left-sided Pr <= F 0.3685 Investigational 54 89 143 Right-sided Pr >= F 0.7194 19.35 31.90 51.25 37.76 62.24 49.54 52.35
Table Probability (P) 0.0880
Control 55 81 136 Two-sided Pr <= P 0.7129 19.71 29.03 48.75 40.44 59.56 50.46 47.65
Total 109 170 279 39.07 60.93 100.00
Frequency Missing = 9
197 Appendix XII – Example SAS output (data checking)
Table 25 Surgery as a result of litigation, checking for balanced randomisation
Table of TRTGRP by LITIGATN
LITIGATN
TRTGRP Yes No Total
Investigational 17 126 143 6.09 45.16 51.25 11.89 88.11 43.59 52.50
Control 22 114 136 7.89 40.86 48.75 16.18 83.82 56.41 47.50
Total 39 240 279 13.98 86.02 100.00
Frequency Missing = 9
Fisher's Exact Test
Cell (1,1) Frequency (F) 17
Left-sided Pr <= F 0.195 0
Right-sided Pr >= F 0.886 0
Table Probability (P) 0.081 0
Two-sided Pr <= P 0.388 1
198 Appendix XII – Example SAS output (data checking) Table 26 Age, checking for balanced randomisation across treatment groups
TRTGRP N Mean Std Dev Std Err Minimum Maximum
Investigational 143 42.8322 9.8624 0.8247 22.0000 78.0000
Control 136 41.7721 9.6999 0.8318 19.0000 70.0000
Diff (1-2) 1.0601 9.7836 1.1718
Equality of Variances
Method Num DF Den DF F Value Pr > F
Folded F 142 135 1.03 0.8464
Table 27 Oswestry score, checking for consitency bertween reported overall score and score as calcualted from raw data
Table of sc_osw by S_OSW
S_OSW
sc_osw Success Failure Total
Success 921 10 931 64.36 0.70 65.06 98.93 1.07 99.57 1.98
Failure 4 496 500 0.28 34.66 34.94 0.80 99.20 0.43 98.02
Total 925 506 1431 64.64 35.36 100.00
Frequency Missing = 809
199 Appendix XIII – Efficacy results chapter
Appendix XIII
Appendices to Efficacy Results Chapter
200 Appendix XIII – Efficacy results chapter
Table 28 Primary and secondary outcomes as specified in Medtronic protocols.
Trial Primary Outcomes Secondary outcomes Other outcomes Trials using INFUSE® INFUSE®/LT-CAGE® pilot RCT Not specified INFUSE®/LT-CAGE® open pivotal RCT Effectiveness: Effectiveness: • Overall success • Fusion • Disc height measurement • Antibody monitoring • Pain/disability status • General health status • Donor site pain status Safety: • Back and leg pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified INFUSE®/LT-CAGE® laparoscopic pivotal single-arm Effectiveness: Effectiveness: • Overall success trial • Fusion • Disc height measurement • Antibody monitoring • Pain/disability status • General health status • Donor site pain status Safety: • Back and leg pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified INFUSE®/bone dowel pilot RCT Effectiveness: Effectiveness: • Bovine collagen antibody response • Fusion • Donor site pain status • Overall success • Disc height management • Patient satisfaction/quality of life status • Pain/disability status • Back/leg pain status • Neurological status • Patient satisfaction Safety: Safety: • Severity and rate of implant-associated • Nature and frequency of adverse events adverse events not associated with the implants • Rate of implant revision, removal, and • Rate of re-operation procedure supplemental fixation procedures • Incidence of permanent adverse events INFUSE®/bone dowel pivotal RCT Effectiveness: Effectiveness: • Overall success • Fusion • Disc height measurement • Antibody monitoring • Pain/disability status • General health status • Donor site pain status Safety: • Back/leg pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified
201 Appendix XIII – Efficacy results chapter
Trial Primary Outcomes Secondary outcomes Other outcomes INFUSE®/INTER FIX™ PLIF RCT Effectiveness: Effectiveness: • Antibody monitoring • Fusion • Donor site pain status • Overall success • Pain/disability status • Disc height measurement • Neurological status • General health status Safety: • Back/leg pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified INFUSE®/MASTERGRAFT® pilot RCT Effectiveness: Effectiveness: • Overall success • Fusion • General health status • Antibody monitoring • Pain/disability status • Back/leg pain status • Donor site pain status Safety: • Patient satisfaction • Doctor’s perception of results Not specified • Patient global perceived effect Safety: Not specified INFUSE®/CORNERSTONE® ACDF pilot RCT Effectiveness: Effectiveness: • Antibody monitoring • Fusion • Time to fusion • Overall success • Pain/disability status • Donor site pain status • Neurological status • General health status Safety: • Neck/arm pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified INFUSE®/CORNERSTONE® ACDF pivotal RCT Effectiveness: Effectiveness: • Overall success • Fusion • Disc height measurement • Antibody monitoring • Pain/disability status • General health status • Donor site pain status Safety: • Neck/arm pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified INFUSE®/INTER FIX™ ALIF pilot RCT Effectiveness: Effectiveness: • Antibody monitoring • Fusion • Donor site pain status • Overall success • Pain/disability status • Disc height measurement • Neurological status • General health status Safety: • Back/leg pain status Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified
202 Appendix XIII – Efficacy results chapter
Trial Primary Outcomes Secondary outcomes Other outcomes MAVERICK™ disc pivotal RCT Effectiveness: Effectiveness: • Overall success • Pain/disability status • General health status • Antibody monitoring • Disc height measurement • Back/leg pain status • Adjacent level motion Safety: • Patient satisfaction • Return to work Not specified • Patient global perceived effect • Doctor’s perception of results • Motion and fusion measurements • Heterotopic ossification Safety: Not specified INFUSE®/TELAMON PEEK instrumented PLIF pilot Effectiveness: Effectiveness: • Overall success single-arm study • Fusion • Disc height measurement • Antibody monitoring • Pain/disability status • General health status • Return to work Safety: • Back/leg pain status • Doctor’s perception of results Not specified • Patient satisfaction • Patient global perceived effect Safety: Not specified Trials using AMPLIFY™ AMPLIFY™ (rhBMP-2/CRM) pivotal RCT Effectiveness: Effectiveness: • Overall success • Fusion • General health status • Antibody monitoring • Pain/disability status • Back/leg pain status • Donor site pain status Safety: • Patient satisfaction Not specified • Patient global perceived effect Safety: Not specified rhBMP-2/CRM 2-level pilot single-arm study Not specified • Patient satisfaction • Patient global perceived effect • Antibody monitoring Trials using rhBMP-2/BCP rhBMP-2/BCP Mexico pilot Not specified rhBMP-2/BCP US pilot Effectiveness: Effectiveness: • Antibody monitoring • Fusion • Neurological status • Overall success • Pain/disability status • Donor site pain status Safety: • General health status Not specified • Back/leg pain status • Patient satisfaction • Patient global perceived effect Safety: Not specified
203 Appendix XIII – Efficacy results chapter
Trial Primary Outcomes Secondary outcomes Other outcomes rhBMP-2/BCP Canada pivotal RCT Effectiveness: Effectiveness: • Antibody monitoring • Fusion • Neurological status • Pain/disability status • Donor site pain Safety: • General health status Not specified • Back/leg pain status • Patient satisfaction • Patient global perceived effect Safety: Not specified
204 Appendix XIII – Efficacy results chapter
Figure 47 Forest plot of back pain 24 months after surgery.
Figure 48 Forest plot of leg pain 24 months after surgery.
205 Appendix XIII – Efficacy results chapter
Figure 49 Meta-analyses of the eight sub-sections of the SF-36 questionnaire.
206 Appendix XIII – Efficacy results chapter
Table 29 Summary of interactions between patient-level factors and ODI, SF-36 PCS, back pain, and leg pain.
Covariate Result ODI SF-36 PCS Back pain Leg pain
Age Baseline No effect SF36 improves with age (0.05% Improves with age (0.04 points Improves with age (0.06 points per year per year, p<0.0001) per year, p<0.0001) p < 0.0001) rhBMP2 Interaction No effect No effect No effect Additional reduction in pain (0.03 points per year, p<0.02)
Sex Baseline No effect Men do worse by 1-2% Men do worse by 0.8 points (p = No effect (p = 0.02) 0.0006) rhBMP2 Interaction No effect No effect No effect Men on rhBMP-2 do worse by 0.7 points (p=0.05)
History Baseline No effect History of surgery reduces benefit History of surgery reduces No effect of surgery by benefit by 1.6 - 1.8% points 1.3 points (p =0.0003) (p <0.0001) rhBMP2 Interaction Significant (p = 0.022) No effect No effect No effect Interaction MD 2.35% points, 95%CI: 0.35 to 4.34 i.e. BMP2 is no better than iliac crest if had previous surgery
Smoking Baseline Reduces benefit of surgery by 3% Reduces benefit of surgery by Reduces benefit of surgery by Reduces benefit of surgery by points (p = 0.0001) 2 - 3% points (p < 0.0001) 0.6 points (p =0.05) 0.7 points (p =0.04) rhBMP2 Interaction No effect No effect No effect No effect
Alcohol Baseline No effect No effect No effect Reduces benefit of surgery by 0.8 points (p=0.016) rhBMP2 Interaction No effect Drinkers on BMP do better by 2 – No effect No effect 3% points (p = 0.01) at 12 months and later
Diabetes Baseline Diabetics do better by 6% points No effect No effect No effect (p < 0.0001) rhBMP2 Interaction Beneficial effect not seen in rhBMP-2 No effect No effect No effect arm (MD = -6% points,p = 0.059)
BMI Baseline No effect No effect Higher BMI improves pain (0.07 No effect points per unit BMI, p=0.0007) rhBMP2 Interaction No effect No effect Beneficial effect not seen in No effect rhBMP-2 arm
207 Appendix XIII – Efficacy results chapter
Table 30 Summary of interactions between patient-level factors and successful fusion.
Covariate Result Successful fusion
Age Baseline No effect rhBMP2 Interaction No effect
Sex Baseline No effect rhBMP2 Interaction No effect
History Baseline No effect of surgery rhBMP2 Interaction (RR 0.60, p = 0.003) i.e. BMP2 is no better than iliac crest if had previous surgery
Smoking Baseline Reduces benefit of surgery (RR 0.55, p < 0.0001) rhBMP2 Interaction Significant interaction (RR 1.6, p = 0.004) i.e. smokers with rhBMP-2 achieve fusion as often as non-smokers
Alcohol Baseline No effect rhBMP2 Interaction No effect
Diabetes Baseline Reduces benefit of surgery (RR 0.03 p < 0.0007) rhBMP2 Interaction No effect
BMI Baseline No effect rhBMP2 Interaction No effect
208 Appendix XIII – Efficacy results chapter
Figure 50 Sensitivity analysis: Oswestry score at 24 months including additional patients from BCP US trial.
Figure 51 Sensitivity analysis: Fusion success at 24 months including additional patients from BCP US trial.
209 Appendix XIII – Efficacy results chapter
Figure 52 Sensitivity analysis: ODI score at 24 months excluding cervical surgery trial (Cornerstone pilot).
Figure 53 Sensitivity analysis: Fusion at 24 months excluding cervical surgery trial (Cornerstone pilot).
210 Appendix XIII – Efficacy results chapter
Figure 54 ODI at 24 months according to pilot or pivotal study.
Figure 55 ODI at 24 months INFUSE® vs AMPLIFY™/BCP.
211 Appendix XIII – Efficacy results chapter
Figure 56 Spinal fusion at 24 months INFUSE® vs AMPLIFY™/BCP.
Figure 57 One-stage meta-analysis results for pain and functionality outcomes.
212 Appendix XIII – Efficacy results chapter
Figure 58 Analysis of efficacy outcomes comparing complete case and multiple imputation analyses.
Figure 59 Sensitivity analysis of spinal fusion comparing complete case and multiple imputation analyses.
213 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Appendix XIV
Table 31 All adverse events categories in the Medtronic studies
Adverse event type
accidental injury/muscle strain lower extremity pain not of back allergic reaction malpositioned implant anatomical/technical difficulty neck and/or arm pain arthritis/bursitis neurological back and/or leg pain non-union cancer non-union - outcome pending cardiovascular non-union pending carpal tunnel syndrome other cervical spinal event other pain cold symptoms respiratory death retrograde ejaculation dural injury shoulder pain dysphagia/dysphonia skin disorder electrolyte imbalance spinal event elevated temperature spinal event at other lumbar level endocrine spinal event at target level foot pain spinal event at thoracic level gastrointestinal subsidence graft site related systemic gynecological trauma headache upper extremity pain not of neck hypokalemia urogenital implant breakage vascular intra-operative implant displacement/loosening vertebral fracture incision related vertigo infection wound complication infection uti wound infection insomina
low hemoglobin
Adverse events summary tables from Medtronic CSRs:
The tables below were copied as seen in the clinical trial reports; no errors were corrected. We marked any numbers where we believe an error might have been made with an asterisk (*) in the respective tables.
214 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Adverse events tables of INFUSE® trials from clinical study reports
Table 32 Adverse events summary table from INFUSE®/LT-CAGE® pilot RCT. TABLE V – ADVERSE EVENTS
Complication Operative Post-operative 5 Weeks 3 Month 6 Month 12 Month 24 Month Investigationa Contro [Number (%)] (1 day - ≤ 5 (> - ≤ 7 Weeks) (>7 Wks - ≤5 Mos) (>5 - ≤9 months) (>9 - ≤19 months) (>19 - ≤30 l l weeks) [Number (%)] [Number (%)] [Number (%)] [Number (%)] months) BMP/TIF TIF [Number (%)] [Number (%)] (% of 11) (% of 3) BMP/TI Contro BMP/TI Contro BMP/TI Contro BMP/TI Contro BMP/TI Contro BMP/TI Contro BMP/TI Contro F l F l F l F l F l F l F l N=11 N = 3 N=11 N = 3 N=11 N = 3 N=11 N = 3 N=11 N = 3 N=11 N = 3 N=11 N = 3 Ileus 1 (9.1) 1 1 (9.1) 1 (33.3) (33.3) 1 (9.1) 1 (9.1) 1 1 (9.1) 1 (33.3) (33.3) Wound 1 (9.1) 1 (9.1) Dehiscence Back Pain 1 (9.1) 1 (9.1) 2 (18.2) Trauma 2 (18.2) 1 (9.1) 1 (9.1) 4 (36.4) Back Strain 1 (9.1) 1 (9.1) Endcap Event 2 (18.2) 1 1 1 (9.1) 3 (27.3) 2 (33.3) (33.3) (66.7) Urinary 1 1 Retention (33.3) (33.3) Fracture 1 (9.1) 1 (9.1) 2 (18.2) Pseudoarthrosi 1 (9.1) 1 1 (9.1) 1 s (33.3) (33.3) Shortness of 1 (9.1) 1 (9.1) Breath Numbness/SLE 1 (9.1) 1 (9.1) Drug Use 1 (9.1) 1 (9.1) Facet Joint 1 (9.1) 1 (9.1) Pain Rectal 1 (9.1) 1 (9.1) Bleeding/ Hemorrhoids
215 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Removals 0 (0.0) 0 (0.0) Revisions 0 (0.0) 0 (0.0) Reoperative 0 (0.0) 0 (0.0) Supplemental 1 (33.3) 0 (0.0) 1 (33.3) Fixations
8 of 11 evaluable investigational patients (72.7%) had a total of 21 adverse events. As to date, none of these have required additional surgery. 3 of 3 evaluable control patients (100%) had a total of 7 adverse events. To date, one patient has required supplemental fixation.
216 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 33 Adverse events summary table from INFUSE®/LT-CAGE® open pivotal RCT.
TABLE 7 – ADVERSE EVENTS1 – INFUSETM BONE GRAFT/LT-CAGETM DEVICE – OPEN APPROACH Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months Total Adverse Events # of Patients Reporting (1 day - <4 Weeks) (≥4 Wks - <9 Weeks) (≥9 Wks - <5 Months) (≥5 Mos - <9 (≥9 Mos - <19 Months) (≥19 Mos - <30 Months) Months) Complication InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETMGraft/ Control Graft/LT- Graft/LT- Graft/LT- Graft/LT- Graft/LT- Graft/LT- Graft/LT- Graft/LT- LT-CAGETM #(% of CAGETM CAGETM CAGETM CAGETM CAGETM CAGETM CAGETM CAGETM Device 136) Device Device Device Device Device Device Device Device # (% of 143) Anatomical/ 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 (0.0) 2 (1.5) Technical Difficulty Back and/or Leg 0 0 5 4 4 5 8 5 6 4 14 7 4 8 41 33 38 (26.6) 30 (22.1) Pain Cancer 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 1 (0.7) 1 (0.7) Cardio/Vascular 0 0 2 5 2 2 1 3 2 1 1 2 0 1 8 14 8 (5.6) 12 (8.8) Death 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 (0.0) 1 (0.7) Dural Injury 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 (0.0) 1 (0.7) Gastrointestinal 0 0 21 21 2 0 3 1 6 1 4 3 1 5 37 31 30 (21.0) 26 (19.1) Graft Site Related 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 8 0 (0.0) 8 (5.9) Implant 0 0 0 0 1 0 1 0 0 0 0 0 0 0 2 0 2 (1.4) 0 (0.0) Displacement/ Loosening Infection 0 0 11 9 4 4 2 1 2 1 2 0 0 2 21 17 18 (12.6) 16 (11.8) Malpositioned 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 (0.7) 0 (0.0) Implant Neurological 0 0 5 5 4 3 3 2 2 2 6 3 2 7 22 22 19 (13.3) 21 (15.4) Non-Union 0 0 0 0 0 0 0 0 0 2 1 0 0 1 1 3 1 (0.7) 3 (2.2) (OUTCOME PENDING) Other 4 6 8 10 3 2 1 4 3 4 7 8 5 6 31 40 24 (16.8) 35 (25.7) Other Pain 0 0 1 1 2 0 3 2 3 1 6 6 4 3 19 13 15 (10.5) 12 (8.8) Respiratory 0 0 2 2 1 0 0 0 0 0 0 1 0 1 3 4 3 (2.1) 4 (2.9) Retrograde 0 0 3 1 1 0 1 0 0 0 0 0 0 0 5 1 5 (3.5) 1 (0.7) Ejaculation (6.4)2 (1.5)3 Spinal Event 0 0 0 2 0 0 6 2 5 3 7 8 3 2 21 17 19 (13.3) 16 (11.8) Subsidence 0 0 3 2 1 0 1 0 1 0 0 0 0 0 6 2 6 (4.2) 2 (1.5) Trauma 0 0 1 4 0 3 4 6 12 5 18 9 4 7 39 34 31 (21.7) 29 (21.3) Urogenital 0 0 13 4 1 0 1 2 4 1 2 1 3 2 24 10 20 (14.0) 9 (6.6) Vascular Intra-Op 6 5 0 0 0 0 0 0 0 0 0 0 0 0 6 5 6 (4.2) 5 (3.7) Vertebral Fracture 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 1 (0.7) 0 (0.0) Any Adverse Event 113 (79.0) 109 (80.1) 1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table. 2 Percent for 78 males 3 Percent for 68 males
217 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 34 Adverse events summary table from INFUSE®/LT-CAGE® laparoscopic pivotal single-arm study.
