Review: Safety and Effectiveness of Recombinant Human Bone Morphogenetic Protein-2 For Spinal Fusion

Annals of Internal Medicine Review
Safety and Effectiveness of Recombinant Human Bone Morphogenetic

Protein-2 for Spinal Fusion
A Meta-analysis of Individual-Participant Data
Mark C. Simmonds, PhD, MA; Jennifer V.E. Brown, MSc, BA; Morag K. Heirs, MSc, MA; Julian P.T. Higgins, PhD, BA;
Richard J. Mannion, PhD; Mark A. Rodgers, MSc, BSc; and Lesley A. Stewart, PhD, MSc, BSc
Background: Recombinant human bone morphogenetic protein-2 adverse events. At 24 months, ODI scores were 3.5% lower (bet-
(rhBMP-2) is widely used to promote fusion in spinal surgery, but ter) with rhBMP-2 than with ICBG (95% CI, 0.5% to 6.5%) and
its safety has been questioned. radiographic fusion was 12% higher (CI, 2% to 23%). At or shortly
after surgery, pain was more common with rhBMP-2 (odds ratio,
Purpose: To evaluate the effectiveness and safety of rhBMP-2. 1.78 [CI, 1.06 to 2.95]). Cancer was more common after rhBMP-2
Data Sources: Individual-participant data obtained from the spon- (relative risk, 1.98 [CI, 0.86 to 4.54]), but the small number of
sor or investigators and data extracted from study publications events precluded definite conclusions.
identified by systematic bibliographic searches through June 2012. Limitation: The observational studies were diverse and at risk of
Study Selection: Randomized, controlled trials of rhBMP-2 versus bias.
iliac crest bone graft (ICBG) in spinal fusion surgery for degenera- Conclusion: At 24 months, rhBMP-2 increases fusion rates, re-
tive disc disease and related conditions and observational studies in duces pain by a clinically insignificant amount, and increases early
similar populations for investigation of adverse events. postsurgical pain compared with ICBG. Evidence of increased can-
Data Extraction: Individual-participant data from 11 eligible of 17 cer incidence is inconclusive.
provided trials sponsored by Medtronic (Minneapolis, Minnesota) Primary Funding Source: Yale University Open Data Access
(n ⫽ 1302) and 1 of 2 other eligible trials (n ⫽ 106) were included. Project.
Additional aggregate adverse event data were extracted from 35
published observational studies.
Data Synthesis: Primary outcomes were pain (assessed with the Ann Intern Med. 2013;158:877-889. www.annals.org
Oswestry Disability Index [ODI] or Short Form-36), fusion, and For author affiliations, see end of text.
R ecombinant human bone morphogenetic protein-2

(rhBMP-2) is widely used as an alternative to iliac crest
bone graft (ICBG) to promote fusion in spinal surgery (1,
are presented in our full report and have been submitted
for publication elsewhere (5a).
2). Since the U.S. Food and Drug Administration (FDA) METHODS
approved rhBMP-2 for anterior lumbar interbody fusion Eligibility, Search, Data Collection, and
(ALIF) surgery (3), its use has grown rapidly, including Critical Assessment
off-label indications (2, 4). A review of publicly available Methods were prespecified (in advance of detailed
data suggesting that the risk for adverse events is 10 to 50 knowledge of the IPD to be provided) in a protocol (Sup-
times higher than reported in trial publications (5) raised plement 1, available at www.annals.org) that was registered
concerns about the safety of rhBMP-2. in PROSPERO in February 2012 (CRD42012001907)
The Yale University Open Data Access (YODA) Proj- (6).
ect team invited Medtronic (Minneapolis, Minnesota) to All RCTs that compared rhBMP-2 with ICBG in
provide full data from all of its trials of rhBMP-2 to allow spinal fusion surgery regardless of spinal level or surgical
independent reanalysis. The project team subsequently in- approach were eligible for inclusion in our principal anal-
vited proposals to undertake independent evaluation and ysis. We included trials of the licensed INFUSE formula-
funded the Centre for Reviews and Dissemination and 1
other group to do so, thus enabling meta-analysis of
individual-participant data (IPD), which is regarded as a See also:
“gold standard” approach to evidence synthesis.
We embedded our IPD meta-analysis within a system- Print
atic review and sought to examine all relevant evidence. Editorial comments . . . . . . . . . . . . 910, 912, 914, 916
Investigation of comparative effectiveness was restricted to Related article. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 890
randomized, controlled trials (RCTs). In addition, to in- Web-Only
vestigate the safety of rhBMP-2, we sought all observa- Supplements (protocol and Annals peer review
tional studies of its use in spinal surgery that reported ad- materials)
verse events. Findings of our investigation of reporting bias
© 2013 American College of Physicians 877
Downloaded from https://annals.org by guest on 04/06/2019

Review Recombinant Human Bone Morphogenetic Protein-2 for Spinal Fusion
tion (rhBMP-2 concentration, 1.5 g/L) and unlicensed Statistical Methods

AMPLIFY (rhBMP-2 concentration, 2 g/L) and biphasic Appendix 2 provides details of statistical methods and
calcium phosphate (BCP) formulations. For supporting analyses. In the analyses of effectiveness for continuously
analyses of safety, we included all studies of more than 10 distributed outcomes (such as ODI score), we calculated
adult participants that compared rhBMP-2 with any other mean differences between treatment groups in the change
spinal fusion technique and reported adverse events. in score from preoperative values. For dichotomous out-
Our reanalysis of the Medtronic data was done in the comes (such as successful fusion), we calculated relative
context of a full systematic review. We performed a sys- risks (RRs). Both were calculated separately for every trial
tematic literature search of the Cochrane Central Register at every time point. We then used standard random-effects
of Controlled Trials, MEDLINE, EMBASE, and Science meta-analytic techniques (10, 11) to combine effect esti-
Citation Index in January 2012 and automated “current mates across trials. Separate meta-analyses were done for
awareness” searches up to June 2012 to identify eligible each of the specified time points. Linear and logistic
studies not provided by Medtronic. We also searched random-effects regression models were used to combine all
ClinicalTrials.gov to identify ongoing or unpublished ran- data from all trials in “1-stage” meta-analyses as sensitivity
domized trials and published a call for evidence. Detailed analyses (12, 13). We used multiple imputation to explore
search strategies are provided in Appendix 1 (available at the influence of missing observations.
www.annals.org). Heterogeneity was assessed in all meta-analyses by us-
One researcher collated IPD from across the SAS data ing the Higgins I2 statistic (14) and the Cochran Q test.
files (SAS Institute, Cary, North Carolina) provided by We performed a subgroup analysis (stratified by trial) to
Medtronic, and a second researcher checked the process. examine whether effects varied according to the type of
The data were checked for completeness, internal consis- spinal surgery or by rhBMP-2 formulation (INFUSE or
tency, improbable values, and balance of patient character- AMPLIFY). We investigated whether patient-level factors
istics across treatment groups (see Appendix 2, available at (age, sex, smoking, alcohol consumption, body mass index,
www.annals.org, for details). Three researchers indepen- diabetic status, and history of spinal surgery for back pain)
dently working in pairs performed study selection and data were associated with the effectiveness of rhBMP-2 surgery
extraction from published reports. Disagreements were re- by using a 1-stage random-effects regression model (12)
solved by discussion or by referring to the third researcher. that included interaction terms between patient-level fac-
We assessed risk of bias by using the Cochrane Collabora- tors and treatment.
tion’s “risk-of bias” tool in the RCTs and a modified form Safety Analysis
of the Newcastle-Ottawa Scale for nonrandomized studies We examined numbers of adverse events, including
(8, 9) (see Appendix 2 for details). Risk of bias was assessed cancer, provided in the IPD according to the Medtronic
by at least 2 researchers independently, with disagreements classifications of adverse events. Medtronic did not provide
resolved by discussion. case report forms, but summaries of cases were given for
Effectiveness Analysis
most patients in the clinical study reports. A full clinical
assessment of these narratives was beyond the scope of this
Our prespecified primary outcomes were those likely
project, but checks in 1 trial (Inter Fix PLIF [15]) showed
to be important to patients. The Oswestry Disability Index
that the Medtronic classifications of events seemed appro-
(ODI) and Neck Disability Index for cervical spinal sur-
priate. To our knowledge, these data represent all adverse
gery measure lower back and neck pain, respectively, on a
events that occurred during the follow-up periods of these
scale from 0% (no pain) to 100% (extreme pain). The
trials. Because the number of specific adverse events in
Short Form-36 (SF-36) Physical Component Summary
most trials was generally small, we used 1-stage random-
(PCS) assesses pain and physical function on a scale from
effects logistic regression meta-analysis models (13) to an-
0% (worst) to 100% (best). Medtronic also provided data
alyze these data (see Appendix 2 for model details). Results
on back and leg pain, which were measured on a scale from
of these analyses are presented as odds ratios.
0 (no pain) to 20 (extreme pain). Spinal fusion was as-
We extracted data on number of adverse events from
sessed as success or failure according to Medtronic’s radio-
other published studies that compared spinal fusion sur-
graphic definition, which required evidence of bridging
gery using rhBMP-2 with ICBG and other comparators in
trabeculae, no evidence of motion (⬍3-mm difference in
at least 10 adult participants, as specified in our protocol.
translation and ⬍5-degree difference in angular motion),
and no evidence of radiolucency. We analyzed these out- Role of the Funding Source
comes at 6 weeks and 3, 6, 12, and 24 months after sur- This review was funded by the YODA Project, which
gery. We did not analyze the limited data that were avail- provided the IPD for and other materials relating to the
able for longer follow-up times for few participants. Medtronic trials but was not involved in the analyses of
We also considered 4 secondary outcomes: duration of these data or in the production of this paper. There was no
hospital stay, operating time, successful return to work or direct contact with Medtronic or the other evaluation
usual activity, and use of pain medication. team. Our full report to the YODA Project is publicly
878 18 June 2013 Annals of Internal Medicine Volume 158 • Number 12 www.annals.org

Recombinant Human Bone Morphogenetic Protein-2 for Spinal Fusion Review
Figure 1. Summary of evidence search and selection.
References identified in database search (n = 6807)
Excluded during de-duplication (n = 4073)

Additional records identified from bibliography
hand searches and electronic update searches
(n = 103)
Screening of titles and abstracts (n = 2837)
Materials provided by Medtronic

Excluded (n = 2521)
IPD for 17 trials (efficacy and
safety data)
Protocols and publications for
Full-text screening (n = 316)
the trials
Excluded (n = 236)
Full text could not be
Excluded Medtronic trials (n = 6) retrieved: 3
Single-group trial: 4 Inclusion criteria not met: 233
Trial with no ICBG group: 1
Trial of 3 patients: 1
Effectiveness analysis* Safety analyses