TABLE 7 – ADVERSE EVENTS1 – INFUSETM BONE GRAFT/LT-CAGETM DEVICE – LAPAROSCOPIC APPROACH
Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months 48 Months Total # of Patients (1 day – (≥4 Weeks – (≥9 Weeks – (≥5 Months – (≥9 Months – (≥19 Months – (≥30 Months – Adverse Events Reporting <4 Weeks) <9 Weeks) <5 Months) <9 Months) <19 Months) <60 Months) <60 Months) Complication InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control InFUSETM/ Control # LT Device LT Device LT Device LT Device LT Device LT Device LT Device LT Device LT Device LT Device (% of 136) # (% of 134) Anatomical/Technical 12 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 2 12 (8.8) 2 (1.5) Difficulty Back and/or Leg Pain 0 0 6 4 6 5 3 5 7 4 6 7 2 9 1 5 31 34 26 (19.1) 34 (25.0)
Cancer 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 (0.0) 1 (0.7)
Cardio/Vascular 2 0 2 5 4 2 0 3 0 1 2 2 0 1 0 0 10 14 7 (5.1) 12 (8.8)
Death 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 2 0 (0.0) 2 (1.5)
Dural Injury 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 (0.0) 1 (0.7)
Gastrointestinal 1 0 19 21 0 0 2 1 2 1 5 3 3 5 0 0 32 31 24 (17.6) 26 (19.1)
Graft Site Related 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 (0.0) 8 (5.9)
Implant Displacement/ 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 2 0 2 (1.5) 0 (0.0) Loosening Infection 0 0 8 9 4 4 2 1 3 1 1 0 0 2 1 0 19 17 17 (12.5) 16 (11.8)
Malpositioned Implant 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 4 (2.9) 0 (0.0)
Neurological 0 0 2 5 3 3 2 2 3 2 4 4 3 7 1 2 18 23 18 (13.2) 24 (17.7)
Non-Union 0 0 0 0 0 0 0 0 0 1 3 0 1 1 0 1 4 3 4 (2.9) 3 (2.2) (OUTCOME PENDING) Other 2 6 9 10 4 2 2 4 5 4 6 8 7 6 1 2 36 40 30 (22.4) 36 (26.5)
Other Pain 0 0 1 1 0 1 1 2 3 1 1 7 4 3 0 1 10 15 10 (7.4) 14 (10.3)
Respiratory 0 0 1 2 0 0 0 0 0 0 0 1 0 1 0 0 1 4 1 (0.7) 4 (2.9)
Retrograde Ejaculation 0 0 2 1 3 0 0 0 0 0 2 0 0 0 0 0 7 1 6 (4.4) 1 (0.7) (10.3)2 (1.5)3 Spinal Event 0 0 1 2 0 0 0 2 3 4 1 8 2 2 0 3 7 18 6 (4.4) 20 (14.7)
Subsidence 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 1 2 1 (0.7) 2 (1.5)
Trauma 0 0 1 5 4 3 7 7 1 5 9 12 8 8 0 1 30 40 28 (20.6) 35 (25.7)
Urogenital 1 0 9 4 1 0 1 3 2 3 0 1 2 2 0 0 16 13 16 (11.8) 12 (8.8)
Vascular Intra-Op 9 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 5 8 (5.9) 5 (3.7)
Any Adverse Event 102 (75.0) 114 (83.8)
218 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table. 2 Percent for 58 males 3 Percent for 68 Males
219 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 35 Adverse events summary table from INFUSE®/bone dowel pilot RCT.
ADVERSE EVENTS1 – rhBMP-2/ACS/ALLOGRAFT BONE DOWEL PILOT – as of 01/28/04 Postoperative 12 Months 24 Months 48 Months 72 Months 6 Weeks 3 Months 6 Months Total Discrete # of Patients Surgery (1 day - <1 (≥9 - <19 (≥19 - <30 (≥30 - <60 (≥60 - <84 (≥1 – 2 Months) (≥2 - <5 Months) (≥5 - <9 Months) Adverse Events Reporting Month) Months) Months) Months) Months) InFUSE® Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control BD #(% of Complication N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 N=24 N=22 #(% of 22) 24) Anatomical/Technical 0(0.0) 0(0.0) Difficulty Back and/or Leg 2 2 4 1 1 2 9 3 9(37.5) 3(13.6) Pain Cancer 1 1 2 1(4.2) 0(0.0) Cardio/Vascular 1 1 1(4.2) 0(0.0) Death 1 1 0(0.0) 1(4.5) Dural Injury 0(0.0) 0(0.0) Gastrointestinal 1 3 1 2 3 2(8.3) 3(13.6) Graft Site Related 1 1 0(0.0) 1(4.5) Implant Displacement/ 1 1 1(4.2) 0(0.0) Loosening/Breakage Infection 1 1 0(0.0) 1(4.5) Malpositioned 0(0.0) 0(0.0) Implant Neurological 1 2 1 1 4 1 4(16.7) 1(4.5) Non-Union (OUTCOME 1 2 1 2 2 1 6 1(4.2) 6(27.3) PENDING) Other 1 1 1 1 2 4 3 7 3(12.5) 6(27.3) Other Pain 1 1 1 1 2 1 4 2 9 4 7(29.2) 3(13.6) Respiratory 0(0.0) 0(0.0) Retrograde 0(0.0) 0(0.0) Ejaculation Spinal Event 1 1 1 5 1 2 1 7 5 5(20.8) 4(18.2) Subsidence 0(0.0) 0(0.0) Trauma 1 1 1 3 1 1 2 2 5 7 10 5(20.8) 8(36.4) Urogenital 1 1 1 1 3 1 3(12.5) 1(4.5) Vascular Intra-Op 2 3 2 3 2(8.3) 3 13.6) Vertebral Fracture 0(0.0) 0(0.0)
1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table.
220 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 36 Adverse events summary table from INFUSE®/bone dowel pivotal RCT. SECTION II – TABLE 3 – IDE #G970124
ADVERSE EVENTS1 – rhBMP-2/ACS/ALLOGRAFT BONE DOWEL PIVOTAL (As of 1/28/04) Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months Total Discrete # of Patients Reporting (1 day - <1 (≥1 - <2 (≥2 – 5 Months) (≥5 - <9 (≥9 - <19 (≥19 - <3 Adverse Events Month) Months) Months) Months) Months) Complication INV Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control InfuseTM BD Control N=55 N=30 #(% of 55) #(% of 30) Anataomical/Technical 1 2 1 2 1(1.8) 2(6.7) Difficulty Back and/or Leg Pain 2 2 1 2 3 2 3 4 4 2 1 11 15 9(16.4) 12(40.0) Cancer 1 1 0 1(1.8) 0(0.0) Cardio/Vascular 1 1 2 1 3 2 3(5.5) 2(6.7) Gastrointestinal 8 6 1 1 8 8 8(14.5) 8(26.7) Graft Site Related 1 0 1 0(0.0) 1(3.3) Implant Displacement/ 1 1 0 1(1.8) 0(0.0) Loosening/Breakage Infection 1 3 2 1 1 1 5 4 5(9.1) 3(10.0) Neurological 3 1 1 2 6 3 3 1 16 4 15(27.3) 4(13.3) Non-Union 1 2 1 1 3 1(1.8) 3(10.0) (OUTCOME PENDING) Other 1 2 7 1 2 1 2 5 7 14 7(12.7) 11(36.7) Other Pain 1 2 3 1 4 4 2 3 1 16 5 12(21.8) 4(13.3) Respiratory 2 3 1 2 4 2(3.6) 4(13.3) Spinal Event 1 1 2 3 1 3(5.5) 1(3.3) Subsidence 1 1 0 1(1.8) 0(0.0) Trauma 2 2 2 5 4 2 5 2 2 12 14 10(18.2) 9(30.0) Urogenital 4 2 1 2 7 2 7(12.7) 2(6.7) Vascular Intra-Op 1 1 0 1(1.8) 0(0.0)
1 Since fusion is a primary effectiveness endpoint, non-unions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These non-unions are captured in the secondary surgery table and the fusion table.
221 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 37 Adverse events summary table from INFUSE®/INTER FIX™ PLIF RCT.
TABLE 5 ADVERSE EVENTS1 – INFUSETM Bone Graft/INTER FIXTM Device PLIF Study
Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months Total Adverse # of Patients Reporting2 (1 day - <1 (≥1- <2 (≥2 - <5 (≥5 - <9 (≥9 - <19 (≥19 - <30 Events Month) Months) Months) Months) Months) Months) Complication INV Control INV Control INV Control INV Control INV Control INV Control INV Control INV Control InFUSETM/ Control INTER (% of 33) FIXTM (% of 34) Back and/or 0 0 0 2 4 1 0 1 4 3 4 2 1 2 13 11 12(35.3) 9(27.3) Leg Pain Cardio/Vascular 1 5 6 5 1 0 0 0 0 0 0 0 1 1 9 11 6(17.6) 10(30.3) Death 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 1 1(2.9) 1(3.0) Dural Injury 3 2 0 0 0 0 0 0 0 0 0 0 0 0 3 2 3(8.8) 2(6.1) Gastrointestinal 0 1 8 7 1 0 0 0 1 0 1 3 0 0 11 11 9(26.5) 11(33.3) Graft Site 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 2 0(0.0) 2(6.1) Related Infection 0 0 3 4 0 1 2 0 1 1 2 0 0 0 8 6 7(20.6) 5(15.2) Neurological 0 0 4 3 1 0 2 2 2 4 5 5 2 4 16 18 14(41.2) 14(42.4) Pending Non- 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 2 (0.0) 2(6.1) Union Other 0 0 15 13 2 0 1 2 1 1 1 2 1 5 21 23 17(50.0) 15(45.5) Other Pain 0 0 3 4 2 0 1 2 2 1 3 3 3 1 14 11 9(26.5) 9(27.3) Respiratory 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 2 0(0.0) 2(6.1) Spinal Event 0 0 0 0 0 1 2 0 0 1 2 2 1 1 5 5 5(14.7) 5(15.2) Trauma 0 0 0 1 0 0 3 1 1 1 4 3 0 3 8 9 7(20.6) 5(15.2) Urogenital 0 0 0 4 0 0 1 0 0 0 0 1 0 0 1 5 1(2.9) 4(12.1)
1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table. 2 Once investigational patient did not receive the device intraoperatively, and three control patients did not receive the device intraoperatively. These complications are explained in the Adverse Event Case Histories and are categorized as “Anatomical or Technical Difficulties”. These complications are not included in this table.
222 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 38 Adverse events summary table from INFUSE®/MASTERGRAFT® pilot RCT.
INFUSE Bone Graft/MasterGraftTM Granules/CD HORIZON Spinal System Pilot Study G020056 – IDE Final Progress Report 2006-2007 – TOTAL ADVERSE EVENTS Postop 6 Weeks 3 Months 6 Months 12 Months 24 Months 36 Months Total Total # of Patients EVENT Operative (1 day - <1 (≥1-<2 (≥2-<5 (≥5-<9 (≥9-<19 (≥19-<30 (≥30-<42 Adverse Reporting AEs Month) Months) Months) Months) Months) Months) Months) Events #(% of N) Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl Inv Ctrl N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 N=25 N=21 Accidental 10(40.0% Injury/Muscle 0 0 1 0 2 1 2 1 0 1 2 0 5 1 0 0 12 4 4(19.0%) ) Strain Allergic Reaction 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1(4.0%) 0(0.0%) Back &/or Leg Pain 0 0 1 1 1 0 2 0 2 2 2 1 1 1 0 0 9 5 8(32.0%) 5(23.8%) Cancer 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1(4.0%) 0(0.0%) Cardiovascular 0 0 0 1 1 0 0 1 0 1 1 1 0 1 0 0 2 5 2(8.0%) 3(14.3%) Cervical Spinal Event 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 1 1(4.0%) 1(4.8%) Dural Injury 1 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 2 2 2(8.0%) 2(9.5%) Electrolyte Imbalance 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0(0.0%) 1(4.8%)
Endocrine 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0(0.0%) 1(4.8%) Gastrointestinal 0 0 1 4 0 1 1 1 0 0 1 4 1 0 0 1 4 11 3(12.0%) 8(38.1%) Graft Site Related 0 0 0 1 0 0 0 0 0 1 0 1 0 1 0 0 0 4 0(0.0%) 4(19.0%) Implant Displacement/ 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1(4.0%) 0(0.0%) Loosening/Breaking Incision Related 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1(4.0%) 1(4.8%) Infection 0 0 1 1 0 1 0 0 0 2 2 1 1 0 1 1 5 6 4(16.0%) 5(23.8%) Lower Extremity Pain Not of 0 0 1 1 0 0 0 0 1 1 4 0 0 0 1 0 7 2 6(24.0%) 2(9.5%) Back Etiology Malpositioned Implant 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1(4.0%) 0(0.0%) Neck &/or Arm Pain 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0(0.0%) 1(4.8%) Neurological 0 0 2 2 0 0 3 0 0 1 0 1 0 0 0 0 5 4 4(16.0%) 3(14.3%) Non-Union (Pending) 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 2 0(0.0%) 2(9.5%) Other 0 0 1 3 0 0 0 0 1 0 0 0 2 2 0 1 4 6 3(12.0%) 6(28.6%) Other Pain 0 0 1 0 0 0 2 0 0 0 2 0 0 0 1 0 6 0 4(16.0%) 0(0.0%) Respiratory 0 0 0 1 0 0 0 0 0 1 1 0 1 0 0 0 2 2 2(8.0%) 2(9.5%) Skin Disorder 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 2 0(0.0%) 2(9.5%) Spinal Event at Other 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 1 2 2 2(8.0%) 2(9.5%) Lumbar Level(s) Spinal Event at 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 2 0 2(8.0%) 0(0.0%) Target Level Spinal Event at 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0(0.0%) 1(4.8%) Thoracic Level Trauma 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1(4.0%) 0(0.0%) Upper Extremity Pain 0 0 0 0 0 0 1 0 1 0 0 0 0 1 0 1 2 2 2(8.0%) 1(4.8%) Not of Neck Etiology Urogenital 0 0 1 0 0 0 0 0 1 1 0 1 0 0 0 0 2 2 2(8.0%) 2(9.5%) TOTAL 22(88.0% 1 1 13 17 4 3 12 6 7 13 18 14 14 7 4 6 73 67 18(85.7%) ) N = Study patient sample size – Final as of February 28, 2007 CRD: Due to the poor print quality of the original Medtronic CSR, the top row of this table was difficult to read and inaccuracies might have occurred.
223 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 39 Adverse events summary table from INFUSE®/CORNERSTONE® ACDF pilot RCT.
Table V – ADVERSE EVENTS1 – INFUSETM Bone Graft/CORNERSTONE-SRTM/ATLANTISTM ACP – Pilot Study (as of 3/4/03)
Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months 48 Months Total Adverse # of Patients Reporting (1 day - <4 (≥4- <9 Weeks) (≥9 Wks -<5 (≥5 Mos - <9 (≥9 Mos - <19 (≥19 Mos - <30 (≥ 30 Mos - <60 Events Weeks) Months) Months) Months) Months) Months) Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Invest. Control Complication (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) (N=18) (N=15) # (% of 18) # (% of 15) Back and/or Leg 1 2 1 1 1 4 2 4(22.2) 2(13.3) Pain Cardiovascular 1 2 1 3 1 3(16.7) 1(6.7) Death 1 1 0 1(5.5) 0(0.0) Dysphagia/ 1 2 1 2 2 2(11.1) 2(13.3) Dysphonia Gastrointestional 1 1 1 2 5 0 5(27.8) 0(0.0) Headache 1 1 0 2 0(0.0) 2(13.3) Infection 2 2 4 0 4(22.2) 0(0.0) Neck and/or Arm 1 1 1 1 1(5.5) 1(6.7) Pain Neurological 1 3 3 2 1 8 2 8(44.4) 2(13.3) Other 2 1 1 1 2 1 7 1 7(38.9) 1(6.7) Other Pain 1 1 1 2 1 2(11.1) 1(6.7) Respiratory 1 1 0 1(5.5) 0(0.0) Spinal Event 1 3 3 1 3(16.7) 1(6.7) Trauma 1 1 1 1 3 1 3(16.7) 1(6.7) Urogenital 1 1 1 2 1 2(11.1) 1(6.7)
1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table.
224 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 40 Adverse events summary table from INFUSE®/CORNERSTONE® ACDF pivotal RCT.
Table IV. Adverse Events
ADVERSE Operative Postoperative 6 Week 3 Month 6 Month 12 Month 24 Month DISCRETE ADVERSE EVENTS [Number (%)] (1 Day <4 (≥4 <9 Weeks) (≥9 Weeks <5 (≥5 <9 Months) (≥9 <19 (≥19 <30 EVENTS Weeks) [Number (%)] Months) [Number (%)] Months) Months) [Number (%)] [Number (%)] [Number (%)] [Number (%)] Invest Control Invest Control Invest Control Invest Control Invest Control Invest Control Invest Control Invest. Control N=2 N=1 N=2 N=1 N=2 N=1 N=2 N=1 N=2 N=1 N=2 N=1 N=2 N=1 (% of 2) (% of 1) Back and/or Leg 1 1 1(50%) 1(100%) Pain Other 1 1(50%) 0 Trauma 1 0 1(100%)
225 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 41 Adverse events summary table from INFUSE®/INTER FIX™ ALIF pilot RCT.
Table V – ADVERSE EVENTS1 – INFUSETM BONE GRAFT/INTER FIXTM DEVICE – Pilot (as of 1/27/03) Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months 48 Months Total Adverse Events # of Patients (1 day - <4 Weeks) (≥4 Wks - <9 Weeks) (≥9 Wks - <5 Months) (≥5 Mos - <9 Months) (≥9 Mos - <19 (≥19 Mos - <30 (≥30 Mos - <60 Reporting Months) Months) Months) Complication InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control InFUSETM Control Graft/ Graft/ Graft/ Graft/ Graft/ Graft/ Graft/ Graft/ Graft/ Graft/ #(% of INTER INTER INTER INTER INTER INTER INTER INTER INTER INTER 20) FIXTM FIXTM FIXTM FIXTM FIXTM FIXTM FIXTM FIXTM FIXTM FIXTM Device Device Device Device Device Device Device Device Device Device #(% of 25) Anatomical/ 1 1 1 1 1(4.0) 1(5.0) Technical Difficulty Back and/or 1 1 1 1 2 2 1 1 5 5 5(20.0) 5(25.0) Leg Pain Cardiovascular 1 1 1 2 1 2(8.0) 1(5.0) Death 1 0 1 0(0.0) 1(5.0) Gastrointestinal 2 1 1 1 1 4 2 5(20.0) 1(5.0) Graft Site 1 0 1 0(0.0) 1(5.0) Related Implant 1 2 1 2 1(4.0) 2(10.0) Displacement/ Loosening Neck and/or 1 1 0 1(4.0) 0(0.0) Arm Pain Neurological 1 2 1 2 2 2(8.0) 2(10.0) Other 1 1 2 3 1 2 6 2(8.0) 6(30.0) Respiratory 1 1 0 1(4.0) 0(0.0) Spinal Event 1 1 1 1 2 2 2(8.0) 2(10.0) Trauma 1 1 3 1 1 1 2 8 2 8(32.0) 2(10.0) Vertebral 1 1 0 1(4.0) 0(0.0) Fracture
1Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table.
226 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Table 42 Adverse events summary table from MAVERICK™ Disc Pivotal RCT MaverickTM Total Disc Replacement – GO10354 Final Report – Table 4. Time Course Summary of All Adverse Events, Pivotal Trial
Investigational (N=405)
Operative 1 day – 4 wks 6 weeks 3 months 6 months 12 months 24 months Adverse Event Type Patient Patient Patient Patient Patient Patient Patient Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%)
Patients Who Had Any Adverse Events 65 59(14.6) 230 167(41.2) 106 90(22.2) 150 125(30.9) 140 110(27.2) 242 155(38.3) 206 125(30.9) ALLERGIC REACTION 0 0(0.0) 2 2(0.8) 3 3(0.7) 0 0(0.0) 2 2(0.5) 1 1(0.2) 4 3(0.7) ANATOMICAL/TECHNICAL DIFFICULTY 8 9(2.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) BACK AND LEG PAIN 0 0(0.0) 7 7(1.7) 3 3(0.7) 8 8(2.0) 5 5(1.2) 5 5(1.2) 7 7(1.7) BACK PAIN 0 0(0.0) 6 6(1.5) 11 11(2.7) 18 18(4.4) 12 12(3.0) 27 27(6.7) 13 13(3.2) CANCER 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) 0 0(0.0) 2 2(0.5) CARDIOVASCULAR 1 1(0.2) 2 2(0.5) 0 0(0.0) 1 1(0.2) 1 1(0.2) 5 5(1.2) 4 4(1.0) CERVICAL SPINE EVENT 0 0(0.0) 0 0(0.0) 1 1(0.2) 2 2(0.5) 3 3(0.7) 7 7(1.7) 5 5(1.2) DEATH 0 0(0.0) 1 1(0.2) 1 1(0.2) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) DEPRESSION 0 0(0.0) 3 3(0.7) 3 3(0.7) 5 5(1.2) 6 6(1.5) 4 4(1.0) 4 4(1.0) ELEVATED TEMPERATURE WITHOUT INFECTION 0 0(0.0) 6 6(1.5) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) GASTROINTESTINAL – ILEUS 7 7(1.7) 13 13(3.2) 0 0(0.0) 1 1(0.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) GASTROINTESTINAL – OTHER 4 4(1.0) 21 20(4.9) 5 5(1.2) 8 7(1.7) 6 6(1.5) 15 13(3.2) 19 18(4.0) IMPLANT DISPLACEMENT/LOOSENING/MALPOSITION 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) INCISION – RELATED 1 1(0.2) 18 18(4.4) 13 13(3.2) 7 7(1.7) 3 3(0.7) 4 4(1.0) 4 4(1.0) INFECTION 1 1(0.2) 6 6(1.5) 4 4(1.0) 3 3(0.7) 5 5(1.2) 5 4(1.0) 3 2(0.5) LEG PAIN 0 0(0.0) 20 20(4.9) 4 4(1.0) 7 7(1.7) 1 1(0.2) 5 5(1.2) 4 4(1.0) LOWER EXTREMITY PAIN 0 0(0.0) 11 11(2.7) 2 2(0.5) 5 5(1.2) 4 4(1.0) 10 10(2.5) 8 9(2.0) NECK AND/OR ARM PAIN 0 0(0.0) 1 1(0.2) 1 1(0.2) 1 1(0.2) 3 3(0.7) 3 3(0.7) 3 3(0.7) NEUROLOGICAL 10 9(2.2) 43 42(10.4) 17 17(4.2) 29 29(7.2) 25 22(5.4) 39 35(8.6) 22 20(4.9) NON-UNION 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) OTHER 4 4(1.0) 19 19(4.7) 7 7(1.7) 11 11(8.7) 9 8(2.0) 21 19(4.7) 22 19(4.7) OTHER PAIN 1 1(0.2) 6 6(1.5) 7 7(1.7) 6 6(1.5) 6 6(1.5) 21 20(4.9) 19 19(4.7) PAIN MANAGEMENT CONTROL 0 0(0.0) 3 3(0.7) 0 0(0.0) 1 1(0.2) 0 0(0.0) 1 1(0.2) 2 2(0.5) PERITONEAL TEAR 3 3(0.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) RASH 1 1(0.2) 3 3(0.7) 0 0(0.0) 0 0(0.0) 1 1(0.2) 1 1(0.2) 1 1(0.2) RESPIRATORY 1 1(0.2) 3 3(0.7) 0 0(0.0) 0 0(0.0) 2 2(0.5) 2 2(0.5) 1 1(0.2) RETROGRADE EJACULATION 0 0(0.0) 3 3(0.7) 0 0(0.0) 1 1(0.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) SPINAL EVENT 3 3(0.7) 2 2(0.5) 3 3(0.7) 0 0(0.0) 2 2(0.5) 3 3(0.7) 1 1(0.2) SPINAL EVENT AT OTHER LUMBAR LEVEL (S) 0 0(0.0) 3 3(0.7) 1 1(0.2) 2 2(0.5) 2 2(0.5) 2 2(0.5) 8 7(1.7) SPINAL EVENT AT TARGET LEVEL 1 1(0.2) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) 3 3(0.7) 0 0(0.0) SPINAL EVENT AT THORACIC LEVEL 1 1(0.2) 1 1(0.2) 0 0(0.0) 1 1(0.2) 0 0(0.0) 3 3(0.7) 3 3(0.7) SUBSIDENCE 0 0(0.0) 2 2(0.5) 2 2(0.5) 1 1(0.2) 3 3(0.7) 4 4(1.0) 2 2(0.5) TRAUMA 0 0(0.0) 10 9(2.2) 12 12(3.0) 20 20(4.9) 30 27(6.7) 37 33(8.1) 26 29(7.2) UPPER EXTREMITY PAIN 0 0(0.0) 1 1(0.2) 2 2(0.5) 4 4(1.0) 1 1(0.2) 3 3(0.7) 3 3(0.7) UROGENITAL 2 2(0.5) 11 11(2.7) 3 3(0.7) 6 6(1.5) 6 6(1.5) 8 8(2.0) 5 4(1.0) VASCULAR EVENT POST – OP PERIPHERAL 0 0(0.0) 2 2(0.5) 0 0(0.0) 0 0(0.0) 2 2(0.5) 0 0(0.0) 0 0(0.0) VASCULAR INJURY INTRA-OP 14 14(3.5) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) VERTEBRAL FRACTURE 1 1(0.2) 1 1(0.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2)
227 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Investigational (N=405) Total 36 months 48 months 60 months 72 months 84 months Total (All) (<=24 months) Adverse Event Type Patient Patient Patient Patient Patient Patient Patient Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Patients Who Had Any Adverse Events 1139 345(85.2) 204 118(29.1) 131 88(21.7) 145 92(22.7) 73 43(10.6) 16 13(3.2) 1709 370(91.4) ALLERGIC REACTION 12 10(2.5) 1 1(0.2) 0 0(0.0) 0 0(0.0) 1 1(0.2) 1 1(0.2) 15 12(3.0) ANATOMICAL/TECHNICAL DIFFICULTY 9 9(2.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) 10 10(2.5) BACK AND LEG PAIN 35 34(.8.4) 8 8(2.0) 5 5(1.2) 1 1(0.2) 1 1(0.2) 0 0(0.0) 50 48(11.9) BACK PAIN 87 78(19.3) 17 17(4.2) 5 5(1.2) 19 19(4.7) 6 6(1.5) 1 1(0.2) 135 112(27.7) CANCER 3 3(0.7) 2 2(0.5) 0 0(0.0) 2 2(0.5) 0 0(0.0) 0 0(0.0) 7 7(1.7) CARDIOVASCULAR 14 14(3.5) 8 8(2.0) 5 5(1.2) 4 4(1.0) 0 0(0.0) 1 1(0.2) 32 27(6.7) CERVICAL SPINE EVENT 18 18(4.4) 5 5(1.2) 4 4(2.0) 3 3(0.7) 1 1(0.2) 2 2(0.5) 33 31(7.7) DEATH 3 3(0.7) 0 0(0.0) 1 1(0.2) 3 3(0.7) 0 0(0.0) 0 0(0.0) 7 7(1.7) DEPRESSION 25 25(6.2) 2 2(0.5) 2 2(0.5) 1 1(0.2) 3 3(0.7) 0 0(0.0) 33 32(7.9) ELEVATED TEMPERATURE WITHOUT INFECTION 6 6(1.5) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 6 6(1.5) GASTROINTESTINAL – ILEUS 21 21(5.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 21 21(5.2) GASTROINTESTINAL – OTHER 78 59(14.6) 12 12(3.0) 10 7(1.7) 7 6(1.5) 7 7(1.7) 1 1(0.2) 115 78(19.3) IMPLANT DISPLACEMENT/LOOSENING/MALPOSITION 1 1(0.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2) INCISION – RELATED 50 48(11.9) 3 3(0.7) 1 1(0.2) 1 1(0.2) 0 0(0.0) 0 0(0.0) 55 52(12.8) INFECTION 28 24(5.9) 8 6(1.5) 5 5(1.2) 7 7(1.7) 5 4(1.0) 2 1(0.2) 55 37(9.1) LEG PAIN 41 41(10.1) 2 2(0.5) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 43 42(10.4) LOWER EXTREMITY PAIN 40 37(9.1) 13 12(3.0) 7 7(1.7) 8 8(2.0) 5 4(1.0) 2 1(0.2) 75 62(15.3) NECK AND/OR ARM PAIN 12 12(3.0) 4 4(1.0) 5 5(1.2) 2 2(0.5) 0 0(0.0) 0 0(0.0) 23 22(5.4) NEUROLOGICAL 185 138(34.1) 12 11(2.7) 15 13(3.2) 12 11(2.7) 4 3(0.7) 0 0(0.0) 228 153(37.8) NON-UNION 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) OTHER 93 75(18.5) 28 25(6.2) 25 21(5.2) 26 21(5.2) 12 10(2.5) 1 1(0.2) 186 119(29.4) OTHER PAIN 86 60(14.8) 15 12(3.0) 5 5(1.2) 6 6(1.5) 7 4(1.0) 2 2(0.5) 101 80(19.8) PAIN MANAGEMENT CONTROL 7 7(1.7) 1 1(0.2) 1 1(0.2) 1 1(0.2) 0 0(0.0) 0 0(0.0) 10 10(2.5) PERITONEAL TEAR 3 3(0.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 3 3(0.7) RASH 7 7(1.7) 0 0(0.0) 1 1(0.2) 0 0(0.0) 1 1(0.2) 0 0(0.0) 9 9(2.2) RESPIRATORY 9 8(2.0) 3 3(0.7) 1 1(0.2) 1 1(0.2) 2 1(0.2) 0 0(0.0) 18 14(3.5) RETROGRADE EJACULATION 4 4(1.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(1.0) SPINAL EVENT 14 14(3.5) 0 0(0.0) 0 0(0.0) 2 2(0.5) 1 1(0.2) 0 0(0.0) 17 15(3.7) SPINAL EVENT AT OTHER LUMBAR LEVEL (S) 18 17(4.2) 5 5(1.2) 4 4(1.0) 4 4(1.0) 3 3(0.7) 1 1(0.2) 35 33(9.1) SPINAL EVENT AT TARGET LEVEL 5 5(1.2) 1 1(0.2) 2 2(0.5) 0 0(0.0) 0 0(0.0) 0 0(0.0) 8 8(2.0) SPINAL EVENT AT THORACIC LEVEL 9 9(2.2) 1 1(0.2) 2 2(0.5) 0 0(0.0) 2 2(0.5) 0 0(0.0) 14 13(3.2) SUBSIDENCE 14 14(3.5) 3 3(0.7) 0 0(0.0) 3 3(0.7) 1 1(0.2) 0 0(0.0) 21 21(5.2) TRAUMA 145 110(27.2) 30 25(6.2) 20 18(4.4) 20 16(4.0) 7 7(1.7) 1 1(0.2) 223 143(35.3) UPPER EXTREMITY PAIN 14 13(3.2) 10 10(2.5) 2 2(0.5) 5 5(1.2) 1 1(0.2) 0 0(0.0) 32 28(7.2) UROGENITAL 41 39(9.6) 9 9(2.2) 3 3(0.7) 8 7(1.7) 1 1(0.2) 1 1(0.2) 63 54(13.3) VASCULAR EVENT POST – OP PERIPHERAL 4 4(1.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.2) 0 0(0.0) 5 5(1.2) VASCULAR INJURY INTRA-OP 15 15(3.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 13 15(3.7) VERTEBRAL FRACTURE 3 3(0.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 3 3(0.7)
228 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Control (N=172) Operative 1 day – 4 wks 6 weeks 3 months 6 months 12 months 24 months Adverse Event Type Patient Patient Patient Patient Patient Patient Patient Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Patients Who Had Any Adverse Events 19 15(8.7) 78 61(35.5) 44 39(22.7) 57 54(31.4) 70 54(31.4) 108 77(44.8) 71 48(27.9) ALLERGIC REACTION 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 1 1(0.6) 1 1(0.6) ANATOMICAL/TECHNICAL DIFFICULTY 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) BACK AND LEG PAIN 0 0(0.0) 2 2(1.2) 1 1(0.6) 4 4(2.3) 1 1(0.6) 4 4(2.3) 3 3(1.7) BACK PAIN 0 0(0.0) 4 4(2.3) 2 2(1.2) 9 9(5.2) 8 8(4.7) 16 16(9.3) 7 7(4.1) CANCER 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 2 2(1.2) CARDIOVASCULAR 0 0(0.0) 2 2(1.2) 0 0(0.0) 0 0(0.0) 1 1(0.6) 3 3(1.7) 2 2(1.2) CERVICAL SPINE EVENT 1 1(0.6) 0 0(0.0) 0 0(0.0) 1 1(0.6) 1 1(0.6) 4 4(2.3) 2 2(1.2) DEATH 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) DEPRESSION 0 0(0.0) 1 1(0.6) 2 2(1.2) 0 0(0.0) 4 4(2.3) 2 2(1.2) 0 0(0.0) ELEVATED TEMPERATURE WITHOUT INFECTION 0 0(0.0) 3 3(1.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) GASTROINTESTINAL – ILEUS 2 2(1.2) 2 2(1.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) GASTROINTESTINAL – OTHER 0 0(0.0) 10 10(5.8) 1 1(0.6) 0 0(0.0) 1 1(0.6) 2 2(1.2) 1 1(0.6) IMPLANT DISPLACEMENT/LOOSENING/MALPOSITION 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) INCISION – RELATED 0 0(0.0) 6 6(3.5) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) INFECTION 0 0(0.0) 3 3(1.7) 1 1(0.6) 1 1(0.6) 2 2(1.2) 2 2(1.2) 6 4(2.3) LEG PAIN 0 0(0.0) 4 4(2.3) 1 1(0.6) 2 2(1.2) 3 3(1.7) 0 0(0.0) 3 3(1.7) LOWER EXTREMITY PAIN 0 0(0.0) 8 8(.47) 3 3(1.7) 0 0(0.0) 3 3(1.7) 4 4(2.3) 3 3(1.7) NECK AND/OR ARM PAIN 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 3 3(1.7) 1 1(0.6) NEUROLOGICAL 2 1(0.6) 9 9(5.2) 14 13(7.6) 13 13(7.6) 14 11(6.4) 9 9(5.2) 10 10(5.8) NON-UNION 0 0(0.0) 0 0(0.0) 2 2(1.2) 0 0(0.0) 1 1(0.6) 4 4(2.3) 0 0(0.0) OTHER 1 1(0.6) 4 3(1.7) 5 5(2.9) 1 1(0.6) 3 3(1.7) 11 10(5.8) 3 3(1.7) OTHER PAIN 0 0(0.0) 0 0(0.0) 1 1(0.6) 3 3(1.7) 3 3(1.7) 3 2(1.2) 4 2(1.2) PAIN MANAGEMENT CONTROL 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 2 2(1.2) 1 1(0.6) 0 0(0.0) PERITONEAL TEAR 2 2(1.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) RASH 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) RESPIRATORY 0 0(0.0) 1 1(0.6) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) RETROGRADE EJACULATION 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) SPINAL EVENT 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) SPINAL EVENT AT OTHER LUMBAR LEVEL (S) 0 0(0.0) 1 1(0.6) 0 0(0.0) 4 4(2.3) 1 1(0.6) 6 6(3.5) 3 3(1.7) SPINAL EVENT AT TARGET LEVEL 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) SPINAL EVENT AT THORACIC LEVEL 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) SUBSIDENCE 0 0(0.0) 3 3(1.7) 4 4(2.3) 3 3(1.7) 3 3(1.7) 1 1(0.6) 0 0(0.0) TRAUMA 0 0(0.0) 7 7(4.1) 3 3(1.7) 7 7(4.1) 13 12(7.0) 20 19(11.0) 15 14(8.1) UPPER EXTREMITY PAIN 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 1 1(0.6) 5 5(2.9) 2 2(1.2) UROGENITAL 2 2(1.2) 4 4(2.3) 1 1(0.6) 4 4(2.3) 1 1(0.6) 2 2(1.2) 3 2(1.2) VASCULAR EVENT POST – OP PERIPHERAL 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 1 1(0.6) 0 0(0.0) VASCULAR INJURY INTRA-OP 8 8(4.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) VERTEBRAL FRACTURE 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0)
229 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Control (N=172) Total 36 months 48 months 60 months 72 months 84 months Total (All) (<=24 months) Adverse Event Type Patient Patient Patient Patient Patient Patient Patient Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Event n (%) Patients Who Had Any Adverse Events 447 153(89.0) 79 47(27.3) 55 39(22.7) 71 45(28.2) 27 17(9.9) 7 6(3.5) 696 165(95.9) ALLERGIC REACTION 4 4(2.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 5 5(2.9) ANATOMICAL/TECHNICAL DIFFICULTY 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.5) BACK AND LEG PAIN 15 14(8.1) 3 3(1.7) 4 4(2.3) 2 2(1.2) 1 1(0.6) 0 0(0.0) 25 22(12.8) BACK PAIN 46 46(26.7) 8 8(4.7) 8 8(4.7) 6 6(3.5) 3 3(1.7) 1 1(0.6) 72 65(37.8) CANCER 3 3(1.7) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 4 3(1.7) CARDIOVASCULAR 8 8(4.7) 3 3(1.7) 1 1(0.6) 2 2(1.2) 0 0(0.0) 0 0(0.0) 14 12(7.0) CERVICAL SPINE EVENT 9 8(4.7) 2 2(1.2) 1 1(0.6) 3 3(1.7) 0 0(0.0) 0 0(0.0) 15 13(7.6) DEATH 1 1(0.6) 0 0(0.0) 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 2 2(1.2) DEPRESSION 9 9(5.2) 2 2(1.2) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 13 12(7.0) ELEVATED TEMPERATURE WITHOUT INFECTION 3 3(1.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 3 3(1.7) GASTROINTESTINAL – ILEUS 4 4(2.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(2.3) GASTROINTESTINAL – OTHER 15 15(8.7) 5 5(2.9) 1 1(0.6) 4 4(2.3) 1 1(0.6) 0 0(0.0) 26 23(13.4) IMPLANT DISPLACEMENT/LOOSENING/MALPOSITION 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.6) INCISION – RELATED 8 8(4.7) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 10 10(5.8) INFECTION 15 12(7.0) 2 2(1.2) 0 0(0.0) 5 4(2.3) 0 0(0.0) 0 0(0.0) 22 17(9.9) LEG PAIN 13 13(7.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 14 14(8.1) LOWER EXTREMITY PAIN 21 21(12.2) 2 2(1.2) 2 2(1.2) 3 3(1.7) 2 2(1.2) 0 0(0.0) 30 27(15.7) NECK AND/OR ARM PAIN 5 5(2.9) 0 0(0.0) 2 2(1.2) 3 3(1.7) 0 0(0.0) 0 0(0.0) 10 10(5.8) NEUROLOGICAL 71 59(34.3) 14 11(6.4) 8 7(4.1) 9 9(5.2) 4 4(2.3) 1 1(0.6) 107 76(44.2) NON-UNION 7 7(4.1) 0 0(0.0) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 8 8(4.7) OTHER 28 24(14.0) 10 7(4.1) 10 10(5.8) 7 7(4.1) 4 4(2.3) 1 1(0.6) 60 42(24.4) OTHER PAIN 14 11(8.4) 3 3(1.7) 5 5(2.9) 2 2(1.2) 0 0(0.0) 1 1(0.6) 25 19(11.0) PAIN MANAGEMENT CONTROL 4 4(2.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(2.3) PERITONEAL TEAR 2 2(1.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(1.2) RASH 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) RESPIRATORY 3 3(1.7) 1 1(0.6) 1 1(0.6) 0 0(0.0) 2 2(1.2) 0 0(0.0) 7 7(4.1) RETROGRADE EJACULATION 2 2(1.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(1.2) SPINAL EVENT 2 2(1.2) 1 1(0.6) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 3 3(1.7) SPINAL EVENT AT OTHER LUMBAR LEVEL (S) 15 15(6.7) 2 2(1.2) 3 3(1.7) 2 2(1.2) 2 2(1.2) 2 2(1.2) 26 25(14.5) SPINAL EVENT AT TARGET LEVEL 1 1(0.6) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 3 3(1.7) SPINAL EVENT AT THORACIC LEVEL 2 2(1.2) 0 0(0.0) 1 1(0.6) 1 1(0.6) 0 0(0.0) 0 0(0.0) 4 4(2.3) SUBSIDENCE 14 14(8.1) 0 0(0.0) 1 1(0.6) 2 2(1.2) 0 0(0.0) 0 0(0.0) 17 17(9.9) TRAUMA 65 55(32.0) 12 12(7.0) 6 6(3.5) 7 6(3.5) 3 1(0.6) 0 0(0.0) 93 69(40.1) UPPER EXTREMITY PAIN 9 9(5.2) 3 3(1.7) 1 1(0.6) 4 4(2.3) 1 1(0.6) 1 1(0.6) 19 19(11.0) UROGENITAL 17 16(9.3) 5 4(2.3) 5 5(2.9) 4 4(2.3) 2 2(1.2) 0 0(0.0) 33 29(16.8) VASCULAR EVENT POST – OP PERIPHERAL 2 2(1.2) 0 0(0.0) 2 2(1.2) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(2.3) VASCULAR INJURY INTRA-OP 8 8(4.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 8 8(4.7) VERTEBRAL FRACTURE 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) CRD: Numbers presented in italic font were difficult to read in the original Medtronic CSR due to poor print quality. Inaccuracies may have occurred during copying.