IPD from 11 Medtronic RCTs IPD from 11 Medtronic RCTs
IPD from 1 additional RCT (18) 43 publications reporting on 35
additional controlled adverse
events studies (not IPD)
ICBG ⫽ iliac crest bone graft; IPD ⫽ individual-patient data; RCT ⫽ randomized, controlled trial.
* One additional RCT (19) was eligible, but IPD were not provided.
available and examines a wide range of issues around the not contribute to our analyses. We identified 1 ongoing
effectiveness and safety of rhBMP-2 in spinal surgery (16, 17). trial for which recruitment had been suspended, but
we could not include it because it had not been closed
RESULTS (20).
The YODA project team provided IPD from 17 Our IPD meta-analysis was based on data from 1302
Medtronic trials. Eleven of these were RCTs comparing patients in 11 Medtronic RCTs that compared rhBMP-2
rhBMP-2 with ICBG surgery and were eligible for inclu- with ICBG surgery, as well as data from 106 participants
sion in our principal evaluation of effectiveness. Of the in 1 additional RCT (18) (Table 1).
others, 4 were single-group trials of rhBMP-2 and 1 (7) Appendix Table 1 (available at www.annals.org) pro-
used a different comparator. We included these only in our vides details of publications associated with each
supporting consideration of adverse events. We did not Medtronic trial. However, our analyses used the supplied
consider 1 trial that was stopped early after recruiting only IPD rather than data reported in these publications.
3 patients. For all of the included trials, data on all pain outcomes
Figure 1 shows search results. In addition to the trials were available at all time points from 6 weeks to 24
supplied by Medtronic, we identified 2 eligible randomized months. Data on spinal fusion were available for all
trials not conducted by Medtronic that compared Medtronic trials except the LT-CAGE pilot trial (31) at all
rhBMP-2 with ICBG surgery. We requested IPD from the times from 6 months onward. The trial by Glassman and
authors and obtained them from 1 study (18). Data from colleagues (18) provided fusion data in a different format
the other trial, which involved 40 patients having single- and is not included in these analyses. Appendix Table 2
level bilateral posterior lateral interbody fusion, were un- (available at www.annals.org) summarizes the levels of
available (19). This trial reported 100% fusion in both missing data for the main pain and fusion outcomes. At 24
groups and no difference in back pain or ODI score at 12 months after surgery, outcome data were not available for
months and therefore these published aggregate data could approximately 15% of participants.
www.annals.org 18 June 2013 Annals of Internal Medicine Volume 158 • Number 12 879

Table 1. Summary of Trials That Provided Individual-Participant Data Used in the Analyses
Trial (Reference) Date of First Surgical Approach Intervention/Control Patients, Patients Patients Mean Male, Previous
Patient n Excluded, Evaluated, Age, y % Spinal
Enrollment n n Surgery, %
LT-CAGE January ALIF (single-level rhBMP-2/ACS plus LT-CAGE 11 0 11 42 45 0
pilot (21) 1997 open/laparoscopic) (1.5 mg/mL)
ICBG plus LT-CAGE 3 0 3 40 67 0
LT-CAGE August 1998 ALIF (single-level rhBMP-2/ACS plus LT-CAGE 145 2 143 43 55 38
open (22) open) (1.5 mg/mL)
ICBG plus LT-CAGE 137 1 136 42 50 40
Bone dowel April 1998 ALIF (single-level rhBMP-2/ACS plus cortical 24 0 24 41 33 46
pilot (23) open) bone dowels (1.5 mg/mL)
ICBG plus cortical bone 23 1 22 45 45 32
dowels
Bone dowel September ALIF (single-level rhBMP-2/ACS plus cortical 64 9 55 39 44 33
pivotal (24) 2000 open) bone dowels (1.5 mg/mL)
ICBG plus cortical bone 31 1 30 42 30 33
dowels
Inter Fix March 1999 PLIF (single-level) rhBMP-2/ACS plus Inter Fix 35 1 34 46 50 35
PLIF (15) cages (1.5 mg/mL)
ICBG plus Inter Fix cage 36 3 33 46 45 39
Cornerstone September ACIF (1- or 2-level) rhBMP-2/ACS plus 18 0 18* 51 44 6
pilot (25) 1999 Cornerstone implant plus
Atlantis plate (1.5 mg/mL)
ICBG/Cornerstone implant 15 0 15† 47 47 0
plus Atlantis plate
Mastergraft April 2003 PLF (single-level) rhBMP-2/ACS/Mastergraft/ 27 2 25 56 0 24
pilot (26) CD Horizon (12 mg)
ICBG plus CD Horizon 23 2 21 57 0 29
Inter Fix ALIF February ALIF (single-level rhBMP-2/ACS plus Novus 25 0 25 45 44 44
pilot (27) 1999 open) LC Device
ICBG plus Novus LC 20 0 20 45 45 35
BCP U.S. (28) May 1999 PLF (single-level) rhBMP-2/BCP (3 mg/mL); 22 0 11 50 55 18
rhBMP-2/BCP plus TSRH 11
(2 mg/mL)
ICBG plus TSRH 5 0 5 52 40 0
BCP Canada (29) September PLF (1- or 2-level) rhBMP-2/BCP plus CD 102 4 98 53 36 19
1999 Horizon or rhBMP-2/BCP
plus TSRH
ICBG plus CD Horizon or 105 6 99 53 48 20
ICBG plus TSRH
AMPLIFY (30) March 2002 PLF (open bilateral) rhBMP-2/CRM/CD Horizon 262 23 239 53 45 31
(2 mg/mL)
ICBG plus CD Horizon 256 32 224 52 42 28
Glassman et al May 2004 PLF (single-level or rhBMP-2/ACS plus fixation NA NA 52 69 30 32
(18)‡ multilevel) ICBG plus fixation NA NA 54 70 33 37
ACIF ⫽ anterior cervical interbody fusion; ACS ⫽ absorbable collagen sponge; ALIF ⫽ anterior lumbar interbody fusion; BCP ⫽ biphasic calcium phosphate; CRM ⫽
compression-resistant matrix; ICBG ⫽ iliac crest bone graft; NA ⫽ not applicable; PLF ⫽ posterior lumbar fusion; PLIF ⫽ posterior lumbar interbody fusion; rhBMP-2 ⫽
recombinant human bone morphogenetic protein-2; TSRH ⫽ Texas Scottish Rite Hospital.
* Ten patients had single-level surgery.
† Eight patients had single-level surgery.
‡ Not a Medtronic trial.
Our assessment of risk of bias was the same for all after surgery onward, the use of rhBMP-2 generally
Medtronic trials. Randomization and allocation conceal- achieved greater pain reduction (from preoperative values)
ment procedures were adequate for all trials. Neither pa- than did ICBG. Among rhBMP-2 recipients, the ODI
tients nor physicians were blinded to the treatment re- score was approximately 3.5 percentage points better
ceived, and all pain and function outcomes were patient- (mean difference, ⫺3.48 percentage points [95% CI,
assessed, so there was a potential for bias in these ⫺6.47 to ⫺0.49 percentage points]; I2 ⫽ 38%) and back
outcomes. Successful fusion was assessed by researchers pain was better by more than 1 point on the 20-point scale
blinded to the treatment received. used (mean difference, ⫺1.58 [CI, ⫺2.65 to ⫺0.51]; I2 ⫽
Effectiveness 44%) at 24 months after surgery. The SF-36 PCS score
Pain was 1.93 percentage points higher for rhBMP-2 recipients
Figure 2 shows results of meta-analyses across the 12 (CI, 0.63 to 3.22 percentage points; I2 ⫽ 0%) at 24
RCTs for 4 pain outcomes (the SF-36 PCS also incorpo- months. We found no evidence of a difference in leg pain
rates an assessment of physical function). From 6 months reduction between treatment groups (mean difference,

⫺0.59 [CI, ⫺1.27 to 0.09]; I2 ⫽ 0%). In general, patients (I2 ⫽ 97%, 80%, and 76% at 6, 12, and 24 months,
in both groups improved considerably over time such that respectively).
the extra benefit of rhBMP-2 over ICBG surgery was small Results from the 1-stage logistic regression meta-
in comparison (Appendix Figure 1, available at www analysis model for successful fusion were almost identical
.annals.org). The ODI score improved by approximately to the results presented earlier, as were those from models
26 percentage points at 24 months for ICBG recipients that used multiple imputation of missing fusion data.
and by 30 percentage points for rhBMP-2 recipients.
Results of the 1-stage linear regression meta-analysis
Investigation by Surgical Approach
models of pain outcomes were almost identical to the re-
sults presented earlier, as were those from models that used We performed subgroup analyses to investigate
multiple imputation of missing pain data (see Appendix whether the effectiveness of rhBMP-2 varied among pa-
Figure 2, available at www.annals.org). tients who had anterior lumbar fusion, posterior lumbar
fusion, or anterior cervical fusion (Table 1). Appendix Fig-
ure 3 (available at www.annals.org) shows the results of
Fusion these analyses for ODI score and successful fusion 24
Figure 3 shows a forest plot for successful fusion 24 months after surgery. A test for heterogeneity showed
months after surgery, at which time rhBMP-2 increased moderate evidence of a difference between surgery types for
fusion rates by 12% (RR, 1.12 [CI, 1.02 to 1.23]). In- ODI score (P ⫽ 0.065), but this was primarily due to the
creased fusion rates were also identified at 6 months (RR, large benefit of rhBMP-2 on the Neck Disability Index
1.20 [CI, 1.00 to 1.44]) and 12 months (RR, 1.11 [CI, score observed in the single small cervical surgery trial (n ⫽
1.00 to 1.22]) after surgery. However, we found substantial 23). Excluding this trial resulted in no clear difference in
heterogeneity of the RR for successful fusion across trials the effectiveness of rhBMP-2 between anterior or posterior
Figure 2. Meta-analyses of pain outcomes at 6 wk and 3, 6, 12, and 24 mo after surgery.
A. Back pain B. Leg pain

Mean Difference in Outcome

(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)
2 2
0 0
–2 –2
–4 –4
–6 –6
0 5 10 15 20 25 0 5 10 15 20 25
Time Since Surgery, mo Time Since Surgery, mo
C. Oswestry Disability Index D. SF-36 PCS


(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)
2 2
0 0
–2 –2
–4 –4
–6 –6
0 5 10 15 20 25 0 5 10 15 20 25
Points on the plot represent mean differences in changes in scores (from preoperative values) at each time point, and vertical lines show the 95% CIs. For
the first 3 outcomes, points below the dotted line indicate a benefit of rhBMP-2; for the SF-36 PCS score, points above the line indicate a benefit of
rhBMP-2. ICBG ⫽ iliac crest bone graft; PCS ⫽ Physical Component Summary; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2;
SF-36 ⫽ Short Form-36.