230 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 43 Adverse events summary table from INFUSE®/TELAMON PEEK Instrumented PLIF pilot, single-arm study.
TELAMON P(TM) Implant/INFUSE™ Bone Graft/CD HORIZON® Spinal System Pilot Study Summary of All Adverse Events Investigational (N=30) Operative 1 day – 4 wks 6 weeks 3 months 6 months Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 1 1(3.3) 17 15(50.0) 12 11(36.7) 12 12(40.0) 18 16(53.3) Accidental Injury/Muscle Strain 0 0(0.0) 1 1(3.3) 2 2(6.7) 1 1(3.3) 2 2(6.7) Back and/or Leg Pain 0 0(0.0) 3 3(10.0) 4 4(13.3) 2 2(6.7) 7 7(23.3) Cardiovascular (MI, CVA, TIA) 0 0(0.0) 2 2(6.7) 0 0(0.0) 0 0(0.0) 0 0(0.0) Depression 0 0(0.0) 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) Dural Inuury 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Elevated Temperature w/o Infection 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) Gastrointestinal (Other) 0 0(0.0) 2 2(6.7) 1 1(3.3) 1 1(3.3) 0 0(0.0) Hematological 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) Incision Related 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Infection 0 0(0.0) 2 2(6.7) 0 0(0.0) 0 0(0.0) 1 1(3.3) Lower Extremity Pain not of Back Etiology 0 0(0.0) 0 0(0.0) 1 1(3.3) 2 2(6.7) 1 1(3.3) Neurological 0 0(0.0) 0 0(0.0) 1 1(3.3) 0 0(0.0) 1 1(3.3) Other 0 0(0.0) 3 3(10.0) 2 2(6.7) 3 3(10.0) 1 1(3.3) Other Pain 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) Respiratory 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 0 0(0.0) Spinal Event (at other cervical level) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) Spinal Event (at other lumbar levels) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Spinal Event (at target level (s)) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) Trauma 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Upper Extemity Pain not of Back Etiology 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Urogenital 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) Vascular Event Postop Peripheral 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0)
231 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Investigational (N=30) 12 months 24 months 36 months 48 months Total (All) Total (<-24 months) Event Patient Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 24 14(46.7) 13 8(26.7) 5 5(16.7) 1 1(3.3) 103 29(96.7) 97 29(96.7) Accidental Injury/Muscle Strain 3 3(10.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 9 8(26.7) 9 8(26.7) Back and/or Leg Pain 7 6(20.0) 1 1(3.3) 1 1(3.3) 1 1(3.3) 26 22(73.3) 24 20(66.7) Cardiovascular (MI, CVA, TIA) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(6.7) 2 2(6.7) Depression 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 2 2(6.7) 2 2(6.7) Dural Injury 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Elevated Temperature w/o Infection 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Gastrointestinal (Other) 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 5 5(16.7) 5 5(16.7) Hematological 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Incision Related 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Infection 1 1(3.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 3(10.0) 4 3(10.0) Lower Extremity Pain not of Back Etiology 2 2(6.7) 2 2(6.7) 0 0(0.0) 0 0(0.0) 8 8(26.7) 8 8(26.7) Neurological 5 4(13.3) 1 1(3.3) 0 0(0.0) 0 0(0.0) 8 6(20.0) 8 6(20.0) Other 1 1(3.3) 4 3(10.0) 3 3(10.0) 0 0(0.0) 17 13(43.3) 14 11(36.7) Other Pain 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Respiratory 1 1(3.3) 1 1(3.3) 0 0(0.0) 0 0(0.0) 3 2(6.7) 3 2(6.7) Spinal Event (at other cervical level) 0 0(0.0) 1 1(3.3) 0 0(0.0) 0 0(0.0) 2 2(6.7) 2 2(6.7) Spinal Event (at other lumbar levels) 2 2(6.7) 1 1(3.3) 0 0(0.0) 0 0(0.0) 3 3(10.0) 3 3(10.0) Spinal Event (at target level (s)) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Trauma 1 1(3.3) 0 0(0.0) 1 1(3.3) 0 0(0.0) 4 3(10.0) 3 3(10.0) Upper Extemity Pain not of Back Etiology 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(6.7) 2 2(6.7) Urogenital 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3) Vascular Event Postop Peripheral 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(3.3) 1 1(3.3)
232 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Adverse events tables of AMPLIFY™ trials from clinical study reports
Table 44 Adverse events summary table from AMPLIFY™ (rhBMP-2/CRM) pivotal RCT/
Summary of All Adverse Events
Investigational (N=239) Operative 1 day – 4 wks 6 weeks 3 months 6 months 12 months Event Patient Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 21 20 (8.4) 184 114 (47.7) 33 31(13.0) 82 62(25.9) 96 72 (30.1) 198 107(44.8) Anatomical/Technical Difficulty 1 1 (0.4) 0 0 (0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Arthritis/Bursitis 0 0 (0.0) 3 3(1.3) 1 1(0.4) 7 7(2.9) 4 4(1.7) 6 6(2.5) Back and/or Leg Pain 0 0 (0.0) 18 18 (7.5) 11 10(4.2) 15 15(6.3) 21 20(8.4) 36 33(13.8) Cancer 0 0(0.0) 0 0 (0.0) 1 1(0.4) 1 1(0.4) 2 2(0.8) 4 4(1.7) Cardiovascular 2 2 (0.8) 43 37 (15.5) 0 0(0.0) 4 4(1.7) 2 2(0.8) 15 13(5.4) Carpal Tunnel Syndrome 0 0 (0.0) 0 0 (0.0) 0 0(0.0) 0 0(0.0) 2 2(0.8) 4 4(1.7) Death 0 0 (0.0) 0 0 (0.0) 1 1(0.4) 0 0(0.0) 1 1(0.4) 1 1(0.4) Dural Injury 13 13 (5.4) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Gastrointestinal 0 0 (0.0) 19 18(7.5) 0 0(0.0) 4 4(1.7) 5 5(2.1) 10 10(4.2) Graft Site Related 0 0 (0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Implant Displacement/Loosening 0 0 (0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) Infection 0 0 (0.0) 20 19(7.9) 4 3(1.3) 4 4(1.7) 6 4(1.7) 11 11(4.6) Malpositioned Implant 1 1 (0.4) 3 3(1.3) 0 0(0.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) Neurological 0 0 (0.0) 9 9(3.8) 2 2(0.8) 19 17(7.1) 17 17(7.1) 21 19(7.9) Non-Union 0 0 (0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) 0 0(0.0) 4 4(1.7) Non-Union Pending 0 0 (0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(1.7) Other 1 1 (0.4) 43 35(14.6) 7 7(2.9) 7 6(2.5) 7 6(2.5) 18 14(5.9) Other Pain 0 0 (0.0) 2 2(0.8) 2 2(0.8) 1 1(0.4) 6 6(2.5) 11 11(4.6) Respiratory 0 0 (0.0) 8 8(3.3) 0 0(0.0) 1 1(0.4) 0 0(0.0) 5 4(1.7) Spinal Event - All 0 0 (0.0) 3 3(1.3) 0 0(0.0) 4 4(1.7) 6 5(2.1) 7 7(2.9) Spinal Event - Cervical 0 0 (0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) 4 4(1.7) 3 3(1.3) Spinal Event – Lumbar 0 0 (0.0) 3 3(1.3) 0 0(0.0) 3 3(1.3) 2 2(0.8) 3 3(1.3) Spinal Event - Thoracic 0 0 (0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) Trauma 0 0 (0.0) 2 2(0.8) 2 2(0.8) 8 8(3.3) 13 12(5.0) 35 32(13.4) Urogenital 0 0 (0.0) 10 10(4.2) 2 2(0.8) 5 5(2.1) 4 4(1.7) 5 5(2.1) Vertebral Fracture 3 3 (1.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0)
233 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Investigational (N=239) 24 months Total 36 months 48 months 60 months Total (<=24 months) (All) Event Patient Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 154 93(38.9) 768 209(87.4) 167 94(39.3) 175 94(39.3) 105 66 (27.6) 1215 223(93.3) Anatomical/Technical Difficulty 0 0(0.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) Arthritis/Bursitis 5 3(1.3) 26 24(10.0) 9 9(3.8) 5 5(2.1) 4 4(1.7) 44 33(13.8) Back and/or Leg Pain 35 33(13.8) 136 103(43.1) 34 29(12.1) 32 30(12.6) 14 12(5.0) 216 133(55.6) Cancer 3 3(1.3) 11 10(4.2) 4 4(1.7) 0 0(0.0) 0 0(0.0) 15 12(5.0) Cardiovascular 4 4(1.7) 70 53(22.2) 18 16(6.7) 12 11(4.6) 8 7(2.9) 108 77(32.2) Carpal Tunnel Syndrome 3 3(1.3) 9 9(3.8) 0 0(0.0) 0 0(0.0) 1 1(0.4) 10 10(4.2) Death 0 0(0.0) 3 3(1.3) 0 0(0.0) 3 3(1.3) 1 1(0.4) 7 7(2.9) Dural Injury 0 0(0.0) 14 14(5.9) 0 0(0.0) 0 0(0.0) 1 1(0.4) 15 15(6.3) Gastrointestinal 8 7(2.9) 46 40(16.7) 14 13(5.4) 15 13(5.4) 11 10(4.2) 86 65(27.2) Graft Site Related 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Implant Displacement/Loosening 0 0(0.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) Infection 7 6(2.5) 52 39(16.3) 2 2(0.8) 7 7(2.9) 3 3(1.3) 64 48(20.1) Malpositioned Implant 0 0(0.0) 5 5(2.1) 0 0(0.0) 0 0(0.0) 0 0(0.0) 5 5(2.1) Neurological 16 15(6.3) 84 68(28.5) 14 12(5.0) 13 13(5.4) 9 9(3.8) 120 88(36.8) Non-Union 0 0(0.0) 5 5(2.1) 1 1(0.4) 0 0(0.0) 0 0(0.0) 6 6(2.5) Non-Union Pending 1 1(0.4) 5 5(2.1) 0 0(0.0) 0 0(0.0) 0 0(0.0) 5 5(2.1) Other 20 17(7.1) 103 70(29.3) 28 20(8.4) 42 21(8.8) 20 14(5.9) 193 89(37.2) Other Pain 8 8(3.3) 30 29(12.1) 6 6(2.5) 12 12(5.0) 10 9(3.8) 58 48(20.1) Respiratory 4 4(1.7) 18 16(6.7) 0 0(0.0) 2 2(0.8) 1 1(0.4) 21 19(7.9) Spinal Event - All 9 9(3.8) 29 24(10.0) 7 7(2.9) 8 8(3.3) 6 6(2.5) 50 43(18.0) Spinal Event - Cervical 3 3(1.3) 11 10(4.2) 2 2(0.8) 1 1(0.4) 4 4(1.7) 18 17(7.1) Spinal Event – Lumbar 6 6(2.5) 17 14(5.9) 5 5(2.1) 7 7(2.9) 2 2(0.8) 31 28(11.7) Spinal Event - Thoracic 0 0(0.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) Trauma 29 27(11.3) 89 68(28.5) 24 22(9.2) 18 16(6.7) 14 13(5.4) 145 99(41.4) Urogenital 2 2(0.8) 28 27(11.3) 6 5(2.1) 6 6(2.5) 2 2(0.8) 42 37(15.5) Vertebral Fracture 0 0(0.0) 3 3(1.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 3 3(1.3)
234 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Control (N=224) Operative 1 day – 4 wks 6 weeks 3 months 6 months 12 months Event Patient Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 22 21(9.4) 161 90(40.2) 49 35(15.6) 74 65(29.0) 111 88(39.3) 153 100(44.6) Anatomical/Technical Difficulty 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Arthritis/Bursitis 0 0(0.0) 1 1(0.4) 1 1(0.4) 2 2(0.9) 2 2(0.9) 6 6(2.7) Back and/or Leg Pain 0 0(0.0) 8 7(3.1) 5 5(2.2) 12 12(5.4) 30 28(12.5) 26 24(10.7) Cancer 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) 1 1(0.4) Cardiovascular 0 0(0.0) 43 39(17.4) 2 2(0.9) 3 3(1.3) 8 8(3.6) 9 8(3.6) Carpal Tunnel Syndrome 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) 3 3(1.3) Death 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(0.9) 1 1(0.4) Dural Injury 18 18(8.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Gastrointestinal 0 0(0.0) 16 15(6.7) 2 2(0.9) 3 3(1.3) 2 2(0.9) 11 6(2.7) Graft Site Related 0 0(0.0) 4 4(1.8) 3 3(1.3) 5 5(2.2) 3 3(1.3) 2 2(0.9) Implant Displacement/Loosening 0 0(0.0) 0 0(0.0) 0 0(0.0) 1 1(0.4) 0 0(0.0) 1 1(0.4) Infection 0 0(0.0) 27 24(10.7) 6 6(2.7) 2 2(0.9) 1 1(0.4) 10 10(4.5) Malpositioned Implant 0 0(0.0) 1 1(0.4) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) Neurological 0 0(0.0) 6 5(2.2) 8 7(3.1) 13 12(5.4) 17 17(7.6) 14 14(6.3) Non-Union 0 0(0.0) 0 0(0.0) 0 0(0.0) 7 7(3.1) 4 4(1.8) 5 5(2.2) Non-Union Pending 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(0.9) 2 2(0.9) 1 1(0.4) Other 1 1(0.4) 34 25(11.2) 7 4(1.8) 8 8(3.6) 11 10(4.5) 14 11(4.9) Other Pain 0 0(0.0) 3 3(1.3) 0 0(0.0) 1 1(0.4) 5 5(2.2) 4 4(1.8) Respiratory 0 0(0.0) 7 7(3.1) 1 1(0.4) 1 1(0.4) 0 0(0.0) 3 3(1.3) Spinal Event - All 0 0(0.0) 3 3(1.3) 3 2(0.9) 4 4(1.8) 2 2(0.9) 18 16(7.1) Spinal Event - Cervical 0 0(0.0) 0 0(0.0) 2 1(0.4) 1 1(0.4) 1 1(0.4) 9 9(4.0) Spinal Event – Lumbar 0 0(0.0) 2 2(0.9) 1 1(0.4) 3 3(1.3) 1 1(0.4) 9 8(3.6) Spinal Event - Thoracic 0 0(0.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Trauma 0 0(0.0) 2 2(0.9) 8 8(3.6) 7 7(3.1) 17 16(7.1) 18 16(7.1) Urogenital 0 0(0.0) 6 6(2.7) 2 2(0.9) 3 3(1.3) 3 3(1.3) 6 5(2.2) Vertebral Fracture 3 3(1.3) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0)