Figure 3. Forest plot of RR for successful fusion 24 mo after surgery.
Trial (Reference) Fused/Total, n/N RR (95% CI)

rhBMP-2 ICBG
LT-CAGE open (22) 123/130 106/119 1.06 (0.99–1.15)

Bone dowel pilot (23) 24/24 13/19 1.44 (1.07–1.94)
Bone dowel pivotal (24) 43/44 22/27 1.20 (1.00–1.44)
Inter Fix ALIF pilot (27) 16/20 13/14 0.86 (0.66–1.12)
Cornerstone pilot (25) 11/11 11/11 1.00 (0.84–1.19)
Inter Fix PLIF (15) 24/26 21/27 1.19 (0.94–1.49)
Mastergraft pilot (26) 18/19 14/20 1.35 (1.00–1.84)
BCP U.S. (28) 17/21 3/3 0.81 (0.67–1.00)
BCP Canada (29) 86/89 64/91 1.37 (1.20–1.58)
AMPLIFY (30) 186/194 151/169 1.07 (1.01–1.14)
Combined 1.12 (1.02–1.23)
0.50 0.75 1.00 1.50 2.00
ICBG better rhBMP-2 better
ALIF ⫽ anterior lumbar interbody fusion; BCP ⫽ biphasic calcium phosphate; ICBG ⫽ iliac crest bone graft; PLIF ⫽ posterior lumbar interbody
fusion; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2; RR ⫽ relative risk.
approaches (P ⫽ 0.171). We found no evidence of a dif- and leg pain) (Appendix Figure 6, available at www.annals
ference in the RRs for successful fusion (P ⫽ 0.88) or any .org). We found no evidence of a consistent relationship
other outcome at 24 months across surgery types. between improvements in fusion due to rhBMP-2 and im-
Evidence of interactions between rhBMP-2 and the provements in pain or function. If successful fusion re-
patient-level factors (age, sex, smoking, alcohol consump- sulted in reduced pain, we would expect trials that showed
tion, body mass index, diabetic status, and history of spinal higher fusion rates with rhBMP-2 to show greater im-
surgery) was generally lacking—that is, each factor bene- provement in pain scores, but this did not seem to be the
fited from rhBMP-2 to the same extent. One possible ex- case. In particular, trials where fusion was less common in
ception was that, for persons with a previous spinal sur- the rhBMP-2 recipients (Inter Fix ALIF pilot [27] and
gery, there was no difference in the effectiveness of BCP U.S. [32]) still showed a benefit of rhBMP-2 in terms
rhBMP-2 and ICBG at reducing ODI score or improving of improved SF-36 PCS score. The BCP U.S. trial also
fusion rates. Given the number of analyses done, this may showed a benefit of rhBMP-2 on ODI score and back
be a chance finding. These results are available in our full pain. Therefore, the apparent small benefits of rhBMP-2 in
report (17). pain reduction do not seem to be due to increased fusion
rates.
Secondary Outcomes
We found no evidence of difference in duration of Safety
hospital stay (mean difference, ⫺0.15 days [CI, ⫺0.33 to All Medtronic trials provided data on adverse events at
0.03 days]) or that rhBMP-2 surgery increased the proba- all specified time points and also at or shortly after (that is,
bility of returning to work or usual activity earlier com- up to 4 weeks after) surgery. Because reporting of adverse
pared with ICBG (RR at 24 months, 1.01 [CI, 0.88 to events in the trial by Glassman and colleagues (18) was not
1.17]). Using rhBMP-2 shortened operating times by 21 consistent with that in the Medtronic trials, it was not
minutes (CI, 15 to 27 minutes) (Appendix Figure 4, avail- included in these analyses; however, the data are included
able at www.annals.org) from an average of 135 minutes. in our full report (17).
We found no evidence that analgesic use differed between We note that pain was reported as an adverse event in
the rhBMP-2 and ICBG groups at any time (Appendix the Medtronic IPD as well as being assessed as an effective-
Figure 5, available at www.annals.org). ness outcome using the pain scales discussed earlier. The
We investigated the association between successful fu- reasons for this were not clear, but for completeness, we
sion and change in pain score by comparing the mean analyze pain reported as an adverse event, particularly be-
difference in pain at 24 months after surgery with the RR cause pain immediately after surgery was not recorded on
for successful fusion at the same time point for each of the the pain scales—its presence or absence was recorded only
4 pain outcomes (ODI score, SF-36 PCS score, back pain, as an adverse event.

Figure 4 shows the results of the 1-stage meta-analyses pain reduction was greater in the rhBMP-2 recipients from
for adverse events across the 11 Medtronic RCTs at or 3 months after surgery onward.
shortly after surgery. Risks for arthritis and bursitis, Cancer
implant-related events, neurologic events, other pain, ret- Table 2 summarizes the cancer cases observed in the
rograde ejaculation, wound complications, and vascular 11 Medtronic RCTs, 5 of which observed at least 1 case. It
events increased by at least 50% among rhBMP-2 recipi- excludes preexisting cancer but includes 3 cases in the LT-
ents. Because there were few events, CIs were wide and CAGE open pivotal trial (33) that were identified during
findings were inconclusive. For back and leg pain, we extended follow-up of only the rhBMP-2 recipients. How-
found clear evidence of a higher incidence among ever, these 3 cases were not included in the quantitative
rhBMP-2 recipients (odds ratio, 1.92 [CI, 1.14 to 3.25]; analyses because this would have biased against rhBMP-2
P ⫽ 0.004). Appendix Figure 7 (available at www.annals because any equivalent cases occurring in the ICBG group
.org) shows the results of the 1-stage meta-analyses for ad- had not been sought. A 1-stage random-effects meta-
verse events over all times up to 24 months after surgery. analysis model found that cancer was nearly twice as com-
As in the analysis in Figure 4, results were generally mon among rhBMP-2 recipients (RR, 1.98 [CI, 0.86 to
inconclusive. 4.54]), but the 95% CIs were consistent with risk in
Appendix Figure 8 (available at www.annals.org) rhBMP-2 recipients being anywhere from 14% lower to
shows the results of meta-analyses for 4 key adverse event 454% higher. The absolute risk for cancer was low (3% in
categories (implant-related, infections, neurologic, and any rhBMP-2 recipients). A forest plot for the equivalent
pain) across all periods. Events were uncommon, so find- 2-stage analysis is shown in Figure 5 (RR, 1.84 [CI, 0.81
ings are mostly inconclusive. The only clear evidence of a to 4.16]). The RR for cancer was similar across trials. In
difference was for pain at or shortly after surgery, which particular, the RR for cancer in the AMPLIFY trial (34,
was more common in rhBMP-2 recipients (odds ratio, 35), which used a different preparation of rhBMP-2 at a
1.78 [CI, 1.06 to 2.95]; P ⫽ 0.007). This contrasts with higher dose, was no greater than in trials that used
the results seen in the analyses of ODI score, SF-36 PCS INFUSE (P ⫽ 0.82). We note that in addition to the 3
score, and back pain in the effectiveness analyses, where cancer cases identified during additional follow-up, 3 cases
Figure 4. Meta-analysis of adverse events at or shortly after surgery in 11 Medtronic trials.
Adverse Event ICBG, n rhBMP-2, n OR (95% CI)
Arm and neck pain 2 1 0.65 (0.06–7.24)

Arthritis/bursitis 1 3 2.84 (0.29–27.51)
Back and leg pain 24 39 1.92 (1.13–3.25)
Cardiovascular 65 60 0.99 (0.68–1.44)
Dural injury 10 11 1.49 (0.62–3.62)
Dysphagia 2 1 0.30 (0.02–3.59)
Gastrointestinal 73 67 1.04 (0.73–1.48)
Implant 6 10 2.11 (0.73–6.07)
Infection 47 53 1.23 (0.81–1.85)
Neurologic 20 29 1.73 (0.96–3.12)
Other pain 9 12 1.68 (0.69–4.07)
Respiratory 16 19 1.27 (0.64–2.51)
Retrograde ejaculation 1 3 3.00 (0.31–29.15)
Spinal 7 4 0.53 (0.15–1.85)
Trauma 8 8 0.93 (0.34–2.55)
Urogenital 35 42 1.38 (0.86–2.21)
Vascular 4 6 1.94 (0.53–7.18)
Wound complication 3 4 1.76 (0.39–8.06)
0.25 0.50 1.00 2.00 5.00 10.00
More common with ICBG More common with rhBMP-2
Details on the Medtronic trials are available in Appendix Table 1 (available at www.annals.org). ICBG ⫽ iliac crest bone graft; OR ⫽ odds ratio;
rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2.

(7), we did not have IPD from any of them. The studies
Table 2. Incidence of Cancer in the Medtronic Trials*
are summarized in Appendix Table 3 (available at www
.annals.org). Quality assessment found that all of the stud-
Cancer rhBMP-2 (n ⴝ 694) ICBG (n ⴝ 608)
ies included patients who were representative of those
Skin
Melanoma 2 0
likely to receive treatment in practice and clearly estab-
Basal cell carcinoma 0 1 lished exposure to treatment. However, most made no at-
Squamous cell carcinoma 1 1 tempt to match or control for potential confounding fac-
Breast 2 2
Colon 1 1
tors, and data on the comparability of the treatment groups
Larynx 1 0 were generally unreported or limited. Given the heteroge-
Leukemia 1 0 neity of these studies and their potential for bias, we did
Lung 1 0
Non-Hodgkin lymphoma 1 1
not meta-analyze these data.
Ovarian 1 0 Despite the methodological issues, we found evidence
Pancreatic 2 0 suggestive of higher rates of particular adverse events
Prostate 2 1
Stomach 1 0
among rhBMP-2 recipients (Figure 6). Among studies re-
Testicular 1 0 porting at least 1 event, heterotopic bone formation (re-
Thyroid 3 1 ported in 5 studies) was more common among rhBMP-2
Total 20 8
recipients, although whether this led to any clinical conse-
ICBG ⫽ iliac crest bone graft; rhBMP-2 ⫽ recombinant human bone morpho- quences for these patients was unclear. Leg pain and radic-
genetic protein-2. ulitis (4 studies) seemed to be more common, as had been
* The data shown are the number of cases and include 3 cases (1 of thyroid cancer,
1 of testicular cancer, and 1 of melanoma) in rhBMP-2 recipients identified observed in the Medtronic trials. Osteolysis was more com-
through extended follow-up of the LT-CAGE open trial (22). mon, but only 2 studies reported on this event. Dysphagia
(6 studies) seemed to be more common among rhBMP-2
recipients having cervical spinal surgery, although results of
were seen among rhBMP-2 recipients in the Maverick trial
these studies were inconsistent. Comer and colleagues (39)
(7) and 2 were seen in a single-group Medtronic trial (36 –
compared 4 consecutive-patient cohorts undergoing ALIF
38). These were not included in our analyses because nei-
with or without rhBMP-2 and reported a higher rate of
ther trial had an ICBG comparator.
retrograde ejaculation among rhBMP-2 recipients (6.3%
Adverse Events Reported in the Literature vs. 0.9%; P ⫽ 0.001). Further adverse event outcomes
We identified 35 observational studies of at least 10 were examined in our full report.
adult patients (in 43 publications) that reported adverse
effects of rhBMP-2. There were 14 studies of posterior
lumbar fusion, 5 of anterior lumbar fusion, 10 of cervical DISCUSSION
fusion, and 8 that used multiple spinal fusion procedures. Our principal analyses were based on data from 1408
These studies used various spinal fusion techniques as con- individual participants in 11 eligible RCTs, including all
trols, including ICBG, local bone graft, allograft, and bone trials sponsored by Medtronic (published and unpub-
marrow aspirates. Other than the Medtronic Maverick trial lished) and 1 additional trial. We found the randomization
Figure 5. Forest plot of cancer incidence in the Medtronic trials.
Trial (Reference) Cancer/Total, n/N RR (95% CI)

rhBMP-2 ICBG
LT-CAGE open (22) 2/143 1/136 1.92 (0.18–20.89)

Bone dowel pilot (23) 1/24 0/22 2.76 (0.12–64.23)
Bone dowel pivotal (24) 1/55 0/30 1.65 (0.07–39.25)
BCP Canada (29) 1/98 2/99 0.51 (0.05–5.48)
AMPLIFY (30) 12/239 5/224 2.25 (0.81–6.29)
Combined 1.84 (0.81–4.16)
0.1 1.0 5.0 10.0 20.0
rhBMP-2 better ICBG better
The number of cancer cases is the total number occurring during the 2-y follow-up. Details on the Medtronic trials are available in Appendix Table 1,
available at www.annals.org. BCP ⫽ biphasic calcium phosphate; ICBG ⫽ iliac crest bone graft; rhBMP-2 ⫽ recombinant human bone morphogenetic
protein-2; RR ⫽ relative risk.