235 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Control (N=224) 24 months Total 36 months 48 months 60 months Total (<=24 months) (All) Event Patient Event Patient Event Patient Event Patient Event Patient Event Patient Adverse Event Type n(%) n(%) n(%) n(%) n(%) n(%) Patients Who Had Any Adverse Events 128 87(38.8) 698 197(87.9) 146 84(37.5) 138 71(31.7) 85 57(25.4) 1067 211(94.2) Anatomical/Technical Difficulty 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) 0 0(0.0) Arthritis/Bursitis 6 6(2.7) 18 16(7.1) 7 7(3.1) 8 7(3.1) 5 5(2.2) 38 28(12.5) Back and/or Leg Pain 24 24(10.7) 105 86(38.4) 35 30(13.4) 25 23(10.3) 18 17(7.6) 183 122(54.5) Cancer 0 0(0.0) 2 2(0.9) 1 1(0.4) 2 2(0.9) 0 0(0.0) 5 5(2.2) Cardiovascular 3 3(1.3) 68 55(24.6) 3 3(1.3) 13 10(4.5) 4 4(1.8) 88 65(29.0) Carpal Tunnel Syndrome 2 2(0.9) 6 6(2.7) 1 1(0.4) 0 0(0.0) 2 2(0.9) 9 9(4.0) Death 1 1(0.4) 4 4(1.8) 1 1(0.4) 1 1(0.4) 2 2(0.9) 8 8(3.6) Dural Injury 0 0(0.0) 18 18(8.0) 1 1(0.4) 0 0(0.0) 1 1(0.4) 20 19(8.5) Gastrointestinal 9 9(4.0) 43 33(14.7) 11 11(4.9) 14 10(4.5) 6 5(2.2) 74 55(24.6) Graft Site Related 0 0(0.0) 17 17(7.6) 2 2(0.9) 0 0(0.0) 0 0(0.0) 19 19(8.5) Implant Displacement/Loosening 0 0(0.0) 2 2(0.9) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(0.9) Infection 5 5(2.2) 51 45(20.1) 6 6(2.7) 7 7(3.1) 3 3(1.3) 67 54(24.1) Malpositioned Implant 0 0(0.0) 2 2(0.9) 0 0(0.0) 0 0(0.0) 0 0(0.0) 2 2(0.9) Neurological 13 11(4.9) 71 58(25.9) 14 11(4.9) 9 9(4.0) 3 2(0.9) 97 69(30.8) Non-Union 2 2(0.9) 18 18(8.0) 1 1(0.4) 0 0(0.0) 0 0(0.0) 19 19(8.5) Non-Union Pending 1 1(0.4) 6 6(2.7) 0 0(0.0) 0 0(0.0) 2 2(0.9) 8 8(3.6) Other 18 15(6.7) 93 62(27.7) 26 20(8.9) 25 15(6.7) 13 11(4.9) 157 83(37.1) Other Pain 18 17(7.6) 31 28(12.5) 13 12(5.4) 7 7(3.1) 8 7(3.1) 59 44(19.6) Respiratory 1 1(0.4) 13 12(5.4) 0 0(0.0) 5 4(1.8) 2 2(0.9) 20 15(6.7) Spinal Event - All 2 2(0.9) 32 28(12.5) 5 5(2.2) 5 5(2.2) 3 2(0.9) 45 36(16.1) Spinal Event - Cervical 0 0(0.0) 13 12(5.4) 3 3(1.3) 2 2(0.9) 1 1(0.4) 19 18(8.0) Spinal Event – Lumbar 2 2(0.9) 18 17(7.6) 2 2(0.9) 2 2(0.9) 2 2(0.9) 24 23(10.3) Spinal Event - Thoracic 0 0(0.0) 1 1(0.4) 0 0(0.0) 1 1(0.4) 0 0(0.0) 2 2(0.9) Trauma 17 16(7.1) 69 58(25.9) 16 14(6.3) 15 12(5.4) 8 7(3.1) 108 79(35.3) Urogenital 5 5(2.2) 25 22(9.8) 3 3(1.3) 2 2(0.9) 5 5(2.2) 35 30(13.4) Vertebral Fracture 1 1(0.4) 4 4(1.8) 0 0(0.0) 0 0(0.0) 0 0(0.0) 4 4(1.8)
236 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 45 Adverse events summary table from rhBMP-2/CRM 2-level pilot, single-arm study. rhBMP-2/CRM/CD HORIZON® Spinal System Pilot Study (2-Level) (G040219) IDE Final Progress Report 2010 Table 2: Summary of All Adverse Events
Complication Operative Postop 6 Weeks 3 Months 6 Months 12 Months 24 Months 36 Months Total #of (1 day – (≥4 Wks - (≥2 Months - (≥5 Months - (≥9 Months - (≥19 Months (≥30 Months Adverse Patients <4 Weeks) <2 Months) <5 Months <9 Months) <19 Months) - <30 - <42 Months Events Reporting Months) # (% of 29) Accidental Injury/Muscle Strain 2 2 1 1 2 8 7 24.1 Allergic Reaction 2 2 2 6.9 Back and/or Leg Pain 1 2 6 5 3 1 18 11 37.9 Cardiovascular 1 1 2 2 6.9 Cervical Spine Event 2 1 3 6 4 13.8 Dural Injury 1 1 1 3.4 Electrolyte Imbalance 1 1 1 3.4 Elevated Temperature without Infection 1 1 1 3.4 Gastrointestinal – Other 2 1 1 3 7 6 20.7 Hematological 1 1 2 2 6.9 Incision Related 2 3 1 1 7 5 17.2 Infection 1 1 1 3.4 Lower Extremity Pain Not of Back Etiology 2 3 1 1 7 6 20.7 Neck and/or Arm Pain 1 1 2 2 6.9 Neurological 2 2 2 1 4 11 9 31.0 Non-union Outcome Pending 1 1 1 3.4 Other 3 3 1 7 7 24.1 Respiratory 1 1 1 3.4 Spinal Event at Other Lumbar Levels 4 1 2 7 7 24.1 Spinal Event at Target Level 2 1 3 2 6.9 Trauma 1 1 2 2 6.9 Upper Extremity Pain Not of Neck Etiology 1 1 1 1 1 1 6 5 17.2 Urogenital 1 1 2 4 4 13.8 Vascular Event Post-op Peripheral 1 1 1 3.4 TOTAL 3 17 11 9 11 25 21 11 108 26 89.7
237 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Adverse events tables of rhBMP-2/BCP trials from clinical study reports
Table 46 Adverse events summary table from rhBMP-2/BCP Mexico pilot.
TABLE V rhBMP-2/BCP Mexico – Cohort # 1 & 2 ADVERSE EVENTS
TOTALS Complication Operative Post-operative 6 Weeks 3 Month 6 Month 12 Month Cohort #1 Cohort #2 [Number (%)] (1 day - ≤5 (>5 - ≤7 weeks) (>7 wks - ≤5 mos) (>5 - ≤9 months) (>9 - ≤12 months) weeks) [Number (%)] [Number (%)] [Number (%)] [Number (%)] [Number (%)] Cohort Cohort Cohort Cohort Cohort Cohort Cohort Cohort Cohort Cohort Cohort Cohort Total No. of Total No. of #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 Events Patients Events Patients N=7 N=8 N=7 N=8 N=7 N=8 N=7 N=8 N=7 N=8 N=7 N=8 (% of 7) (% of 8) Loose Screw 1 1 2 1 (14.3) 0 0 (0.0) (14.3) (14.3) Gastric Ulcer 1 1 1 (14.3) 0 0 (0.0) (14.3) Sacroiliitis 1 1 1 3 2 (28.6) 0 0 (0.0) (14.3) (14.3) (14.3) Stenosis 1 1 1 (14.3) 0 0 (0.0) (14.3) Bone Fracture 1 1 1 1 (14.3) 1 1 (12.5) (12.5) (14.3)
Removals 0 0 (0.0) 0 0 (0.0) Revisions 0 0 (0.0) 0 0 (0.0) Reoperations 1 1 1 (14.3) 0 0 (0.0) (14.3)
• 4 of 7 (57%) evaluable Cohort #1 patients had a total of 5 adverse events. One (1) of 8 (12.5%) evaluable patients from Cohort #2 had one adverse event. • If the same adverse event in a patient persisted at different time intervals, it was counted at each time period, e.g. the same sacroileitis for patient no 84011 was counted at two time intervals.
238 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 47 Adverse events summary table from rhBMP-2/BCP US pilot RCT.
Adverse Event Information rhBMP-2/BCP/TSRH Spinal System – U.S. IDE G980320 (as of February 25, 2003)
Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months # of Patients Reporting Total AEs (1day - <4Weeks) (≥4Wks - <9Wks) (≥9Wks - <5Mos) (≥5Mos - <9Mos) (≥9Mos - <19Mos) (≥19Mos - <30Mos) BMP/BCP/ BMP/BC Control N=11 TSRH BMP BMP BMP BMP BMP BMP BMP BMP P N=5 BMP BMP BMP BMP BMP BMP BMP BMP N=11 Complication 2 BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP BCP TSRH TSRH TSRH TSRH TSRH TSRH TSRH TSRH # % # % # % Control Control Control Control Control Control Control Control Anatomical/Technical 0 0 0 0 0.0% 0 0.0% 0 0.0% Difficulty Back and/or Leg Pain 2 1 5 2 1 5 6 0 3 27.3% 5 45.5% 0 0.0% Cancer 0 0 0 0 0.0% 0 0.0% 0 0.0% Cardiovascular 1 0 1 0 0 0.0% 1 9.1% 0 0.0% Death 0 0 0 0 0.0% 0 0.0% 0 0.0% Dural Injury 1 1 0 2 0 0 0.0% 2 18.2% 0 0.0% Dysphagia/dysphonia 1 0 1 0 0 0.0% 1 9.1% 0 0.0% Gastrointestinal 1 1 2 1 2 1 1 9.1% 2 18.2% 1 20.0% Graft Site Related 0 0 0 0 0.0% 0 0.0% 0 0.0% Implant Loosening/ 0 0 0 0 0.0% 0 0.0% 0 0.0% Displacement/Breakage Infection 1 1 2 0 0 2 18.2% 0 0.0% 0 0.0% Malpositioned Implant 0 0 0 0 0.0% 0 0.0% 0 0.0% Neurological 2 1 1 1 2 3 0 2 18.2% 0.0% 0 0.0% Non-Union 1 1 1 0 1 1 9.1% 0 0.0% 1 20.0% (OUTCOME PENDING) Other 1 1 1 1 1 1 1 1 2 4 2 2 18.2% 4 36.4% 2 40.0% Other Pain 1 1 1 1 3 1 0 2 18.2% 1 9.1% 0 0.0% Respiratory 1 1 0 0 1 9.1% 0 0.0% 0 0.0% Retrograde Ejaculation 0 0 0 0 0.0% 0 0.0% 0 0.0% Spinal Event 1 1 1 0 3 0 0 0.0% 2 18.2% 0 0.0% Subsidence 0 0 0 0 0.0% 0 0.0% 0 0.0% Trauma 1 1 1 1 2 0 1 9.1% 1 9.1% 0 0.0% Urogenital 1 0 0 1 0 0.0% 0 0.0% 1 20.0% Vascular Intra-Op 0 0 0 0 0.0% 0 0.0% 0 0.0% Vertebral Fracture 0 0 0 0 0.0% 0 0.0% 0 0.0% 1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table. 2 The clinical trial is a three-arm study with the following arms: 1) Investigational without instrumentation (rhBMP-2/BCP), 2) Investigational with instrumentation (rhBMP-2/BCP/TSRH Spinal System), and 3) Control (Autograft with TSRH Spinal System).
239 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs) Table 48 Adverse events summary table from rhBMP-2/BCP Canada pivotal RCT.
Table II rhBMP-2/BCP/TSRH Spinal System or rhBMP-2/BCP/CD Horizon Spinal System CANADA Application File No. 13006 - FINAL Surgery Postoperative 6 Weeks 3 Months 6 Months 12 Months 24 Months (1 day – 4 (≥4 Wks - < 9 (≥9 Wks - < 5 (≥5 Mos - <9 (≥9 Mos - <19 (≥19 Mos - < 30 Total AEs # of Patients Reporting Weeks) Wks) Mos) Mos) Mos) Mos) Invest N=98 Control N=99 Complication Invest Control Invest Control Invest Control Invest Control Invest Control Invest Control Invest Control Invest Control # % # % Anatomical/Technical 3 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 3 3.1% 0 0.0% Difficulty Back and/or Leg Pain 0 0 10 5 12 6 10 10 15 16 23 21 19 23 89 81 70 71.4% 71 71.7% Cancer 0 0 0 0 1 0 0 1 0 0 0 0 0 1 1 2 1 1.0% 2 2.0% Cardio/Vascular 1 0 10 9 1 0 1 0 2 1 2 7 2 1 19 18 16 16.3% 17 17.2% Cold symptoms 0 0 0 0 1 2 0 1 1 2 2 1 0 0 4 6 2 2.0% 4 4.0% Death 0 0 0 1 0 0 0 0 0 0 1 0 0 1 1 2 1 1.0% 2 2.0% Dural Injury 5 7 0 0 0 0 0 0 0 0 0 0 0 0 5 7 5 5.1% 7 7.1% Elevated Temperature 0 0 8 9 0 0 0 0 0 0 0 0 0 0 8 9 8 8.2% 9 9.1% Foot Pain 0 0 0 0 1 0 0 0 0 0 0 0 3 0 4 0 3 3.1% 0 0.0% Gastrointestinal 0 0 8 15 0 0 1 2 2 3 4 4 3 3 18 27 14 14.3% 23 23.2% Graft Site Related 0 0 0 2 0 1 0 1 0 5 0 1 0 1 0 11 0 0.0% 11 11.1% Gynecological 0 0 0 0 0 0 0 0 0 1 2 1 0 0 2 2 2 2.0% 2 2.0% Headache 0 0 1 0 0 0 0 0 2 0 0 0 1 0 4 0 4 4.1% 0 0.0% Hypokalemia 0 0 5 3 0 0 0 0 0 0 0 0 0 0 5 3 5 5.1% 3 3.0% Implant Loosening/ 1 0 0 1 0 0 0 0 0 1 0 0 1 0 2 2 2 2.0% 2 2.0% Displacement/ Breakage Infection 0 0 1 1 0 0 1 5 1 0 1 5 3 0 7 11 6 8.1% 7 7.1% Infection UTI 0 0 7 2 2 2 2 0 0 0 1 0 1 0 13 4 9 9.2% 4 4.0% Insomnia 0 0 1 0 3 0 0 0 0 0 0 0 0 0 4 0 4 4.1% 0 0.0% Low Hemoglobin 1 0 5 9 0 1 0 0 0 0 0 0 0 0 5 10 8 6.1% 10 10.1% Neck and/or Arm Pain 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 0 2 2.0% 0 0.0% Neck Pain 0 0 1 1 0 0 0 1 0 2 2 2 2 0 5 8 5 5.1% 5 8.1% Neurological 0 1 4 2 3 1 2 2 4 5 3 0 4 5 20 14 14 14.3% 14 14.1% Non-Union Outcome 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 2 0 0.0% 2 2.0% Pending Other 0 1 8 8 3 2 1 1 4 2 8 8 8 7 30 29 18 18.3% 23 23.2% Other Pain 0 0 3 0 0 3 1 0 1 2 4 2 5 1 15 8 13 13.3% 7 7.1% Respiratory 0 0 5 1 0 0 1 0 2 0 0 1 0 0 8 2 7 7.1% 2 2.0% Shoulder Pain 0 0 0 1 1 0 0 1 0 2 2 1 0 1 3 6 2 2.0% 5 5.1% Spinal Event 0 1 0 0 0 0 1 1 1 1 0 0 1 1 3 4 3 3.1% 4 4.0% Systemic 1 0 0 2 0 0 0 0 0 0 0 0 1 0 2 2 2 2.0% 2 2.0% Trauma 0 0 0 0 3 1 2 7 4 8 8 10 4 8 19 30 15 15.3% 21 21.2% Urogenital 1 0 12 17 1 0 0 0 0 1 4 0 2 3 20 21 19 19.4% 18 18.2% Vascular/Intra-Op 3 2 0 0 0 0 0 0 0 0 0 0 0 0 3 1 3 3.1% 1 1.0% Vertebral Fracture 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1.0% 0 0.0% Vertigo 0 0 1 0 0 0 0 0 2 0 0 0 0 0 3 0 3 3.1% 0 0.0% Wound complication 0 4 4 0 4 1 0 0 0 0 0 0 0 1 8 5 8 8.2% 8 6.1% Wound Infection 0 0 8 4 0 0 0 0 0 0 1 0 0 0 9 4 9 9.2% 4 4.0% 1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the non-union resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table.