Figure 6. Heterotopic bone formation, radiculitis, and dysphagia in nonrandomized studies of rhBMP-2.
Author, Year (Reference) Comparator Surgery Events/Total, n/N RR (95% CI)

rhBMP-2 Comparator
Heterotopic bone formation

Rihn et al, 2009 (93) ICBG TLIF 2/86 0/33 1.94 (0.10–39.29)
Mannion, 2010 (91) ICBG PLIF/TLIF 2/17 0/19 5.57 (0.29–108.27)
Gray and Rampersaud, 2010 (99) Autologous bone PLF 43/82 10/39 2.05 (1.15–3.63)
Joseph and Rampersaud, 2007 (100) Autologous bone PLIF/TLIF 5/23 1/10 2.17 (0.29–16.30)
Pimenta et al, 2011 (110) SCP LIF 1/15 0/15 3.00 (0.13–68.09)
Radiculitis
Rihn et al, 2009 (93) ICBG TLIF 12/86 1/33 4.60 (0.62–34.03)
Gray and Rampersaud, 2010 (99) Autologous bone PLF 2/82 0/39 2.39 (0.12–48.69)
Mindea et al, 2009 (98) Autologous bone TLIF 4/35 0/8 2.15 (0.13–36.31)
Rowan et al, 2011 (107) No rhBMP-2 TLIF 11/64 3/40 2.29 (0.68–7.72)
Dysphagia
Longley et al, 2009 (101) No rhBMP-2 ACDF 1/84 1/84 1.00 (0.06–15.73)
Smucker et al, 2006 (102) No rhBMP-2 ACDF 5/69 2/165 5.98 (1.19–30.08)
Vaidya et al, 2007 (103) No rhBMP-2 ACDF 13/20 4/18 2.92 (1.16–7.36)
Xu et al, 2011 (105) No rhBMP-2 PCF 3/48 6/156 1.62 (0.42–6.25)
Williams et al, 2011 (104) No rhBMP-2 ACDF 10/4532 4/352 0.19 (0.06–0.62)
Yaremchuk 2010 (106) No rhBMP-2 CSA 18/260 17/515 2.10 (1.10–4.00)
0.25 1.00 5.00 10.00 20.00 40.00
More common with comparator More common with rhBMP-2
ACDF ⫽ anterior cervical discectomy and fusion; ICBG ⫽ iliac crest bone graft; CSA ⫽ cervical spinal arthrodesis; LIF ⫽ lumbar interbody fusion; PCF
⫽ posterior cervical fusion; PLF ⫽ posterolateral lumbar fusion; PLIF ⫽ posterior lumbar interbody fusion; rhBMP-2 ⫽ recombinant human bone
morphogenetic protein-2; RR ⫽ relative risk; SCP ⫽ silicated calcium phosphate; TLIF ⫽ transforaminal lumbar interbody fusion.
procedures to be adequate in all trials, but participants In general, successful fusion and pain reduction do not
were not blinded to treatment. Although assessment of seem to be strongly correlated. Trials with higher fusion
some outcomes, such as radiologic assessment of fusion, rates for rhBMP-2 did not also achieve greater pain reduc-
was blinded, patient-reported outcomes related to pain tion. In the trials with lower fusion rates among rhBMP-2
were not. Follow-up was reasonably complete up to our recipients, pain reduction was still greater among these pa-
final analysis time point of 24 months. Although there is tients. It therefore seems that either rhBMP-2 surgery has
some potential for bias associated with patient-reported an effect on pain beyond that from fusion—which seems
outcomes, in general, we consider the body of evidence for medically unlikely— or the interpretation of pain was bi-
comparative effectiveness to be strong. ased. Because participants were not blinded to the treat-
We found clear evidence that rhBMP-2 improves rates ment received or their fusion status, they may have re-
of fusion compared with ICBG; however, the Medtronic ported exaggerated benefits of the “new” treatment, thus
definitions of fusion that we used may have been stringent biasing assessment in favor of rhBMP-2.
given that only 69% of ICBG recipients achieved fusion In contrast, the analysis of adverse events reported in
within 24 months, which is lower than would be expected the IPD showed an increased risk for pain associated with
generally. Inconsistency across trials was high, with large I2 rhBMP-2 in the immediate postsurgical period. Although
values at all time points. this may seem to contradict the finding that rhBMP-2
We also found that rhBMP-2 improves back pain and reduces pain from 6 months onward, rhBMP-2 surgery
quality of life compared with ICBG at between 6 and 24 may lead to increased pain shortly after surgery but re-
months after surgery. However, these improvements in duced pain in the longer term.
pain fall below previously described, clinically meaningful The IPD also indicate that rhBMP-2 may be associ-
thresholds (estimated as between 4 and 17 percentage ated with an increased risk for cancer, with nearly double
points for ODI score and ⱖ5.4 points for SF-36 PCS [40, the number of new cancer cases compared with ICBG re-
41]). cipients. The overall absolute risk for cancer is low in both

groups, however, so whether this increased risk is genuine .acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum⫽M12

is uncertain, but it is consistent with the literature suggest- -2603.
ing a possible link between BMP and cancer (42).
Adverse event data in the literature raise concerns that Reproducible Research Statement: Study protocol: Available in Supple-
rhBMP-2 may increase the risk for heterotopic bone for- ment 1 (available at www.annals.org). Statistical code: Not available be-
cause of large voulme. Data set: Available from the YODA Project (http:
mation, osteolysis, radiculitis, and retrograde ejaculation.
//medicine.yale.edu/core/projects/yodap/index.aspx).
However, these findings should be interpreted cautiously
because they are based on only published nonrandomized
Requests for Single Reprints: Lesley A. Stewart, PhD, MSc, BSc, Cen-
studies, most of which provided little information about
tre for Reviews and Dissemination, University of York, York YO10
the comparability of groups. 5DD, United Kingdom.
Our review differs from the existing review (5) in that
we had access to more extensive and detailed data than did
Current author addresses and author contributions are available at www
Carragee and colleagues, who used aggregate data extracted .annals.org.
from publications of industry-sponsored trials and publicly
available FDA summaries and public meeting documents.
The FDA materials seem to provide incomplete outcome References
data from a subset of trials evaluating rhBMP-2. We ana- 1. Deyo RA, Ching A, Matsen L, Martin BI, Kreuter W, Jarvik JG, et al. Use
lyzed IPD from all Medtronic-sponsored trials regardless of of bone morphogenetic proteins in spinal fusion surgery for older adults with
whether they had been published or submitted to the lumbar stenosis: trends, complications, repeat surgery, and charges. Spine (Phila
Pa 1976). 2012;37:222-30. [PMID: 21494195]
FDA. This review included the licensed INFUSE prepara- 2. Cahill KS, Chi JH, Day A, Claus EB. Prevalence, complications, and hospital
tion of rhBMP-2 and the unlicensed AMPLIFY and charges associated with use of bone-morphogenetic proteins in spinal fusion pro-
BCP preparations, which use a higher dose of rhBMP-2. cedures. JAMA. 2009;302:58-66. [PMID: 19567440]
We found no evidence of a difference in effectiveness, 3. U.S. Food and Drug Administration. InFUSE™ Bone Graft/LT-CAGE™
Lumbar Tapered Fusion Device—P000058. Silver Spring, MD: U.S. Food
safety, or cancer risk between licensed and unlicensed and Drug Administration; 2012. Accessed at www.fda.gov/MedicalDevices
preparations. /ProductsandMedicalProcedures/DeviceApprovalsandClearances/Recently
The use of rhBMP-2 in spinal fusion surgery increases -ApprovedDevices/ucm083423.htm on 23 January 2012.
the likelihood of successful fusion at up to 24 months, but 4. Ong KL, Villarraga ML, Lau E, Carreon LY, Kurtz SM, Glassman SD.
Off-label use of bone morphogenetic proteins in the United States using admin-
this does not seem to translate into a clinically significant istrative data. Spine (Phila Pa 1976). 2010;35:1794-800. [PMID: 20700081]
reduction in pain. The small improvements in fusion and 5. Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant
in the level of pain reduction, which manifest after 6 human bone morphogenetic protein-2 trials in spinal surgery: emerging safety
concerns and lessons learned. Spine J. 2011;11:471-91. [PMID: 21729796]
months, also seem to come at the expense of more frequent
5a. Brown JV, Heirs MK, Higgins JP, Mannion RJ, Rogers MA, Seneschall
pain in the immediate postoperative period and, possibly, CC, et al. Systematic review and meta-analysis of the safety and efficacy of
an increased number of cancer cases. We believe that it is recombinant human bone morphogenetic protein-2 (rhBMP-2) for spinal
important that clinicians explain these findings to patients fusion. Report to the YODA Project. Available from http://medicine.yale.edu
/core/projects/yodap/index.aspx.
so that they can make informed choices about the type of 6. Stewart L, Higgins J, Mannion R, Golder S, Simmonds M, Rodgers M,
surgery they would prefer. et al. Systematic review and meta-analysis of the safety and efficacy of recombi-
From the Centre for Reviews and Dissemination, University of York, nant human bone morphogenetic protein-2 (rhBMP-2) for spinal fusion. York,
United Kingdom: Centre for Reviews and Dissemination; 2012. Accessed at
York, and MRC Biostatistics Unit and Addenbrookes Hospital, Cam-
www.crd.york.ac.uk/prospero/display_record.asp?ID⫽CRD42012001907 on 15
bridge, United Kingdom. August 2012.
7. Gornet MF, Burkus JK, Dryer RF, Peloza JH. Lumbar disc arthroplasty with
Note: Annals peer review materials (original and revised manuscripts and MAVERICK disc versus stand-alone interbody fusion: a prospective, random-
communications, including peer reviewer, editorial, statistical, and au- ized, controlled, multicenter investigational device exemption trial. Spine (Phila
thor comments) are available at www.annals.org (see Supplement 2). Pa 1976). 2011;36:E1600-11. [PMID: 21415812]
8. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al;
Disclaimer: This study used data obtained from the YODA Project that Cochrane Bias Methods Group. The Cochrane Collaboration’s tool for assessing
included 17 clinical trials of rhBMP-2 funded by Medtronic. The anal- risk of bias in randomised trials. BMJ. 2011;343:d5928. [PMID: 22008217]
ysis, interpretation, and reporting of these data were completed indepen- 9. Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The
Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised stud-
dently on an unrestricted basis, are solely the responsibility of the au-
ies in meta-analyses. Ottawa, Ontario, Canada: Ottawa Hospital Research Insti-
thors, and do not necessarily represent the official views of the YODA
tute; 2011. Accessed at www.ohri.ca/programs/clinical_epidemiology/oxford.asp
Project. on 9 October 2012.
10. Stewart LA, Clarke MJ. Practical methodology of meta-analyses (overviews)
Acknowledgment: The authors thank Leah Carreon for providing the using updated individual patient data. Cochrane Working Group. Stat Med.
IPD from the trial by Glassman and colleagues (18), Kath Wright for 1995;14:2057-79. [PMID: 8552887]
performing the literature search, and Charlotte Seneschall for assistance 11. Riley RD, Lambert PC, Abo-Zaid G. Meta-analysis of individual partici-
in extracting data for the safety evaluation. pant data: rationale, conduct, and reporting. BMJ. 2010;340:c221. [PMID:
20139215]
Grant Support: By a competitive grant from the YODA Project. 12. Higgins JP, Whitehead A, Turner RM, Omar RZ, Thompson SG. Meta-
analysis of continuous outcome data from individual patients. Stat Med. 2001;
Potential Conflicts of Interest: Disclosures can be viewed at www 20:2219-41. [PMID: 11468761]