240 Appendix XIV – Adverse events tables from Medtronic clinical study reports (CSRs)
Table II 48 and 72 Month Adverse Event Information rhBMP-2/BCP/TSRH Spinal System or rhBMP-2/BCP/CD/Horizon Spinal System CANADA Application File No. 13006 - FINAL 48 Months (≥30 Mos - <52 72 Months (≥52Mos - Total AEs # of Patients Reporting Mos) <84Mos) Complication Invest Control Invest Control Invest Control Invest N=50 Control N=50 # % # % Back and/or Leg Pain 11 10 4 3 15 13 12 24.0% 12 24.0% Cancer 0 1 1 0 1 1 1 2.0% 1 2.0% Cardio/Vascular 3 3 1 2 4 5 3 6.0% 5 10.0% Death 0 1 0 1 0 2 0 0.0% 2 4.0% Gastrointestinal 1 1 0 3 1 4 1 2.0% 4 8.0% Graft Site Related 0 0 0 1 0 1 0 0.0% 1 2.0% Gynecological 0 1 0 1 0 2 0 0.0% 1 2.0% Implant Loosening/ 0 0 0 1 0 1 0 0.0% 1 2.0% Displacement/Breakage Infection 0 1 0 1 0 2 0 0.0% 2 4.0% Neck and/or Arm Pain 1 0 0 0 1 0 1 2.0% 0 0.0% Neck Pain 0 0 1 0 1 0 1 2.0% 0 0.0% Neurological 4 6 0 2 4 8 4 8.0% 7 14.0% Other 3 3 1 2 4 5 4 8.0% 4 8.0% Other Pain 2 0 0 0 2 0 2 4.0% 0 0.0% Respiratory 1 0 0 0 1 0 1 2.0% 0 0.0% Shoulder Pain 0 0 1 0 1 0 1 2.0% 0 0.0% Systemic 0 1 0 0 0 1 0 0.0% 1 2.0% Trauma 2 4 0 4 2 8 2 4.0% 6 12.0% Urogenital 0 3 0 1 0 4 0 0.0% 3 6.0% Wound Infection 1 0 0 0 1 0 1 2.0% 0 0.0% 1 Since fusion is a primary effectiveness endpoint, nonunions reported as adverse events by the investigator are not included in the table if the nonunion resulted in a second surgery. These nonunions are captured in the secondary surgery table and the fusion table.
241 Appendix XV – Adverse events reported in Medtronic IPD
Appendix XV
Summary of all adverse events reported in Medtronic IPD
rhBMP-2 patients ICBG patients
Arm Study Arm Study
- -
Arm Study Arm Study
- -
ACDF Pilot RCT
PLIF RCT 2/CRM) Pivotal RCT 2/CRM) Pivotal RCT - - level Pilot, Single Pilot, level Single Pilot, level - - CAGE® Pilot RCT CAGE® Open Pivotal RCT Single Pivotal Lap CAGE® CAGE® Pilot RCT CAGE® Open Pivotal RCT Single Pivotal Lap CAGE® ------Y™ (rhBMP Y™ 2/BCP Pilot US RCT 2/BCP Canada Pivotal RCT 2/CRM 2 Pilot Mexico 2/BCP 2/BCP Pilot US RCT 2/BCP Canada Pivotal RCT 2/CRM 2 Pilot Mexico 2/BCP Arm Arm ------
Adverse event INFUSE®/LT INFUSE®/LT INFUSE®/LT INFUSE® Bone Dowel Pilot RCT INFUSE® Bone Dowel Pivotal RCT INFUSE®/INTER FIX™ PLIF RCT INFUSE®/CORNERSTONE® INFUSE®/Mastergraft® Pilot RCT INFUSE®/INTER FIX™ ALIF Pilot RCT MAVERICK™ Disc Pivotal RCT Pilot, PLIF Instrumented PEEK INFUSE®/TELAMON Single rhBMP rhBMP (rhBMP AMPLIFY™ rhBMP rhBMP INFUSE®/LT INFUSE®/LT INFUSE®/LT INFUSE® Bone Dowel Pilot RCT INFUSE® Bone Dowel Pivotal RCT INFUSE®/INTER FIX™ INFUSE®/CORNERSTONE® ACDF Pilot RCT INFUSE®/Mastergraft® Pilot RCT INFUSE®/INTER FIX™ ALIF Pilot RCT MAVERICK™ Disc Pivotal RCT Pilot, PLIF Instrumented PEEK INFUSE®/TELAMON Single rhBMP rhBMP AMPLIF rhBMP rhBMP
Arm and neck (upper extremity) pain 0 0 0 0 0 0 1 2 1 14 2 0 9 0 6 0 0 0 0 0 0 0 1 0 0 11 0 Arthritis or bursitis 0 0 0 0 0 0 0 0 0 0 0 0 0 24 0 0 0 0 0 0 0 0 0 0 0 0 16 Back and leg (lower extremity) pain 3 41 26 9 8 12 4 12 5 77 23 7 64 103 15 0 0 33 1 12 8 1 7 5 0 54 86 Cancer 0 2 0 1 1 0 0 1 0 3 0 0 1 10 0 0 0 1 0 0 0 0 0 0 0 1 2 Cardiovascular 0 9 7 1 3 6 2 2 2 8 2 1 16 51 1 0 0 12 0 2 10 1 3 1 0 15 55 Death 0 0 0 0 0 1 0 1 0 1 0 0 1 3 0 0 0 1 1 0 1 0 0 1 0 2 4 Dural injury 0 0 0 0 0 3 0 2 0 0 0 2 5 1 1 0 0 1 0 0 2 0 2 0 0 7 0 Dysphagia 0 0 0 0 0 0 2 0 0 0 0 1 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 Gastrointestinal 2 30 23 1 8 9 4 3 2 18 5 3 15 40 6 0 1 26 3 8 11 0 7 2 1 23 33 Graft related ------0 8 1 1 2 0 4 1 0 11 17 Heterotopic bone formation ------24 ------4 ------Implant 1 8 3 1 2 0 0 2 1 15 0 0 2 5 0 0 1 2 0 0 0 0 0 1 0 2 4 Infection 0 19 17 0 5 7 4 4 1 12 4 2 23 39 1 0 0 16 1 3 5 0 4 0 0 16 45 Neurological 0 20 19 2 15 14 7 4 2 60 6 5 15 68 4 0 0 23 1 4 13 2 3 1 0 14 58 Osteolysis ------12 ------
242 Appendix XV – Adverse events reported in Medtronic IPD
Other pain 0 19 12 6 12 9 2 3 0 12 1 3 14 29 0 0 0 13 1 4 9 1 0 0 1 6 28 Respiratory 1 3 1 0 2 0 1 2 0 3 2 1 8 16 1 0 0 4 0 4 2 0 2 0 0 3 12 Retrograde ejaculation 0 5 6 0 0 0 0 0 0 2 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 Spinal 0 21 9 4 3 5 3 4 1 26 5 2 3 24 10 0 0 17 3 1 5 1 3 2 0 4 27 Trauma 3 34 31 4 10 7 2 1 6 56 3 2 16 68 2 0 0 34 5 9 5 1 0 2 0 23 58 Urogenital 0 20 21 2 7 1 1 2 0 14 1 0 21 27 4 0 1 12 0 2 4 1 2 0 1 18 22 Vascular 0 0 0 2 1 0 0 0 0 2 1 0 3 0 1 0 0 0 3 0 0 0 0 0 0 1 0 Vertebral fracture 0 1 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Wound complication 0 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 0 4 0
243 Appendix XVI – Appendices Safety Results Chapter
Appendix XVI
Appendices Safety Results Chapter
Table 49 Adverse events in the Glassman trial Category ICBG rhBMP-2 Back or leg pain 3 0 Cardiac 7 1 Gastrointestinal 3 2 Incision related 2 1 Line sepsis 0 1 Neurological 1 0 Screw related 2 0 Shingles 0 1 Urinary tract infection 1 1 Wound infection 4 1 Total patients 53 53
244 Appendix XVI – Appendices Safety Results Chapter
Figure 60 Adverse events for INFUSE trials.
245 Appendix XVI – Appendices Safety Results Chapter
Figure 61 Severity of adverse events in INFUSE trials.
Table 50 Reasons for further surgeries in Medtronic trials. Reason for surgery ICBG rhBMP-2 Elective removal 8 11 Non-elective removal 30 16 Removal (unspecified) 4 7 Reoperation 25 28 Revision 11 14 Supplemental fixation 44 31 Other reason 302 407 Total surgeries 424 514 Patients with further surgery 231 268 Total patients 608 694
246 Appendix XVI – Appendices Safety Results Chapter
Figure 62 Forest plot for relative risk of cancer including additional cases from INFUSE®/LT-CAGE® open pivotal RCT.
247 Appendix XVI – Appendices Safety Results Chapter Table 51 Data extracted from ALF/ALIF wider literature papers.
Adverse events not reported in any paper: osteolysis, cancer,
Other, unlikely to be BMP General Adverse Events ALF Related Events related
formation
Study ID Study Intervention Second Surgery Heterotopic bone Neurologic events failureHardware Wound complication (incl infections) wound Infection Urinary retention Retrograde Ejaculation Vascular Events 5/239 22/239 15/239 rhBMP-2 Carragee (2.09%) (9.7%) (6.3%) (2012) 129 4/233 10/233 2/233 ICBG (1.71%) (4.6%) (0.9%) 7/45 (15.5%) 0/45 1/45 0/45 bowel perforations: 0/45 (0%), rhBMP-2 removal of 0/45 (0%) 0/45 (0%) (0%) (2.2%) (0%) urologic injuries: 0/45 (0%) Gerszten screws (2011) 130 2/54 (3.7%) 0/54 1/54 0/54 bowel perforations: 0/54 (0%), BMA removal of 0/54 (0%) 0/54 (0%) (0%) (1.9%) (0%) urologic injuries: 0/54 (0%) screws
McConnell rhBMP-2 0 (0%) (2011) 131 BMA 0 (0%) 39 procedures rhBMP-2 60/172 15/172 0/172 12/172 in 26 Maverick patients Medtronic trial 101 procedures Maverick 133/405 14/405 0/405 23/405 in 77 patients
Pradhan rhBMP-2 3/9 (33%) (2006) 132 ICBG 7/27 (25.9%) deep wound Slosar (2007) intra-operative dural rent: 1/45 rhBMP-2 0/45 (0%) 0/45 infection: 133 (2.2%) 1/45 (2.2%) Control 4/30 (13%) 0/30 superficial
248 Appendix XVI – Appendices Safety Results Chapter
Other, unlikely to be BMP General Adverse Events ALF Related Events related
formation
Study ID Study Intervention Second Surgery Heterotopic bone Neurologic events failureHardware Wound complication (incl infections) wound Infection Urinary retention Retrograde Ejaculation Vascular Events wound infection: 1/30 (3.3%)
249 Appendix XVI – Appendices Safety Results Chapter Table 52 Data extracted from PLF/TLIF aggregate additional papers.
Adverse events not reported in any paper: leg weakness
General Adverse Events PLF/TLIF Related Events
Other, unlikely to be BMP related
surgery
Study ID Study Intervention Second Heterotopic bone formation Osteolysis Infection Neurologic events Cancer failureHardware Wound complication (incl infection) Legpain/radiculitis Inflammatory cyst formation rhBMP-2 0% 0% 3 px total 0% dural tear Anand (2006) 116 ICBG 0% 0% Not reported 0% dural tear
Glassman rhBMP-2 7/91 (2007) 117 ICBG nr symptoms suggestive of radiculitis: 43/82 20/82 55/82 rhBMP-2 6wks: 5/82 (53%) (24%) (67%) (6.1%), 6mths: 2/82 Gray (2010) (2.4%) 118 symptoms suggestive of radiculitis: Autologous 10/39 3/39 32/39 6wks: 2/39 bone (25%) (8%) (82%) (5.1%); 6mths: 0/39 (0%) 5/23 rhBMP-2 Joseph (2007) (27.7%) 119 Autologous 1/10
bone (10%) urinary deep vein thromboses: 3/34 (9%) tract: dural tear: 1/34 (3%) rhBMP-2 1/34 (2.9%) Lee (2010) 120 1/34 cardiac problems: 2/34 (6%) GI (2.9%) problems: 2/34 (6%) ICBG urinary 1/41 (2.4%) deep vein thromboses : 5/41(12.2%)
250 Appendix XVI – Appendices Safety Results Chapter
General Adverse Events PLF/TLIF Related Events
Other, unlikely to be BMP related
surgery
Study ID Study Intervention Second Heterotopic bone formation Osteolysis Infection Neurologic events Cancer failureHardware Wound complication (incl infection) Legpain/radiculitis Inflammatory cyst formation tract: dural tear: 3/41 (7.3%) 2/41 cardiac problems: 4/41 (10%) GI (4.9%) problems: 4/41 (10%) 2/17 Mannion rhBMP-2 2 cases of perineural cyst formation 121 (11.8%) (2010) ICBG 0/19 (0%) rhBMP-2 4/35 (11.4%) Mindea (2009) 122 Allo + 0/8 (0%) autograft rhBMP-2 (+ 1/25 partial CSF leak: 2/25 (8%) (4%) Mummaneni ICBG) (2004) 123 1/19 donor-site pain: 58%, CSF leak: 2/19 ICBG (5.2 (10.4%) %) 1/15 rhBMP Pimenta (6.7%) (2011) 124 SCP 0/15 (0%) urinary 2/86 5/86 tract: 2/86 lumbar: lumbar seroma: dural tear: 4/86 (4.8%), Ileus: 1/86 rhBMP-2 9.3% 12/86 (14%) (2.3%) (5.8%) 2/86 (2.3%) 3/86 (3.5%) 1/86 (1.2%) (1.2%), Retained drain: 1/86 (1.2%) 125 (2.3%) Rihn (2009) urinary lumbar: persistent donor site pain: 10/33 12.1 0/33 tract: 0/33 2/33 (6.1%) lumbar seroma: ICBG 0/33 (0%) 1/33 (3%) (30.3%), dural tear: 0/33 (0%), Ileus: % (0%) 1/33 (0%) donor-site: 0/33 (0%) 1/33 (3%), retained drain: 0/33 (0%) (3%) 1/33 (3%) 6/64 11/64 rhBMP-2 (9.4 Rowan (2011) (16.9%) 126 %) Not BMP 3/40 (7.5%)
251 Appendix XVI – Appendices Safety Results Chapter
General Adverse Events PLF/TLIF Related Events
Other, unlikely to be BMP related
surgery
Study ID Study Intervention Second Heterotopic bone formation Osteolysis Infection Neurologic events Cancer failureHardware Wound complication (incl infection) Legpain/radiculitis Inflammatory cyst formation 1/41 rhBMP + (2.4 0/41 (0%) dural tear: 2/41 (4.9%) ICBG Singh (2006) %) 127 0/11 0/11 (0%) ICBG (0%) (0%) 168/6049 rhBMP-2 15/6049 (0.2%) Williams (2.8%) (2011) 115 496/18267 Control 43/18267 (0.2%) (2.7%) 2/24 2/24 rhBMP (8.3 dural tear: 1/24 (4.2%) (8.3%) Taghavi %%) (2010) 128 6/38 BMA + 3/38 dural tear: 1/38 (2.6%), persistent (15.8 allograft (7.9%) donor site pain: 3/28 (7.9%) %%)
252 Appendix XVI – Appendices Safety Results Chapter Table 53 Data extracted from cervical wider literature papers.
Adverse events not reported in any paper: osteolysis, cancer, airway obstruction
Other, potentially BMP related General Adverse Events Cervical Related Events Other, unlikely to be BMP related classified by the paper
Study ID Study Intervention Second surgery Heterotopic bone formation Infection Neurologic events failureHardware Wound complications (including infections) wound Dysphagia NeckPain Recurrent laryngeal neck palsy rhBMP-2 Butterman (2008) 137 donor site: ICBG 1/30 (3.3%)
rhBMP-2 6/41 (14.6%) medical complications: 0/41 (0%) Crawford 138 1/36 (2.8%) (2009) donor site ICBG deep medical complications: 3/36 (8.3%) infection: 1/36 (2.8%) respiratory distress: 2/16 C5 palsy: pancreatitis: 1/16 (6.3%), hyponatremia: wound (12.6%), neck swelling: 1/16 1/16 2/16 1/16 1/16 (6.3%), atrial fibrillation: 1/16 rhBMP-2 infections: 0 / 16 (0%) (6.3%) (6.3%) (12.5%) (6.3%) (6.3%), GI bleed: 1/16 (6.3%), persistent 0/16 (0%) Hiremath severe pain: 1/16 (6.3%) 139 (2009) congestive heart failure: 1/67 (1.5%), wound respiratory distress: 1/67 3/67 1/67 persistent severe pain: 5/67 (7.5%), Control infections: 0 / 67 (0%) (1.5%) (4.5%) (1.5%) other: ?/67 8/67 (11.9%) deep venous thrombosis: 1/67 (1.5%) esophage al dilation 0/84 feeding tubes placed: 2/84 (2.4%) rhBMP-2 required: (0%) post-surgical ER visits: 5/84 (5.9%) 1/84 Longley (1.2%) (2009) 140 esophage al dilation feeding tubes placed: 1/84 (1.2%) 1/84 Control required: post-surgical ER visits: 3/84 (3.6%) (1.2%) 1/84 (hospitalisation required in 2 controls) (1.2%)
253 Appendix XVI – Appendices Safety Results Chapter
Other, potentially BMP related General Adverse Events Cervical Related Events Other, unlikely to be BMP related classified by the paper
Study ID Study Intervention Second surgery Heterotopic bone formation Infection Neurologic events failureHardware Wound complications (including infections) wound Dysphagia NeckPain Recurrent laryngeal neck palsy Sig. more rhBMP-2 severe Lu (2010) 141 dysphagia Control in rhBMP- 2 patients 5/69 19/69 (27.5%) perioperative rhBMP-2 Smucker (7.2%) swelling 142 (2006) 2/165 6/165 (3.6%) perioperative Control (1.2%) swelling 2 wks: 17/20 2/22 (85%) rhBMP-2 hoarseness: 60% (9%) 6 wks: Vaidya 13/20 (2007a) 143 (65%) 2 wks: 1/24 7/18 (38%) Control hoarseness: 62% (4%) 6 wks: 4/18 (22%) 4451/11 14/4532 10/4532 Epidural hematoma/seroma: rhBMP-2 933 (2.1%) (0.6%) 3/4532 (0.5%) Williams (65%) (2011) 115 2400/43 17/352 4/352 Epidural hematoma/seroma: Control 929 (0.4%) (0.2%) 14/352 (0.3%) (35%) Hematoma: 1/48 (2.2%), C5 palsy: CSF leak: 0/48 (0%), pneumonia: 1/48 Xu (2011) 7/48 0/48 5/48 0/48 wound 3/48 19/48 388 rhBMP-2 3/48 (2.2%), deep vein thrombosis: 0/48 (15.2%) (0%) (10.9%) (0%) dehiscence: (6.3%) (47.5%) (6.5%) (0%), pulmonary embolism: 0/48 (0%) 1/48 (2.2%)
254 Appendix XVI – Appendices Safety Results Chapter
Other, potentially BMP related General Adverse Events Cervical Related Events Other, unlikely to be BMP related classified by the paper
Study ID Study Intervention Second surgery Heterotopic bone formation Infection Neurologic events failureHardware Wound complications (including infections) wound Dysphagia NeckPain Recurrent laryngeal neck palsy hematoma: 3/156 CSF leak: 2/156 (1.3%), pneumonia: (1.9%), C5 palsy: 32/156 0/156 17/156 11/156 6/156 31/156 4/156 (2.6%), deep vein thrombosis: Control wound 7/156 (20.5%) (0%) (11%) (7.1%) (3.8%) (23.3%) 3/156 (1.9%), pulmonary embolism: dehiscence: (4.5%) 2/156 (1.3%) 8/156 (5.1%)
tracheotomies: 8/260 (3.1%), unplanned intubations: 16/260 (6.2%), death: 11/260 (4.2%), readmissions: 18/260 rhBMP-2 dyspnea: 53/260 (20.4%), 23/260 (8.8%), percutaneous (6.9%) hoarseness: 6/260 (2.3%), endoscopic gastrostomy: 6/260 (2.3%), respiratory failure: 34/260 (13.1%) Yaremchuk 145 (2010) tracheotomies: 3/515 (0.6%), unplanned intubations: 8/515 death: 9/515 (1.7%), readmissions: (1.6%), 17/515 26/515 (5%), percutaneous endoscopic Control dyspnea: 41/515 (8%), (3.3%) gastrostomy: 4/515 (0.8%), hoarseness: 6/515 (1.2%),
respiratory failure: 24/515 (4.7%)
255 Appendix XVI – Appendices Safety Results Chapter Table 54 Data extracted from various surgical type papers.
Adverse events not reported in any paper: osteolysis, infection, neurological events Surgery specific General Adverse Events Other, potentially BMP related Other, unlikely to be BMP related events reported
Study ID Study Intervention Second surgery Heterotopic bone formation Cancer failureHardware Wound complocations (including infections) wound Dysphagia Cahill rhBMP-2 716/17,495 152/17,495 other complications: 872/17,495 (2009) 7 (4.1%) (0.9%) (5%) Control 873/52,505 730/52,505 other complications: 2,103/52,505 (1.7%) (1.4%) (4%) Cahill rhBMP-2 repeat (2011) 146 fusion: 55/2,373 (2.3%) ICBG repeat fusion: 80/2,373 (3.4%) Deyo rhBMP-2 183/1,703 re-hospitalisation within 1mth: (2012) 147 (10.8%) 205/1,703 (12%), cardiac/pulmonary/stroke complications: 87/1,703 (5.1%), death within 1mth: 15/1,703 (0.9%) Control 1588/1,5119 re-hospitalisation within 1mth: (10.5%) 1,855/15,119 (12.3%), cardiac/pulmonary/stroke complications: 868/15,119 (5.8%), death within 1mth: 118/15,119 (0.8%) Latzman rhBMP-2 plus 4/24 3/20 BMP patients renal (2010) 148 allo/auto (17%) insufficiency, all 3 later Allo/autograft 8/101 diagnosed with malignancies only (8%) No sig increase in endocrinological, autoimmune, neurologic or neoplastic disorders
256 Appendix XVI – Appendices Safety Results Chapter
Surgery specific General Adverse Events Other, potentially BMP related Other, unlikely to be BMP related events reported
Study ID Study Intervention Second surgery Heterotopic bone formation Cancer failureHardware Wound complocations (including infections) wound Dysphagia Maeda rhBMP-2 1/23 (4.3%) acute tubular necrosis: 1/23 (4.3%) (2009) 149 ICBG 6/32(18.8%) Mines rhBMP-2 8/15460 (2011) 150 (0.05%) Not BMP 83/78194 (0.11%) Vaidya rhBMP-2 4/36 (11.1%) subsidence: cervical fusions: (2007b) 151 24/55 levels 6/11 (55%) (43.6%) between 6wks and 3mths, after 12mths 27/55 levels (49.1%) Control 5 or 6/41 Subsidence: (12.2- after 12mths 14.6%) 4/63 levels (9.3%) Williams rhBMP-2 4451/11933 271/11281 Hematoma/seroma: 23/11281 (2011) 115 (65%) (2.4%) (0.2%) spinal level not reported Control 2400/43929 946/39397 Hematoma/seroma: 79/39397 (35%) (2.4%) (0.2%) spinal level not reported
257 Appendix XVI – Appendices Safety Results Chapter Table 55 Quality assessment for PLF/TLIF papers from the wider literature.
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) 100 Neurologic consecutive events, Mean 30 patients, last Not matched Anand hardware Same months 85 received 116 Representative Secure record No or controlled No details (2006) failure, leg source (minimum 2 rhBMP-2. No for pain/radiculitis years) data on , other comparabilit y of groups. 91 rhBMP-2 patients. Subset of 35 patients from Mean 27 Amplify RCT Not matched Reference Glassman Hardware months used as 117 Representative Unclear Secure record Yes or controlled to secure (2007) failure (minimum 2 controls. for record years) Limited data on comparabilit y and confounding. 121 consecutive Heterotopic patients, 82 bone received formation, Not matched Reference Gray Same rhBMP-2. 118 osteolysis, Representative Secure record Unclear or controlled to secure 2yrs (2010) source Reported no hardware for record significant failure, leg differences pain/radiculitis in mean age, BMI, ASA,
258 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) primary diagnosis or baseline outcomes. 33 Mean 25 consecutive months patients, 23 Heterotopic Not matched Reference (clinical) and chose to Joseph Same 119 bone Representative Secure record Yes or controlled to secure 7.9 months receive (2007) source formation for record (radiological) rhBMP-2. No . Minimum 6 data on months. comparabilit y of groups. Comparison limited to patients aged >65 BMP: 38.3- years (34 39.2mths rhBMP-2
Infection, Sub-group Not matched Reference patients, 41 Lee Same Control: 120 wound (patients aged Secure record No or controlled to secure autograft). (2010) source mean complications >65yrs) for record Higher 34.7mths proportion of (minimum 2 rhBMP-2 years) patients undergoing revision surgery.
259 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) 36 consecutive patients, 17 received Heterotopic Reference rhBMP-2. Mannion bone Same 121 Representative Secure record Unclear Unclear to secure 12 months Similar no. (2010) formation, source record of spinal other levels; no other comparabilit y data. 43 consecutive Not matched Reference Length of patients, 35 Mindea Leg Same 122 Representativ Secure record Yes or controlled to secure follow-up not received (2009) pain/radiculitis source for record reported rhBMP-2. No comparabilit y data. 40 patients followed-up (4 missing patients Not matched Reference Mean 9 suggests Mummanen Neurologic Same 123 Representative Secure record No or controlled to secure months non- i (2004) events source for record (range 3-18) consecutive) , 21 received rhBMP-2. No comparabilit y data.
260 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) Random sequence generation Unclear 15/30 Allocation concealment Unclear patients Unclear Blinding of Low (three blinded radiologists) received participants/personnel Low (100% follow-up both groups) rhBMP-2 in Blinding of outcome Unclear this RCT. heterotopic Pimenta assessment Only an 124 bone (2011) Incomplete outcome data abstract was formation Selective reporting available making it difficult to assess risk of bias 119 patients Second (11 missing surgery, patients heterotopic suggests bone Reference non- formation, to secure Radiographic consecutive) osteolysis, record follow-up , 86 received infection, mean rhBMP-2. Not matched Rihn wound Same Record 19.1mths, Clinical 125 Representative Secure record No or controlled (2009) complication, source linkage clinical follow-up for hardware follow-up was longer failure, leg Self report mean for rhBMP-2 pain/radiculitis (donor site 27.6mths. than , inflammatory pain) autograft cyst patients formation, (38.5 vs. other 24.4 months). No
261 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) significant differences in age, sex, diagnosis, radiographic follow-up, previous surgery, or surgical level. Second Insufficient Rowan No 126 surgery, leg No details Secure record Unclear Unclear No details Not reported details to (2011) details pain/radiculitis assess. 52 patients, 39 received rhBMP-2. Matched on Second sex, though surgery, Reference Singh Same Matched on rhBMP-2 127 heterotopic Representative Secure record No to secure 24 months (2006) source sex patients bone record were formation substantially older (65.3 vs. 54.2 years).
262 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) 62 consecutive patients undergoing revision surgery, 24 received Second rhBMP-2. No Not matched Reference Taghavi surgery, Same Minimum 24 significant 128 Representative Secure record No or controlled to secure (2010) hardware source months differences for record failure between groups on age, sex, smoking, diabetes, osteoporosis of no. of levels. 6049/24316 patients All cases received nd used, not 2 surgery, rhBMP-2 in matched or complications, a controlled Record Williams wound Same retrospective 115 Representative Secure record No for. linkage Not reported (2011) infections, source cohort using (database) inflammatory medical Sig diff on cysts records. No age, matching or diagnosis adjustment for
263 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcome y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) confounders, or details on follow-up period or duration.
264 Appendix XVI – Appendices Safety Results Chapter Table 56 Quality assessment for ALF/ALIF papers from the wider literature.
Selection Comparability Outcomes Demonstrated Confounders that outcome and other Selection Follow-up Study ID AEs assessed Representativeness Ascertainment of interest factors Outcome Comment of control (duration, of exposed cohort? of exposure was not (matched OR assessment group proportion) present at controlled for start of study in analysis) Historical Reference to controls from secure Yes same centre, record Carragee Infection, (retrograde Minimum similar on 129 (retrograde (2012) urinary ejaculation) Not matched of 1 year important Same ejaculation, retention, Representative Secure record or controlled follow-up confounders source urinary Linked retrograde No (infection, for for each for retrograde retention) papers ejaculation urinary patient. ejaculation.
retention) 239 patients No details with rhBMP-2, (infection) 233 without. Heterotopic bone 45/99 formation, participants neurologic drawn from 2 events, different Matched on hardware institutions in “demographic Reference to Gerszten failure, Different different cities 130 Representative Secure record No data, secure 2 years (2011) vascular source received including sex record events, rhBMP-2. No and age” wound data on complications, comparability second other than surgery, age/sex. other. Random sequence generation 34/66 patients Unclear Allocation concealment received McConnell Unclear 131 Second Blinding of High rhBMP-2 in (2011) surgery participants/personnel High this RCT. Only
Blinding of outcome Low (96% follow-up, balanced) an abstract Unclear assessment was available
265 Appendix XVI – Appendices Safety Results Chapter
Selection Comparability Outcomes Demonstrated Confounders that outcome and other Selection Follow-up Study ID AEs assessed Representativeness Ascertainment of interest factors Outcome Comment of control (duration, of exposed cohort? of exposure was not (matched OR assessment group proportion) present at controlled for start of study in analysis) Incomplete outcome data making it Selective reporting difficult to assess risk of bias.
172/577 patients Random sequence generation received Maverick Allocation concealment rhBMP-2 in Low Medtronic Blinding of Low this RCT. trial Second participants/personnel Low Blinding of surgery Blinding of outcome High assessment IPD assessment Low (96% follow-up, balanced) not possible Unclear provided Incomplete outcome data due to the disc Selective reporting replacement system, IPD provided. rhBMP-2: 9/36 patients mean received Not matched Reference to 26mths rhBMP-2. Pradhan Second Same 132 Representative Secure record Yes or controlled secure Comparability (2006) surgery source for record Control: data provided mean for age, sex, 36mths level of fusion. Compared 45 patients Wound Not matched Reference to Minimum Slosar Same receiving 133 complications, Representative Secure record Yes or controlled secure of 12 (2007) source rhBMP-2 with other for record months 30 historical controls from
266 Appendix XVI – Appendices Safety Results Chapter
Selection Comparability Outcomes Demonstrated Confounders that outcome and other Selection Follow-up Study ID AEs assessed Representativeness Ascertainment of interest factors Outcome Comment of control (duration, of exposed cohort? of exposure was not (matched OR assessment group proportion) present at controlled for start of study in analysis) same centre. Similar in terms of age, sex, worker’s compensation, smoking, previous surgery.
267 Appendix XVI – Appendices Safety Results Chapter Table 57 Quality assessment for cervical fusion papers from the wider literature.
Comparabilit Selection Outcomes y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) Not matched or controlled 30/66 for. consecutively Comparabilit recruited Wound y data were patients Butterman complications Same 24 to 36 137 Representative Secure record No provided for Self report received (2008) , dysphagia, source months age, sex, rhBMP-2. other smoking Groups were status, and comparable pre-operative at baseline. diagnosis. Not matched or controlled for. Comparabilit 41/107 Wound y data were consecutively Reference Crawford complications Same provided for recruited 138 Representative Secure record No to a secure 3 months (2009) , infection, source sex, smoking patients record other status, and received age. Groups rhBMP-2.. were comparable at baseline. Not matched Infection, 16/83 or controlled medical consecutively for. Some Reference Hiremath complications Same recruited 139 Representative Secure record No comparability to a secure Not reported (2009) , swelling, source patients data record hardware received provided for failure, other rhBMP-2. age, sex, and
268 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) pre-operative diagnosis.
84/168 patients received rhBMP-2. It is unclear if patients were rhBMP-2: recruited mean 4.8 consecutively Described as nd Reference . Groups Longley 2 surgery, Same “matched” 140 Representative Secure record No to a secure Control: were (2009) other. source but details record mean 5.8 comparable not reported Units not for age at reported baseline. Details of matching not reported and unclear follow-up data.
269 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) 100/150 patients received rhBMP-2. It is rhBMP-2: unclear if mean 35 patients were Described as Reference months recruited Lu (2010) Dysphagia, Same “matched” 141 Representative Secure record No to a secure consecutively other. source but details record Control: . not reported mean 25 Comparability months data and details of the matching process were not reported. Analyses 69/234 controlled for consecutive significant patients predictors. received There were rhBMP-2. significantly The analyses Cervical more Smucker Same Secure controlled for 142 swelling, Representative Secure record No smokers and 6 weeks (2006) source record potential dysphagia patients who confounders had had and the previous follow-up was fusion sufficient for surgery in short term the rhBMP-2 AEs. group, but
270 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) this did not affect the results.
22/46 consecutive patients Dysphagia, Not matched Vaidya nd Same Up to 24 received 143 2 surgery, Representative Secure record No or controlled Self report (2007a) source months rhBMP-2. other for Comparability data were not presented. 652/4532 patients received rhBMP-2 in a All cases retrospective nd used, not 2 surgery, cohort using matched or complications medical controlled Record Williams , wound Same records. No 115 Representative Secure record No for. linkage Not reported (2011) infections, source matching or (database) inflammatory adjustment Sig diff on cysts for age, confounders, diagnosis or details on follow-up period or duration
271 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders Selectio d that and other Study ID AEs reported Representativenes Outcome Follow-up Comment n of Ascertainmen outcome of factors s of exposed assessmen (duration, control t of exposure interest was (matched OR cohort? t proportion) group not present at controlled for start of study in analysis) 48/204 patients received rhBMP-2. It is unclear if consecutive Infection, patients were dysphagia, recruited. wound Not matched Reference Xu (2011) Same Mean 24.2 Comparability 144 complications Representative Secure record No or controlled to a secure nd source months data provided , 2 surgery, for record for age, sex, hardware comorbidities, failure, other. and presenting symptoms. Significantly more men in control group. 260/775 patients received rhBMP-2. It is 2:1 matching 1mth for unclear if Yaremchu Dysphagia, Same based on complications patients were k (2010) Representative Secure record No Unclear 145 other source procedure , 3mths for recruited
type death consecutively . Comparability data were not provided.
272 Appendix XVI – Appendices Safety Results Chapter
273 Appendix XVI – Appendices Safety Results Chapter Table 58 Quality assessment for mixed approach fusion papers from the wider literature.
Comparabilit Selection Outcomes y Demonstrate Confounders d that Selectio and other Follow-up Study ID AEs assessed Representativene Ascertainme outcome of Comment n of factors Outcome (duration, ss of exposed nt of interest was control (matched OR assessment proportion cohort? exposure not present group controlled for ) at start of in analysis) study Retrospectiv e cohort of 17,623 /70, 649 participants, Not matched received or controlled Data rhBMP-2. Dysphagia, for, analysis collected Used wound suggested 2002 – Cahill Same Secure Record linkage medical 7 complications, Representative No associations 2006 but (2009) source record (database) records but other with no follow- analysis complications sex/race/pay up data was not er and BMP reported adjusted for use confounders and duration of follow-up was unclear. Retrospectiv e cohort of Controls 2372/4744 were patients matched Cahill Minimum who Same Secure using Record linkage (2011) Repeat fusion Representative No 12 received 146 source record propensity (database) months rhBMP-2 scores using (details medical provided) records with matched
274 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders d that Selectio and other Follow-up Study ID AEs assessed Representativene Ascertainme outcome of Comment n of factors Outcome (duration, ss of exposed nt of interest was control (matched OR assessment proportion cohort? exposure not present group controlled for ) at start of in analysis) study controls and limited follow-up duration.
Analyses were Retrospectiv 4 years, stratified by e cohort of nd most 2 surgery, surgical 1703/16822 appeared hospitalisation, complexity receiving Deyo to be cardiac/stroke, Same Secure and previous Record linkage rhBMP-2 (2012) Representative No followed 147 wound source record lumbar (database) using up but complications, surgery, medical exact death other records and number variables stratified unclear also analyses. mentioned
275 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders d that Selectio and other Follow-up Study ID AEs assessed Representativene Ascertainme outcome of Comment n of factors Outcome (duration, ss of exposed nt of interest was control (matched OR assessment proportion cohort? exposure not present group controlled for ) at start of in analysis) study 20/121 patients BMP: received mean rhBMP-2 in Not matched 1.48yrs one or controlled institution, for. Notes sig Control: Latzman unclear if Cancer, renal Same Secure diff in Independent/blinded/ mean (2010) Representative No consecutive 148 insufficiency source record operation ref to secure record 4.49yrs patients, details unadjusted (duration and *sig diff analyses cage use) Proportion and sig s not differences reported between groups. BMP: 23/55 mean patients Not matched 2.7yrs received or controlled rhBMP-2 in for. Control: one nd 2 surgery, mean institution, Maeda NS diff in pseudoarthrosi Differen Secure Independent/blinded/ 4.9yrs unclear if (2009) Sub-group No smoking, co 149 s, t source record ref to secure record consecutive morbidity, complications *sig diff patients. BMI, number The vertebrae, All analyses previous patients were surgery had min unadjusted 2yr follow- (but no sig
276 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders d that Selectio and other Follow-up Study ID AEs assessed Representativene Ascertainme outcome of Comment n of factors Outcome (duration, ss of exposed nt of interest was control (matched OR assessment proportion cohort? exposure not present group controlled for ) at start of in analysis) study up differences between groups).