13. Turner RM, Omar RZ, Yang M, Goldstein H, Thompson SG. A multilevel 31. Boden SD, Zdeblick TA, Sandhu HS, Heim SE. The use of rhBMP-2 in
model framework for meta-analysis of clinical trials with binary outcomes. Stat interbody fusion cages. Definitive evidence of osteoinduction in humans: a pre-
Med. 2000;19:3417-32. [PMID: 11122505] liminary report. Spine (Phila Pa 1976). 2000;25:376-81. [PMID: 10703113]
14. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. 32. Boden SD, Kang J, Sandhu H, Heller JG. Use of recombinant human bone
Stat Med. 2002;21:1539-58. [PMID: 12111919] morphogenetic protein-2 to achieve posterolateral lumbar spine fusion in hu-
15. A Pivotal Study of rhBMP-2/ACS/INTER FIX™ Device for Posterior Lum- mans: a prospective, randomized clinical pilot trial: 2002 Volvo Award in clinical
bar Interbody Fusion in Patients With Degenerative Disc Disease [clinical studies. Spine (Phila Pa 1976). 2002;27:2662-73. [PMID: 12461392]
trial]. Accessed at www.clinicaltrials.gov/ct2/results?term⫽NCT01491464& 33. Burkus JK, Gornet MF, Dickman CA, Zdeblick TA. Anterior lumbar in-
Search⫽Search on 29 April 2013. terbody fusion using rhBMP-2 with tapered interbody cages. J Spinal Disord
16. Centre for Reviews and Dissemination. Systematic review and meta-analysis Tech. 2002;15:337-49. [PMID: 12394656]
of the safety and efficacy of recombinant human bone morphogenetic protein-2 34. Dimar JR, Glassman SD, Burkus KJ, Carreon LY. Clinical outcomes and
(rhBMP-2) in spinal fusion. York, United Kingdom: Centre for Reviews fusion success at 2 years of single-level instrumented posterolateral fusions with
and Dissemination; 2012. Accessed at www.york.ac.uk/inst/crd/projects/spinal recombinant human bone morphogenetic protein-2/compression resistant matrix
_fusion.html on 15 August 2012. versus iliac crest bone graft. Spine (Phila Pa 1976). 2006;31:2534-9. [PMID:
17. Yale School of Medicine. Yale University Open Data Access (YODA) 17047540]
35. Dimar JR 2nd, Glassman SD, Burkus JK, Pryor PW, Hardacker JW,
Project. New Haven, CT: Yale School of Medicine; 2012. Accessed at http:
Carreon LY. Clinical and radiographic analysis of an optimized rhBMP-2 for-
//medicine.yale.edu/core/projects/yodap/index.aspx on 15 August 2012.
mulation as an autograft replacement in posterolateral lumbar spine arthrodesis.
18. Glassman SD, Carreon LY, Djurasovic M, Campbell MJ, Puno RM, John-
J Bone Joint Surg Am. 2009;91:1377-86. [PMID: 19487515]
son JR, et al. rhBMP-2 versus iliac crest bone graft for lumbar spine fusion: a
36. Kleeman TJ, Ahn UM, Talbot-Kleeman A. Laparoscopic anterior lumbar
randomized, controlled trial in patients over sixty years of age. Spine (Phila Pa
interbody fusion with rhBMP-2: a prospective study of clinical and radiographic
1976). 2008;33:2843-9. [PMID: 19092613] outcomes. Spine (Phila Pa 1976). 2001;26:2751-6. [PMID: 11740368]
19. Shapiro S, Rodgers RB, Sloan R, Altstadt T, Miller J. A randomized trial of 37. Burkus JK, Heim SE, Gornet MF, Zdeblick TA. Is INFUSE bone graft
instrumented posterior lumbar interbody fusion using machined cortical wedges/ superior to autograft bone? An integrated analysis of clinical trials using the LT-
local bone with or without rhBMP2 in the treatment of degenerative lumbar CAGE lumbar tapered fusion device. J Spinal Disord Tech. 2003;16:113-22.
spondylolisthesis with stenosis. Neurosurgery. 2005;57:398. [PMID: 12679664]
20. INFUSE®Bone Graft in Transforaminal Lumbar Interbody Fusion for De- 38. Burkus JK, Gornet MF, Schuler TC, Kleeman TJ, Zdeblick TA. Six-year
generative Disease of Lumbosacral Spine (TLIF) [clinical trial]. Accessed at http: outcomes of anterior lumbar interbody arthrodesis with use of interbody fusion
//clinicaltrials.gov/ct2/show/NCT01415908?term⫽rhBMP-2&rank⫽24 on 9 cages and recombinant human bone morphogenetic protein-2. J Bone Joint Surg
August 2012. Am. 2009;91:1181-9. [PMID: 19411467]
21. Pilot Study of rhBMP-2/ACS/LT-CAGE® for Anterior Lumbar Interbody 39. Comer GC, Smith MW, Hurwitz EL, Mitsunaga KA, Kessler R, Carragee
Fusion in Patients With Degenerative Disc Disease [clinical trial]. Accessed at EJ. Retrograde ejaculation after anterior lumbar interbody fusion with and with-
http://clinicaltrials.gov/ct2/results?term⫽NCT01491373&Search⫽Search on 29 out bone morphogenetic protein-2 augmentation: a 10-year cohort controlled
April 2013. study. Spine J. 2012;12:881-90. [PMID: 23098617]
22. Pivotal Study of rhBMP-2/ACS/LT-CAGE® Device for Anterior Lumbar 40. Taylor SJ, Taylor AE, Foy MA, Fogg AJ. Responsiveness of common out-
Interbody Fusion in Patients With Symptomatic Degenerative Disc Disease [clin- come measures for patients with low back pain. Spine (Phila Pa 1976). 1999;24:
ical trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT01491386& 1805-12. [PMID: 10488511]
Search⫽Search on 29 April 2013. 41. Ware JE, Kosinski M, Keller SD. SF-36 Physical and Mental Health Sum-
23. A Pilot Study of rhBMP-2/ACS/Allograft Bone Dowel for Anterior Lumbar mary Scales: A User’s Manual. Boston: The Health Institute; 1994.
Interbody Fusion in Patients With Symptomatic Degenerative Disc Disease [clin- 42. Thawani JP, Wang AC, Than KD, Lin CY, La Marca F, Park P. Bone
ical trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT01494428& morphogenetic proteins and cancer: review of the literature. Neurosurgery. 2010;
Search⫽Search on 29 April 2013. 66:233-46. [PMID: 20042986]
24. Pivotal Study of rhBMP-2/ACS/Allograft Bone Dowel for Anterior Lumbar 43. Loke YK, Price D, Herxheimer A; Cochrane Adverse Effects Methods
Interbody Fusion in Patients With Symptomatic Degenerative Disc Disease [clin- Group. Systematic reviews of adverse effects: framework for a structured ap-
ical trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT01494493& proach. BMC Med Res Methodol. 2007;7:32. [PMID: 17615054]
Search⫽Search on 29 April 2013. 44. Golder S, Loke Y. Search strategies to identify information on adverse effects:
25. INFUSE™ BONE GRAFT/CORNERSTONE-SR™ Allograft Ring/ a systematic review. J Med Libr Assoc. 2009;97:84-92. [PMID: 19404498]
ATLANTIS™ Anterior Cervical Plate- Pilot Study [clinical trial]. Accessed at 45. Centre for Reviews and Dissemination. Systematic reviews: CRD’s guidance
http://clinicaltrials.gov/ct2/results?term⫽NCT01491399&Search⫽Search on 29 for undertaking reviews in health care. York, United Kingdom: Centre for Re-
views and Dissemination; 2009.
April 2013.
46. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency
26. INFUSE® Bone Graft and MASTERGRAFT® Granules With CD
in meta-analyses. BMJ. 2003;327:557-60. [PMID: 12958120]
HORIZON® for Posterolateral Lumbar Fusion in Patients With Degenerative
47. Simmonds MC, Tierney J, Bowden J, Higgins JPT. Meta-analysis of time-
Disc Disease - Pilot Study [clinical trial]. Accessed at http://clinicaltrials.gov/ct2
to-event data: a comparison of two-stage methods. Res Synth Methods. 2011;2:
/results?term⫽NCT01491542&Search⫽Search on 29 April 2013.
139-49.
27. A Pilot Study of rhBMP-2/ACS With the INTERFIX™ Device for the 48. Fisher DJ, Copas AJ, Tierney JF, Parmar MK. A critical review of methods
Anterior Lumbar Interbody Fusion in Patients With Degenerative Disc Dis- for the assessment of patient-level interactions in individual participant data meta-
ease [clinical trial]. Accessed at http://clinicaltrials.gov/ct2/results?term analysis of randomized trials, and guidance for practitioners. J Clin Epidemiol.
⫽NCT01491451&Search⫽Search on 29 April 2013. 2011;64:949-67. [PMID: 21411280]
28. Pilot Study of rhBMP/BCP With or Without the TSRH® Spinal System for 49. Thompson SG, Higgins JP. Treating individuals 4: can meta-analysis help
Posterolateral Lumbar Fusion in Patients With Degenerative Disc Disease [clin- target interventions at individuals most likely to benefit? Lancet. 2005;365:341-6.
ical trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT01494441& [PMID: 15664231]
Search⫽Search on 29 April 2013. 50. Simmonds MC, Higgins JP, Stewart LA, Tierney JF, Clarke MJ, Thomp-
29. Clinical Study of rhBMP-2/BCP With the TSRH® Spinal System for Pos- son SG. Meta-analysis of individual patient data from randomized trials: a review
terolateral Lumbar Fusion in Patients With Degenerative Disc Disease [clinical of methods used in practice. Clin Trials. 2005;2:209-17. [PMID: 16279144]
trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT01494454& 51. Whitehead A, Omar RZ, Higgins JP, Savaluny E, Turner RM, Thompson
Search⫽Search on 29 April 2013. SG. Meta-analysis of ordinal outcomes using individual patient data. Stat Med.
30. rhBMP-2/CRM/CD HORIZON® Spinal System Pivotal Study [clinical 2001;20:2243-60. [PMID: 11468762]
trial]. Accessed at http://clinicaltrials.gov/ct2/results?term⫽NCT00707265& 52. Bowden J, Tierney JF, Simmonds M, Copas AJ, Higgins JPT. Individual
Search⫽Search on 29 April 2013. patient data meta-analysis of time-to-event outcomes: one-stage versus two-stage