15460/9365 4 patients received rhBMP-2 in a retrospectiv e cohort Controlled Approx 12 study using Mines for months, Same Secure Record linkage medical (2011) Cancer Representative Yes pancreatic proportion 150 source record (database) records. cancer s not Analysis factors reported was adjusted for some confounders but duration of follow-up was short.
nd Independent/blinded/ BMP: 36/77 2 surgery, Vaidya Not matched ref to secure record mean consecutive 1 subsidence, Same Secure (2007b) Representative No or controlled 24.1mths patients 51 dysphagia in source record for. And self report (no Control: received cervical record) mean rhBMP-2
277 Appendix XVI – Appendices Safety Results Chapter
Comparabilit Selection Outcomes y Demonstrate Confounders d that Selectio and other Follow-up Study ID AEs assessed Representativene Ascertainme outcome of Comment n of factors Outcome (duration, ss of exposed nt of interest was control (matched OR assessment proportion cohort? exposure not present group controlled for ) at start of in analysis) study 24mths but with no adjustment Appears in analyses. to be 100%? 11933/5586 2 patients received rhBMP-2 in a All cases retrospectiv nd used, not 2 surgery, e cohort matched or complications, using Williams controlled wound Same Secure Record linkage Not medical (2011) Representative No for. 115 infections, source record (database) reported records. No
inflammatory matching or Sig diff on cysts adjustment age, for diagnosis confounders , or details on follow-up period or duration.
278 Appendix XVII – Adverse events reported in MedWatch forms
Appendix XVII
Table 59 Number (%) of Medtronic/PCR adverse events reports
Type of surgery* Cervical (ACDF, ALIF Total reported PLIF/TLIF/PLF PCF) % of % of % of % of reporte reporte reporte reporte n/913 d n/171 d n/45 d n/357 d AE description (pre-specified) reports events reports events reports events reports events for this for this for this for this surgery surgery surgery surgery "Unanticipated inflammation and 66 7.2 26 15.2 2 4.4 5 1.4 swelling" Airway obstruction 23 2.5 16 9.4 0 0 0 0 Cancer 1 0.1 0 0 0 0 0 0 Dysphagia 52 5.7 40 23.4 0 0 0 0 Hardware failure 26 2.8 1 0.6 1 2.2 5 1.4 Heterotopic bone formation 95 10.4 7 4.1 3 6.7 57 16 Inflammatory cyst formation 149 16.3 18 10.5 6 13.3 97 27.2 Leg pain/radiculitis 42 4.6 1 0.6 0 0 38 10.6 Leg weakness 3 0.3 0 0 0 0 2 0.6 Neck pain 4 0.4 2 1.2 0 0 0 0 Neurologic events 34 3.7 7 4.1 1 2.2 11 3.1 Non-union 54 5.9 1 0.6 16 35.6 17 4.8 NR 13 1.4 2 1.2 0 0 1 0.3 Osteolysis 72 7.9 5 2.9 0 0 38 10.6 Post-operative pain 16 1.8 0 0 1 2.2 4 1.1 Recurrent laryngeal nerve palsy 3 0.3 3 1.8 0 0 0 0 Retrograde ejaculation 0 0 0 0 0 0 0 0 Urinary retention 1 0.1 0 0 0 0 0 0 Vascular events 1 0.1 0 0 0 0 1 0.3 Wound complications 82 9 9 5.3 3 6.7 16 4.5 Total pre-specified 737 80.7 138 80.7 33 73.3 292 81.8
Not pre specified or more than
one pre-specified event Allergic reaction 1 0.1 0 0 1 2.2 0 0 Allograft resorption 2 0.2 0 0 0 0 0 0 Cardiopulmonary arrest 3 0.3 3 1.8 0 0 0 0 Cauda equina syndrome 1 0.1 0 0 0 0 1 0.3 Dysphagia, possible airway 1 0.1 1 0.6 0 0 0 0 obstruction Either PE or MI 1 0.1 0 0 0 0 1 0.3 Epidural ossification 1 0.1 0 0 0 0 1 0.3 Granuloma 1 0.1 0 0 0 0 1 0.3 Haematoma 20 2.2 10 5.8 2 4.4 3 0.8 Hardware failure, Inflammatory cyst 1 0.1 0 0 0 0 1 0.3 formation
279 Appendix XVII – Adverse events reported in MedWatch forms
Hardware failure, pseudarthrosis 1 0.1 0 0 0 0 0 0 Infection, foot drop, heterotopic 1 0.1 0 0 0 0 1 0.3 bone growth Inflammatory cyst formation, 1 0.1 0 0 0 0 1 0.3 radiculitis Inflammatory cyst formation, leading 1 0.1 0 0 0 0 1 0.3 to radiculitis Inflammatory cyst formation, leg 1 0.1 0 0 0 0 1 0.3 pain Inflammatory cyst formation, leg 1 0.1 0 0 0 0 1 0.3 pain/radiculitis Inflammatory cyst formation, neck 1 0.1 1 0.6 0 0 0 0 pain Inflammatory cyst formation, non 1 0.1 0 0 0 0 1 0.3 radicular hip pain Inflammatory cyst formation, 5 0.5 0 0 0 0 5 1.4 radicular pain Inflammatory cyst formation, 2 0.2 0 0 0 0 2 0.6 radiculopathy Inflammatory cyst formation, 1 0.1 0 0 0 0 1 0.3 radiculopathy, back pain, foot drop Inflammatory cyst formation, symptoms of cauda equina 1 0.1 0 0 0 0 1 0.3 syndrome Inflammatory response 1 0.1 0 0 0 0 0 0 Leg pain, bladder control problems, 1 0.1 0 0 0 0 0 0 multiple cyst formation Leg pain, radiculitis, haematoma 1 0.1 0 0 0 0 1 0.3 Leg weakness, back pain, 1 0.1 0 0 0 0 1 0.3 heterotopic bone growth Leg weakness, haematoma 2 0.2 0 0 0 0 1 0.3 Neurologic events/other 1 0.1 0 0 0 0 1 0.3 Non-union, osteolysis 1 0.1 0 0 0 0 1 0.3 Osteolysis, heterotopic bone 1 0.1 0 0 0 0 1 0.3 formation Pain and numbness right iliac wing 1 0.1 0 0 0 0 1 0.3 Pseudarthrosis/pseudoarthrosis 29 3.2 10 5.8 4 8.9 5 1.4 Pseudarthrosis, hardware failure 2 0.2 0 0 2 4.4 0 0 Quadraplegia 1 0.1 0 0 0 0 0 0 Weakness in all 4 limbs 1 0.1 0 0 0 0 0 0 Other 84 9.2 8 4.7 3 6.7 30 8.4 TOTAL (not pre-specified) 176 19.3 33 19.3 12 26.7 65 18.2
GRAND TOTAL 913 100 171 100 45 100 357 100
280 Appendix XVIII – Information collected on Medtronic trial adverse event forms
Appendix XVIII
Information collected on Medtronic Trials adverse events forms
Adverse event forms for all* trials included the following categories (tick box) with example types of events as shown below: *except CRM Two-level (which collected only a narrative description) and Cornerstone Pilot (which did not include trauma or urogenital categories) • Dural leak or tear • Gastrointestinal: ileus, hernia, constipation • Infection: e.g. wound infection, drainage dehiscence, urinary tract infection, graft site infection etc. • Neurological: e.g. parasthesia, nerve damage, paralysis, loss of bowel, bladder control etc. • Other: describe • Trauma: e.g. fall accidents, fractures etc. (not Cornerstone) • Urogenital: – urinary retention, retrograde ejaculation etc. (not Cornerstone) • Vascular: vessel damage, hemorrhage, thombrophlebitis, blod clots etc. The following categories were additionally included for some trials (tick box) with example types of events as shown: • ATLANTIS related event: breakage, migration etc. (Cornerstone pilot) • Cornerstone bone related event: breakage, migration etc. (Cornerstone pilot) • CD Horizon related event: implant breakage, loosening etc. (Amplify, Mastergraft, BCP Canada) • Maverick related event: implant breakage, implant loosening etc. (Maverick) • LT-CAGE related event: implant breakage, implant loosening etc. (Maverick) • TSRH-related events: implant breakage, implant loosening etc. (BCP US, BCP Canada) • Maverick related event: implant breakage, implant loosening etc. (Maverick) • Systemic: cardiac arrest, reaction to anesthesia, respiratory arrest etc. (BCP US, Bone dowel pivotal, Amplify, Mastergraft , Maverick, Telemon, BCP Canada) • Implant breakage or disassembly (Bone dowel pivotal, Telemon) • Technical difficulties: specify (Bone dowel pivotal, Telemon) • Death: specify (Bone dowel pivotal) • Type of treatment: bed rest/ referral to specialists/ surgical (indicate type)/ physical therapy/ injections (specify)/ other (specifiy)/ no treatment. (Bone dowel pivotal, Amplify, Mastergraft , Maverick, Telemon, CRM) • Whether hospitilisation was required. (Amplify, Mastergraft pilot, CRM) • Outcome of event: resolved/ pending/ permanent disability or condition/death • (Cornerstone, Bone dowel pivotal, Mastergraft, Amplify, Maverick, Telamon, CRM) Adverse effect forms for all trials provided space for description of event, whether additional surgery required (Y/N) and space to describe outcome of event Adverse effect forms for all trials except CRM collected the investigators’ opinion regarding relationship to study: related/surgical procedure related/undetermined (unknown)/not related Adverse effect forms for all trials except CRM collected the investigators’ classification of event: asymptomatic/ mild/ moderate/ severe/ life threatening
281 Appendix XIX – Adverse events occurring in Medtronic trials as reported in published articles
Appendix XIX
Table 60 Adverse events occurring in Medtronic trials, as reported in published journal articles. Journal n n Study Year rhBMP-2 Control Notes article rhBMP-2 Control Postoperative ileus and delay in gait Postoperative ileus and delay in gait training training (1 patient) 9 (1 patient) LT Cage Pilot Boden 11 3 2000 Wound dehiscence (1 patient) Urinary retention (1 patient) Low back pain event before six-month Post-operative traumatic events (3 patients) follow-up (1 patient) Retrograde ejaculation was 10 Vascular events (6 events) Vascular events (5 events) reported, but not by treatment Burkus 143 136 2002 Retrograde ejaculation Retrograde ejaculation group (6 patients in total, 4 permanent) 35 No mention of adverse events in Burkus 143 136 2003 - - results 36 No mention of adverse events in LT Cage Burkus 23† 22† 2003 - - results Open Anatomical /technical difficulty (0 patients) Malpositioned implant (1 patient) Burkus42* 143 - 2009 Implant displacement and/or loosening (2 - Control arm not reported patients) Subsidence (6 patients) 43 Infection (36/266 patients) Burkus 277* 136 2011 - Control arm not reported Non-union (10/266 patients) 47 States "no adverse events Kleeman 22† n/a 2001 - n/a identified" 35 No mention of adverse events in Burkus 134 n/a 2003 - n/a LT Cage results Laparoscopi Anatomical /technical difficulty (9 patients) c Malpositioned implant (4 patients) Burkus42* 134 n/a 2009 Implant displacement and/or loosening (2 n/a patients) Subsidence (1 patient) Reports "no unanticipated Burkus53 24 22 2002 - - adverse events that were related Bone Dowel to the use of InFUSE Bone Graft" Pilot Reports "there were no adverse Burkus37 24 22 2004 - - events due to rhBMP-2" in discussion section only
282 Appendix XIX – Adverse events occurring in Medtronic trials as reported in published articles
Reports 14 patients with "transient localised areas of bone remodeling in the vertebral body Burkus54 79 52 2005 - - adjacent to the allograft dowel" that had "resolved by twenty-four months after surgery". Not reported as an adverse event. Bone Dowel Reports 14 patients with “transient Pivotal localised areas of bone remodelling in the vertebral body 55 adjacent to the allograft dowel” Burkus 79 52 2006 - - but “no unanticipated adverse events related to the use of rhBMP-2 and the collagen sponge carrier occurred” Vascular complication (1 patient, deep vein “New bone formation extending thrombosis) Dural tear (3 patients) outside the disc space and into Dural tear (2 patients) 63 Neurological complications (16 events in the spinal canal or neuroforamina” Interfix PLIF Haid 34 33 2004 Neurological complications (18 events in 14 14 patients) (24 rhBMP-2 patients vs 4 ICBG patients) patients; not explicitly noted as an Iliac crest harvesting complications (1 patient adverse event) pain, 1 patient haematoma) Incidental intraoperative durotomy (1 patient) Incidental intraoperative durotomy (1 patient) Wound infection (surgical site; 1 patient) Wound infection (surgical site; 1 patient) Mastergraft 72 Dawson 25 21 2009 Malpositioned pedicle screws (leading to Wound infection (graft donor site; 1 patient) Pilot 2nd surgery; 1 patient) Pseudarthrosis (leading to 2nd surgery; 2 “Failure of the index procedure” (leading to patients) hardware removal; 1 patient) “Anterior bone formation in “No unanticipated device-related adverse immediately adjacent segments” “No unanticipated device-related adverse events in either treatment group” Cornerstone 67 (2 rhBMP-2 patients vs 1 ICBG Baskin 18 15 2003 events in either treatment group” “Anterior bone formation in immediately Pilot patient; not explicitly noted as an adjacent segments” (1 patient; considered adverse event; considered possibly technique related) possibly technique related) No significant difference in overall Reported all adverse events in 38 Reported all adverse events in 38 categories adverse events (p=0.289). 80 categories Gornet 172 405 2011 345 patients (85.2%) at 24 months Significantly higher proportion of 153 patients (89%) at 24 months 17 (4.2%) possible device-related events possible device-related events in 22 (12.8%) possible device-related events rhBMP-2 group (p<0.001) Maverick‡ Incision related (8/167 patients) Allergic reaction (4/167 patients) Burkus43 172 - 2011 Non-union (7/167 patients) - Control arm not reported Infection (12/167 patients) Elevated temperature (3/167 patients)
283 Appendix XIX – Adverse events occurring in Medtronic trials as reported in published articles
“There were no complications attributable to the rhBMP-2/BCP or TSRH internal fixation” Transient left leg pain for a few weeks after surgery, attributed to irritated nerve root. Subsequently developed right leg pain, underwent decompression one level above the previous surgery for stenosis, with 20 (11 resolution of symptoms (1 patient) 87 with Epidural haematoma. 1 patient with no No adverse events reported for control BCP US Boden 5 2002 TSRH, 9 sequelae after evacuation. 1 patient with group. without residual numbness in both legs after evacuation. 1-year revision decompression performed up to three levels above original surgical level to relieve stenosis from pre- existing ossification of ligamentum flavum. Persistent low back and leg pain, revision with ALIF at 8 months (1 patient). Bilateral solid fusion masses noted at time of revision. 98 No mention of adverse events in Glassman 38† 36† 2005 - - results Gastrointestinal (9 patients) Gastrointestinal (10 patients) Traumas (14 patients) Traumas (9 patients) Cardiovascular (9 patients) Cardiovascular (6 patients) Urogenital (6 patients) Urogenital (6 patients) Dural tears (3 patients) Dural tears (5 patients) 99 Nonsurgical infections (6 patients) Nonsurgical infections (4 patients) Dimar 53† 45† 2006 Malpositioned implants (0 patients) Malpositioned implants (3 patients) Surgical infections (0 patients) Surgical infections (1 patients) Non-unions (0 patients) Non-unions (2 patients) Amplify Respiratory (0 patients) Respiratory (1 patients) Vertebral fractures (0 patients) Vertebral fractures (1 patients) Others (0 patients) Others (17 patients) 101 No mention of adverse events in Glassman 76† 72† 2007 - - results No significant difference in overall Reported all adverse events in 23 Reported all adverse events in 23 categories adverse events (p=0.777). 102 categories Dimar 239 224 2009 198 patients (87%) at 24 months Significantly higher proportion of 209 patients (87%) at 24 months 35 (15.6%) possible device-related events possible device-related events in 21 (8.8%) possible device-related events ICBG group (p<0.032) 43 Cancer (8/234 patients) Burkus 239 - 2011 - Control arm not reported Infection (39/234 patients) *Reports combined data from LT Cage Open and LT Cage Laparoscopic † Results only reported for a subset of patients ‡Control was MAVERICK Disc
284 Appendix XX – Pain and function outcomes reported in publications
Appendix XX
Table 61 Pain and function outcome reporting in publications. Trial Outcome reported Standard deviation/error reported SF36- Back Leg SF36- Back Leg Oswestry PCS pain pain Oswestry PCS pain pain
LT Cage Pilot Y N N N Y N N N LT Cage Open Y Y Y Y Y Y N N Bone Dowel Pilot Y Y Y Y Y Y Y Y Bone Dowel Y N Y Y N N N N Pivotal Interfix PLIF Y Y Y Y N N N N Cornerstone Pilot Y Y Y Y N N N N Mastergraft Pilot Y Y Y Y N N N N BCP US Y Y Y Y N N N N Amplify Y Y Y Y N N N N Glassman Y Y Y Y N N N N
285
Supplementary File
Log of substantive changes made following initial submission of report
Following receipt of peer-review comments on the report • Added an executive summary (pages xiv – xvii)*. • Added timeline (figure 2, pages 25-26)*. • Added data table showing proportion of missing data over time (table 3, page 30). • Added appendix table 28 (page 199-202) showing outcomes listed in each Medtronic protocol*. • Undertaken quality assessment of non-randomised studies from the wider literature using the Newcastle-Ottawa scale in section 8.5 and appendix tables 55 – 57 (pages 258-276)*. • Inclusion of plots for all outcomes in safety results from the wider literature (section 8.5) • All principal analyses of IPD now use complete case analysis approach. • Addition of last observation carried forward analyses undertaken as sensitivity analyses (NB: later replaced with multiple imputation analyses). • Correction to cancer analysis following response from Medtronic (section 8.4). • Addition of sensitivity analyses to compare pilot and pivotal studies (page 41 and appendix XIII figure 54). • Addition of sensitivity analyses to compare INFUSE with AMPLIFY and BCP (page 41). • Addition of sensitivity analysis restricting synthesis of adverse events to INFUSE trials (page 57 and appendix XVI figure 60). • Corrections to some figures and analyses as noted in response to referees comments. • Corrections to values in tables 16 and 17 (section 9.1). • Summary column added to table 19 (section 9.1). • Change of product names in trial identifiers according to information provided in Medtronic’s comments on our draft report in table 1 (pages 21-23) and throughout. • Addition of clarifications as noted in the response to referees comments.
Following receipt of peer review comments on the manuscript submitted to Annals of Internal Medicine • Carried out and incorporated sensitivity analyses that imputed missing values for pain and fusion outcomes (page 42 and appendix XIII figures 58 and 59). • Removed sensitivity analyses of last observation carried forward. • Replaced most figures to match styles used in manuscript. • Some minor text modifications (e.g. Section 7.10 Page 49) to harmonise text between report and manuscript.
During preparation of manuscript reporting on reliability of the evidence • Added figures 42 and 43 illustrating the comparison of IPD with aggregate data from CSRs and publications restricted to published RCTs only. • Added table 20 and cross-referenced this in section 9.3 (page 104) • Corrected total numbers of patients in tables 8 and 9 (pages 61-62). • Expanded some aspects of the wider implications section of the discussion
* These additions were always intended but were not achievable within the initial timeframe
286