approaches for estimating the hazard ratio under a random effects model. Res the ATLANTIS anterior cervical plate. Spine (Phila Pa 1976). 2003;28:1219-24.
Synth Methods. 2011;2:150-62. [PMID: 12811263]
53. Riley RD, Simmonds MC, Look MP. Evidence synthesis combining indi- 73. Bae HW, Stambough J, Glassman SD, Burkus JK. Level-1 data comparing
vidual patient data and aggregate data: a systematic review identified current rhBMP-2/ACS combined with an osteoconductive bulking agent with iliac crest
practice and possible methods. J Clin Epidemiol. 2007;60:431-9. [PMID: bone graft in posterolateral lumbar fusion. Spine J. 2007;7:9S.
17419953] 74. Dawson E, Bae HW, Burkus JK, Stambough JL, Glassman SD. Recombi-
54. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, nant human bone morphogenetic protein-2 on an absorbable collagen sponge
et al. The PRISMA statement for reporting systematic reviews and meta-analyses with an osteoconductive bulking agent in posterolateral arthrodesis with instru-
of studies that evaluate healthcare interventions: explanation and elaboration. mentation. A prospective randomized trial. J Bone Joint Surg Am. 2009;91:
BMJ. 2009;339:b2700. [PMID: 19622552] 1604-13. [PMID: 19571082]
55. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 75. Assiri I, du PS, Hurlbert J, Hu R, Salo P, Whittaker T. A prospective
1986;7:177-88. [PMID: 3802833] randomized clinical study comparing instrumented lumbar fusion rates of recom-
56. Dickman CA, Gornet MF, Burkus JK, Zdeblick TA. A prospective random- binant human bone morphogenetic protein-2 (rhBMP-2) with autogenous iliac
ized lumbar fusion study comparing rhBMP-2 with autograft for cage ALIF. crest bone graft in patients with symptomatic degenerative disc disease. Can
Neurosurgery. 2001;49:521. J Surg. 2004;47:7-8.
57. Gornet M, Burkus J, Dickman C. rhBMP-2 with tapered cages: a prospec- 76. Alexander D, Bailey S, Hurlbert RJ, Abraham E, McBroom R, Mahood J,
tive, randomised lumbar fusion study. Presented at the Annual Meeting of the et al. A long-term radiographic and clinical evaluation of a new rhBMP-2 for-
North American Spine Society, Seattle, Washington, 30 October–3 November mulation in a prospective randomized lumbar posterolateral spine fusion study.
2001. Neurosurgery. 2007;61:200-1.
58. Burkus JK, Dorchak JD, Sanders DL. Radiographic assessment of interbody 77. Abraham E, Alexander D, Bailey S, Hurlbert J, McBroom R, Mahood J,
fusion using recombinant human bone morphogenetic protein type 2. Spine et al. A long-term radiographic and clinical evaluation of a new rhBMP-2 for-
(Phila Pa 1976). 2003;28:372-7. [PMID: 12590213] mulation in a prospective randomized lumbar posterolateral spine fusion study.
59. Burkus JK. Bone morphogenetic proteins in anterior lumbar interbody fu- Can J Surg. 2008;51:S6-7.
sion: old techniques and new technologies. Invited submission from the Joint 78. Abraham EP, Hurlbert J, Alexander D, Bailey S, Fisher C. Evaluation of an
Section Meeting on Disorders of the Spine and Peripheral Nerves, March 2004. rhBMP-2 formulation in 2-level posterolateral lumbar spine arthrodesis. Spine
J Neurosurg Spine. 2004;1:254-60. [PMID: 15478362] J. 2010;10:S105.
60. Chhabra A, Bentley A, Donnell D, Greenberg D, Oliver E, Alt V. A health- 79. Glassman SD, Dimar JR, Carreon LY, Campbell MJ, Puno RM, Johnson
economic evaluation of rhBMP-2 in spine fusion surgery in Germany and UK. JR. Initial fusion rates with recombinant human bone morphogenetic protein-2/
Value Health. 2006;9:A201. compression resistant matrix and a hydroxyapatite and tricalcium phosphate/
collagen carrier in posterolateral spinal fusion. Spine (Phila Pa 1976). 2005;30:
61. Chhabra A, Nicklasson L, Greenberg D, Alt V. A cost-effectiveness analysis
1694-8. [PMID: 16094268]
of rhBMP-2 in spine fusion surgery in Sweden and Denmark. Value Health.
80. Dimar JR, Glassman SD, Burkus JK, Pryor PW, Hardacker JW, Boden
2007;10:A97.
SD. Evaluation of an recombinant human bone morphogenetic protein-2/
62. Van Genugten M, Chhabra A, Alt V. A cost-effectiveness analysis of
ceramic matrix as an iliac crest bone graft replacement in posterolateral fusion: a
rhBMP-2 in spine fusion surgery in the Netherlands. Value Health. 2008;11:
multicenter, level I clinical study. Neurosurgery. 2006;59:459.
A532-3.
81. Glassman SD, Dimar JR 3rd, Burkus K, Hardacker JW, Pryor PW, Boden
63. Alt V, Chhabra A, Franke J, Cuche M, Schnettler R, Le Huec JC. An
SD, et al. The efficacy of rhBMP-2 for posterolateral lumbar fusion in smokers.
economic analysis of using rhBMP-2 for lumbar fusion in Germany, France and
Spine (Phila Pa 1976). 2007;32:1693-8. [PMID: 17621221]
UK from a societal perspective. Eur Spine J. 2009;18:800-6. [PMID: 19301041]
82. McInnis MM, Olchanski N, Kemner JE, Goss T. Budget impact of new
64. Burkus JK, Gornet MF, Glassman SD, Slosar PJ, Rosner MK, Deckey JE,
rhbmp-2 formulation in patients undergoing posterolateral spinal fusion proce-
et al. Blood serum antibody analysis and long-term follow-up of patients treated
dures for degenerative disc disease in randomized controlled trial (RCT). Value
with recombinant human bone morphogenetic protein-2 in the lumbar spine. Health. 2010;13:A305.
Spine (Phila Pa 1976). 2011;36:2158-67. [PMID: 21325990] 83. Gornet MF, Burkus JK, Mathews HH, Dryer RF, Peloza J. Maverick™
65. Burkus JK, Transfeldt EE, Kitchel SH, Watkins R, Balderston RA. total disc replacement versus anterior lumbar interbody fusion with the
rhBMP-2 and cortical dowels in the lumbar spine: long-term outcomes. In: An- INFUSE® Bone Graft/LT-CAGE® device: a prospective, randomized, con-
nual Meeting of the North American Spine Society, Seattle, Washington, 30 trolled, multicenter IDE trial [Abstract]. Spine J. 2007;7:1S.
October–3 November 2001. Philadelphia: Elsevier Science; 2001:145-6. 84. Zdeblick T, Heim S, Kleeman T. Laparoscopic approach with tapered metal
66. Burkus JK, Transfeldt EE, Kitchel SH, Watkins RG, Balderston RA. Clin- cages: rhBMP-2 versus autogenous bone graft. Presented at the Annual Meeting
ical and radiographic outcomes of anterior lumbar interbody fusion using recom- of the North American Spine Society, Seattle, Washington, 30 October–3 No-
binant human bone morphogenetic protein-2. Spine (Phila Pa 1976). 2002;27: vember 2001.
2396-408. [PMID: 12438990] 85. Buttermann GR. Prospective nonrandomized comparison of an allograft
67. Burkus JK, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in with bone morphogenic protein versus an iliac-crest autograft in anterior cervical
combination with structural cortical allografts: clinical and radiographic outcomes discectomy and fusion. Spine J. 2008;8:426-35. [PMID: 17977799]
in anterior lumbar spinal surgery. J Bone Joint Surg Am. 2005;87:1205-12. 86. Crawford CH 3rd, Carreon LY, McGinnis MD, Campbell MJ, Glassman
[PMID: 15930528] SD. Perioperative complications of recombinant human bone morphogenetic
68. Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing protein-2 on an absorbable collagen sponge versus iliac crest bone graft for pos-
patterns associated with allograft interbody constructs in comparison with au- terior cervical arthrodesis. Spine (Phila Pa 1976). 2009;34:1390-4. [PMID:
tograft. Spine (Phila Pa 1976). 2006;31:775-81. [PMID: 16582851] 19440166]
69. Burkus J, Gornet M, Longley M. A large randomized clinical evaluation of 87. Pradhan BB, Bae HW, Dawson EG, Patel VV, Delamarter RB. Graft
rhBMP-2 versus iliac crest bone graft combined with cortical allograft in lumbar resorption with the use of bone morphogenetic protein: lessons from anterior
spine fusion. Spine J. 2004;4:S35-6. lumbar interbody fusion using femoral ring allografts and recombinant human
70. Alexander JT, Branch CL. Recombinant human bone morphogenetic pro- bone morphogenetic protein-2. Spine (Phila Pa 1976). 2006;31:E277-84.
tein 2 in a posterior lumbar interbody fusion construct: 2-year clinical and radio- [PMID: 16648733]
logic outcomes. Neurosurgery. 2002;2:48. 88. Anand N, Hamilton JF, Perri B, Miraliakbar H, Goldstein T. Cantilever
71. Haid RW Jr, Branch CL Jr, Alexander JT, Burkus JK. Posterior lumbar TLIF with structural allograft and RhBMP2 for correction and maintenance of
interbody fusion using recombinant human bone morphogenetic protein type 2 segmental sagittal lordosis: long-term clinical, radiographic, and functional out-
with cylindrical interbody cages. Spine J. 2004;4:527-38. [PMID: 15363423] come. Spine (Phila Pa 1976). 2006;31:E748-53. [PMID: 16985443]
72. Baskin DS, Ryan P, Sonntag V, Westmark R, Widmayer MA. A prospec- 89. Glassman SD, Carreon L, Djurasovic M, Campbell MJ, Puno RM, John-
tive, randomized, controlled cervical fusion study using recombinant human son JR, et al. Posterolateral lumbar spine fusion with INFUSE bone graft. Spine
bone morphogenetic protein-2 with the CORNERSTONE-SR allograft ring and J. 2007;7:44-9. [PMID: 17197332]

90. Lee KB, Taghavi CE, Hsu MS, Song KJ, Yoo JH, Keorochana G, et al. The their risk of adverse events?: a retrospective case-control study [Abstract]. Pre-
efficacy of rhBMP-2 versus autograft for posterolateral lumbar spine fusion in sented at 37th Annual Meeting of the Cervical Spine Research Society, Salt Lake
elderly patients. Eur Spine J. 2010;19:924-30. [PMID: 20041271] City, Utah, 3–5 December 2009. Spine (Phila Pa 1976). 2009.
91. Mannion RJ, Nowitzke AM, Wood MJ. Promoting fusion in minimally 102. Smucker JD, Rhee JM, Singh K, Yoon ST, Heller JG. Increased swelling
invasive lumbar interbody stabilization with low-dose bone morphogenic protein- complications associated with off-label usage of rhBMP-2 in the anterior cervical
2— but what is the cost? Spine J. 2011;11:527-33. [PMID: 20739225] spine. Spine (Phila Pa 1976). 2006;31:2813-9. [PMID: 17108835]
92. Mummaneni PV, Pan J, Haid RW, Rodts GE. Contribution of recombi- 103. Vaidya R, Carp J, Sethi A, Bartol S, Craig J, Les CM. Complications of
nant human bone morphogenetic protein-2 to the rapid creation of interbody anterior cervical discectomy and fusion using recombinant human bone morpho-
fusion when used in transforaminal lumbar interbody fusion: a preliminary re- genetic protein-2. Eur Spine J. 2007;16:1257-65. [PMID: 17387522]
port. Invited submission from the Joint Section Meeting on Disorders of the 104. Williams BJ, Smith JS, Fu KM, Hamilton DK, Polly DW Jr, Ames CP,
Spine and Peripheral Nerves, March 2004. J Neurosurg Spine. 2004;1:19-23. et al; Scoliosis Research Society Morbidity and Mortality Committee. Does
[PMID: 15291015] bone morphogenetic protein increase the incidence of perioperative complica-
93. Rihn JA, Patel R, Makda J, Hong J, Anderson DG, Vaccaro AR, et al. tions in spinal fusion? A comparison of 55,862 cases of spinal fusion with and
Complications associated with single-level transforaminal lumbar interbody fu- without bone morphogenetic protein. Spine (Phila Pa 1976). 2011;36:1685-91.
sion. Spine J. 2009;9:623-9. [PMID: 19482519] [PMID: 21897187]
94. Singh K, Smucker JD, Gill S, Boden SD. Use of recombinant human bone 105. Xu R, Bydon M, Sciubba DM, Witham TF, Wolinsky JP, Gokaslan ZL,
morphogenetic protein-2 as an adjunct in posterolateral lumbar spine fusion: et al. Safety and efficacy of rhBMP2 in posterior cervical spinal fusion for subaxial
a prospective CT-scan analysis at one and two years. J Spinal Disord Tech. degenerative spine disease: Analysis of outcomes in 204 patients. Surg Neurol Int.
2006;19:416-23. [PMID: 16891977] 2011;2:109. [PMID: 21886882]
95. Lu DC, Chou D, Rodts G, Mummaneni PV. Multilevel ACDF with and 106. Yaremchuk KL, Toma MS, Somers ML, Peterson E. Acute airway obstruc-
without BMP: a comparison of outcomes and dysphagia rates in 150 patients. tion in cervical spinal procedures with bone morphogenetic proteins. Laryngo-
J Neurosurg. 2010;113:A406. scope. 2010;120:1954-7. [PMID: 20824786]
96. Slosar PJ, Josey R, Reynolds J. Accelerating lumbar fusions by combining 107. Rowan FE, O’Malley N, Poynton A. Recombinant human bone morpho-
rhBMP-2 with allograft bone: a prospective analysis of interbody fusion rates and genic protein is safe to use in posterolateral spinal fusion surgery and is associated
clinical outcomes. Spine J. 2007;7:301-7. [PMID: 17482113] with better leg pain scores at late follow up. J Bone Joint Surg Brit. 2011;93B:
97. Vaidya R, Weir R, Sethi A, Meisterling S, Hakeos W, Wybo CD. Interbody 490-1.
fusion with allograft and rhBMP-2 leads to consistent fusion but early subsidence. 108. Gerszten PC, Tobler WD, Nasca RJ. Retrospective analysis of L5-S1 axial
J Bone Joint Surg Br. 2007;89:342-5. [PMID: 17356146] lumbar interbody fusion (AxiaLIF): a comparison with and without the use of
98. Mindea SA, Shih P, Song JK. Recombinant human bone morphogenetic recombinant human bone morphogenetic protein-2. Spine J. 2011;11:1027-32.
protein-2-induced radiculitis in elective minimally invasive transforaminal lumbar [PMID: 22122835]
interbody fusions: a series review. Spine (Phila Pa 1976). 2009;34:1480-4. 109. Taghavi CE, Lee KB, Keorochana G, Tzeng ST, Yoo JH, Wang JC. Bone
[PMID: 19525840] morphogenetic protein-2 and bone marrow aspirate with allograft as alternatives
99. Gray RJ, Rampersaud YR. Comparison of the incidence of radiculitis and to autograft in instrumented revision posterolateral lumbar spinal fusion: a min-
radiographic adverse event following minimally invasive lumbar transforaminal imum two-year follow-up study. Spine (Phila Pa 1976). 2010;35:1144-50.
interbody fusions (TLIF) with and without the use of bone morphogenetic pro- [PMID: 20139805]
tein (BMP). Spine J. 2010;10:8S-9S. 110. Pimenta L, Marchi L, Oliveira L, Coutinho E. A prospective, randomized,
100. Joseph V, Rampersaud YR. Heterotopic bone formation with the use of controlled clinical and radiological study to evaluate and compare the use of
rhBMP2 in posterior minimal access interbody fusion: a CT analysis. Spine silicated calcium phosphate and rh-BMP2 in interbody lumbar spine fusion: 36
(Phila Pa 1976). 2007;32:2885-90. [PMID: 18246013] month follow-up. Spine J. 2011;11:130S.
101. Longley MC, Woodward HR, Phillips ED, Fuller JE, McClellan JW, 111. Hiremath GK, Steinmetz MP, Krishnaney AA. Is it safe to use recombi-
Burd TA, et al. Does the use of bone morphogenetic protein (rhBMP-2) in nant human bone morphogenetic protein in posterior cervical fusion? Spine
patients undergoing anterior cervical decompression and fusion (ACDF) increase (Phila Pa 1976). 2009;34:885-9. [PMID: 19531997]

Annals of Internal Medicine
Current Author Addresses: Drs. Simmonds, Higgins, and Stewart; Ms. 4 (rhbmp or rhbmp2 or rhbmp-2).af. (1435)
Brown; Ms. Heirs; and Mr. Rodgers: Centre for Reviews and Dissemi- 5 (rh-bmp or rh-bmp2 or rh-bmp-2).af. (90)
nation, University of York, York YO10 5DD, United Kingdom. 6 (infuse or amplify).af. (17204)
Dr. Mannion: Cambridge University Hospitals NHS Foundation Trust,
7 or/1-6 (38898)
Hills Road, Cambridge CB2 0QQ, United Kingdom.
8 Spine Fusion/ (13270)
Author Contributions: Conception and design: M.C. Simmonds, J.P.T. 9 (spine or spinal).af. (369003)
Higgins, R.J. Mannion, M.A. Rodgers, L.A. Stewart. 10 spondylosyndes$.af. (10)
Analysis and interpretation of the data: M.C. Simmonds, M.K. Heirs, 11 spondylodes$.af. (1709)
J.P.T. Higgins, R.J. Mannion, M.A. Rodgers, L.A. Stewart. 12 lumbar interbody arthrodesis.af. (23)
Drafting of the article: M.C. Simmonds, J.V.E. Brown, M.K. Heirs, R.J. 13 ((lumbar or cervical or posterior or anterior or lumbosa-
Mannion, M.A. Rodgers, L.A. Stewart. cral or transforminal or posterolateral) adj3 fusion$).af. (9892)
Critical revision of the article for important intellectual content: M.C.
14 fusion cage.af. (196)
Simmonds, M.K. Heirs, J.P.T. Higgins, R.J. Mannion, M.A. Rodgers,
L.A. Stewart. 15 or/8-14 (369756)
Final approval of the article: M.C. Simmonds, J.V.E. Brown, J.P.T. 16 7 and 15 (1898)
Higgins, R.J. Mannion, M.A. Rodgers, L.A. Stewart. 17 animal experiment/ (1579120)
Statistical expertise: M.C. Simmonds, J.P.T. Higgins, L.A. Stewart. 18 16 not 17 (1542)
Obtaining of funding: J.P.T. Higgins, R.J. Mannion, L.A. Stewart.
Administrative, technical, or logistic support: R.J. Mannion, M.A.
Rodgers.
Collection and assembly of data: M.C. Simmonds, J.V.E. Brown, M.K. Cochrane Central Register of Controlled Trials (CENTRAL)
Heirs, M.A. Rodgers. #1 MeSH descriptor Bone Morphogenetic Proteins, this
term only
#2 MeSH descriptor Bone Morphogenetic Protein 2, this
APPENDIX 1: SEARCH STRATEGIES AND CALL FOR
term only
EVIDENCE #3 (morphogen* NEXT (protein* or factor* or polypeptide*
Search Strategies or poly-peptide*)) or (osteogen* NEXT (protein* or factor* or
OVID MEDLINE (1948 to Present) polypeptide* or poly-peptide*)) or (osteoinduct* NEXT (pro-
1 bone morphogenetic proteins/ or bone morphogenetic tein* or factor* or polypeptide* or poly-peptide*))
protein 2/ (10872) #4 (bmp or bmp2 or bmp-2)
2 ((bone morphogen$ or osteogen$ or osteoinduct$) adj #5 (rhbmp or rhbmp2 or rhbmp-2)
(protein$ or factor$ or polypeptide$ or poly-peptide$)).af. #6 (rh-bmp or rh-bmp2 or rh-bmp-2)
(16131) #7 (infuse or amplify)
3 (bmp or bmp2 or bmp-2).af. (11384) #8 (#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7)
4 (rhbmp or rhbmp2 or rhbmp-2).af. (1262) #9 MeSH descriptor Spinal Fusion, this term only
5 (rh-bmp or rh-bmp2 or rh-bmp-2).af. (72) #10 (spine or spinal)
6 (infuse or amplify).af. (15590) #11 (spondylosyndes*)
7 or/1-6 (33622) #12 (spondylodes*)
8 Spinal Fusion/ (14504) #13 (“lumbar interbody arthrodesis”)
9 (spine or spinal).af. (322484) #14 ((lumbar or cervical or posterior or anterior or lumbo-
10 spondylosyndes$.af. (7) sacral or transforminal or posterolateral) NEAR/3 fusion*)
11 spondylodes$.af. (631) #15 (“fusion cage”)
12 lumbar interbody arthrodesis.af. (22) #16 (#9 OR #10 OR #11 OR #12 OR #13 OR #14 OR
13 ((lumbar or cervical or posterior or anterior or lumbosa- #15)
cral or transforminal or posterolateral) adj3 fusion$).af. (7931) #17 (#8 AND #16)
14 fusion cage.af. (125)
15 or/8-14 (323015)
16 7 and 15 (1312) Science Citation Index Expanded (SCI-EXPANDED) (1899 to
17 exp animals/ not humans.sh. (3715340) Present)
18 16 not 17 (827) #24 #14 not #23
#23 #22 OR #21 OR #20 OR #19 OR #18 OR #17 OR
EMBASE (1974 to Present) #16 OR #15
1 bone morphogenetic protein/ or bone morphogenetic pro- #22 Title⫽(genera or taxonomy or species or fauna or hab-
tein 2/ (12676) itat or marine or ecology)
2 ((bone morphogen$ or osteogen$ or osteoinduct$) adj #21 Title⫽(cow or cattle or bovine or livestock or swine or
(protein$ or factor$ or polypeptide$ or poly-peptide$)).af. poultry)
(19870) #20 Title⫽(rabbit or rabbits or moss or mosses or fungus or
3 (bmp or bmp2 or bmp-2).af. (11817) fungi)
W-352 18 June 2013 Annals of Internal Medicine Volume 158 • Number 12 www.annals.org

#19 Title⫽(fossil or fossils or lichen or lichens or mushroom ample, trials which have been presented at conferences but not
or mushrooms) fully reported elsewhere.
#18 Title⫽(bat or bats or bee or bees or grass or grasses or We are currently aware of 17 trials funded by the manufac-
bird or birds or avian) turer and have searched the published literature but welcome any
#17 Title⫽(bovine or sheep or fly or flies or fish or fishes or information regarding further unpublished research. If you
fisheries or horse or horses or equine) know of any such trials please contact [CRD details deleted].
#16 350,781 Title⫽(animal or animals or dog or dogs or Link to CRD project page: www.york.ac.uk/inst/crd/projects_in
canine or cat or cats or feline) _progress.cfm.
#15 Title⫽(rat or rats or mouse or mice or hamster or Link to YODA page: http://medicine.yale.edu/core/projects
hamsters) /yodap/index.aspx
#14 #13 AND #6
#13 #12 OR #11 OR #10 OR #9 OR #8 OR #7
APPENDIX 2: METHODS
#12 Topic⫽(“fusion cage”)
Data Sources Used
#11 Topic⫽((lumbar or cervical or posterior or anterior or Analyses of efficacy were restricted to RCTs in spinal fusion
lumbosacral or transforminal or posterolateral) NEAR/3 fusion*) surgery that compared rhBMP-2 with conventional ICBG sur-
#10 Topic⫽(“lumbar interbody arthrodesis”) gery. Single-group trials of rhBMP-2 or trials with comparators
#9 Topic⫽(spondylodes*) other than IBCG were excluded.
#8 Topic⫽(spondylosyndes*) One trial (rhBMP-2/BCP US pilot RCT) had 2 rhBMP-2
#7 Topic⫽(spine or spinal) groups using different fixation procedures. Only the primary
#6 #5 OR #4 OR #3 OR #2 OR #1 group was used in these analyses. The second rhBMP-2 group
#5 Topic⫽(infuse or amplify) (consisting of 11 patients) was combined with the first in sensi-
#4 Topic⫽(rh-bmp or rh-bmp2 or rh-bmp-2) tivity analyses (not presented in this paper).
#3 Topic⫽(rhbmp or rhbmp2 or rhbmp-2) We performed all analyses using the patient-level data sup-
#2 Topic⫽(bmp or bmp2 or bmp-2) plied by Medtronic and from Glassman and colleagues’ (18) trial.
#1 14,240 Topic⫽((“bone morphogen*” or osteogen* or Although intention-to-treat analyses were intended, this was not
osteoinduct*) NEAR/1 (protein* or factor* or polypeptide* or possible because many randomly assigned patients withdrew be-
“poly-peptide*”)) fore surgery; therefore, no outcome data were available for them.
Similar searches were performed in PubMED, DARE (Da-
Outcomes of Interest
tabase of Abstracts of Reviews of Effects), HTA, Biosis Previews,
Primary Outcomes
and Toxfile. Disease-specific pain and functionality:
ODI or Neck Disability Index scores for cervical spinal sur-
Call for Evidence
gery; this measures lower back (or neck) pain on a scale from 0%
Systematic Review of Bone Morphogenic Protein-2 (rhBMP-2)
(no pain) to 100% (extreme pain).
for Spinal Fusion
SF-36 PCS score, which assesses both pain and physical
The Centre for Reviews and Dissemination (CRD) is un-
function on a scale from 0% (worst) to 100% (best).
dertaking a systematic review and individual participant data
Back and leg pain; both measured on a scale from 0 (no
(IPD) meta-analysis of the comparative effectiveness of rhBMP-2
pain) to 20 (extreme pain).
(marketed as INFUSE) for spinal fusion. The review has been
Successful spinal fusion:
commissioned by the Yale University Open Data Access (YODA)
Defined radiographically by Medtronic as requiring all of
initiative as part of an overarching project to systematically re- the following: evidence of bridging trabeculae, no evidence of
view the safety and effectiveness of rhBMP-2, including reanaly- motion (⬍3-mm difference in translation, ⬍5-degree difference
sis of IPD that have been made available to Yale on an unre- in angular motion), and no evidence of radiolucency.
stricted 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 Secondary Outcomes
analyses of such data. Duration of hospital stay
CRD will undertake a comprehensive and rigorous system- Operating time
atic review and meta-analysis of individual participant data (IPD) Successful return to work or usual activity
of all relevant randomised controlled trials that have compared Use of pain medication (not prespecified)
rhBMP-2 with standard bone graft therapy.
We will include all relevant randomised controlled trials ir- Safety Outcomes
respective of whether conducted by the manufacturer or not, and We analyzed adverse events supplied by Medtronic that we
irrespective of whether published or not. We are therefore inter- considered to be potentially related to spinal surgery according to
ested in hearing from anyone who has conducted, or is aware of, the categorizations provided in the Medtronic IPD. We also con-
unpublished or partially published research in this area. For ex- sidered the following broad categories of adverse events:
www.annals.org 18 June 2013 Annals of Internal Medicine Volume 158 • Number 12 W-353

Pain: back, leg, lower extremity, arm, neck, and upper ex- Checking consistency of treatment allocation records
tremity pain Ensuring all demographic data (e.g., ages) were within plau-
Implant-related (hardware failure): displacement, breakage, sible ranges
loosening, malpositioning, and subsidence Ensuring all pain and function measurements (e.g., ODI
Infection score, SF-36 PCS) were within range, with no outliers
Neurologic events: including numbness, tingling, “pins and Checking for balanced randomization in terms of age, sex,
needles” and other patient-level factors
Cancer Checking that summary scores (e.g., ODI score) agreed with
For adverse events not generally reported by Medtronic, we the raw scores from each question from the questionnaire
searched the wider literature and analyzed the following: It was not possible to check judgments of fusion status with-
Leg pain (including radiculitis) out access to the raw radiologic data and a radiologist. However,
Heterotopic bone formation if fusion status was assessed by blinded experts, any possible re-
Dysphagia (in cervical spinal surgery) porting bias or inconsistencies in judgment should be shared
Retrograde ejaculation across both rhBMP-2 and the control groups. We have assumed
Osteolysis (further outcomes were considered in our full that the assessments of spinal fusion provided by Medtronic are
report) valid.
Quality Assessment and Risk of Bias

Patient-Level and Trial-Level Factors Affecting Effectiveness
In addition to the data-checking procedures, we also assessed
We investigated how the effectiveness of rhBMP-2 might be
each randomized trial using the Cochrane risk-of-bias tool (8).
influenced by trial- and patient-level characteristics. We investi-
The Cochrane Collaboration developed this tool to assess aspects
gated the following trial-level factors:
of trial design and conduct that have been demonstrated empir-
Spinal location of surgery (such as cervical or lumbosacral)
ically to affect estimates of treatment effect. The tool does not
Surgical approach (such as anterior lumbar fusion and pos-
address aspects of trial design that relate to applicability or
terior lumbar interbody fusion)
generalizability.
We also investigated the following patient-level factors:
The risk-of-bias tool covers 6 key areas of potential bias:
Previous spinal surgical interventions
random sequence generation, allocation concealment, blinding of
Age
Sex participants and personnel, blinding of outcome assessors, in-
Smoking status complete outcome data, and selective reporting of outcomes.
Diabetic status Each area was given a judgment of high risk of bias, unclear risk
Body mass index of bias, or low risk of bias, and a reason for each judgment was
recorded in the main data extraction spreadsheet. We used the
guidelines from the Cochrane Collaboration on what constituted
Time Points high, low, and unclear risk of bias. Additional details were based
These outcomes were analyzed at a range of different time on discussions with the clinical member of the team.
points after surgery: 6 weeks and 3, 6, 12, and 24 months. Data Two reviewers independently completed risk-of-bias assess-
on successful fusion were available only from 6 months after ments for all trials included in the efficacy analyses. Judgments
surgery onward. Data on adverse events were provided at these were made for each type of outcome reported in the trials: fusion,
times and were also available and analyzed at or immediately after patient-reported, and adverse events.
surgery (up to 4 weeks). For all trials, data were provided at all These decisions were based on the full trial protocols pro-
the described time points. vided by Medtronic and the brief protocol provided by Glassman
Data Management and Checking of IPD and colleagues.
Data were provided by Medtronic for each trial in a range of To address incomplete outcome data, we used standard
separate SAS-format data files according to the types of outcomes data-checking procedures described in the previous section to
reported. From these, we collated individual-level data on all the compare loss to follow-up in each group. We also checked that
available effectiveness outcomes at all the time points listed we had been given IPD for all outcomes that were listed in the
above. trial protocols. We requested from Medtronic any available data
Data from the single non-Medtronic trial (18) were pro- for patients who were recruited but not reported in the trials.
vided as a single Excel (Microsoft, Redmond, Washington) Some tabulated information was provided on why these patients
spreadsheet. did not undergo surgery and so were excluded from the IPD, but
For each trial for which IPD were provided, we checked the further data were unavailable. No postrandomization data had
consistency of the data, both to ensure that the data as provided been collected for these individuals. We also checked loss to
were valid and to check for errors in our data collation process. follow-up and missing outcome data (by treatment group) for
Data-checking procedures included the following: main outcomes at each analysis time point.
Checking uniqueness and consistency of patient identifica- Quality of the nonrandomized studies was assessed by using
tion numbers a domain-based approach, based on a modified version of the

Newcastle-Ottawa scale (9). However, we avoided scoring or as- (that is, they crashed). For these models, results were therefore
signing star values to the studies because doing so risks producing calculated in terms of the odds ratio, with its corresponding 95%
an uninformative summary score. Instead, we tabulated the rele- CI.
vant information for each of the following domains:
Representativeness of the exposed cohort (did the patients Two-Stage Meta-Analyses
require spinal fusion surgery? Did they form a particular sub- We combined the effect estimates from each trial (mean
group [for example, all smokers]?) difference or RR) across trials using a standard DerSimonian–
Selection of the control group (were they drawn from the Laird random-effects meta-analysis to account for potential
same source [for example, the same hospital or database]?) heterogeneity in effects across trials (55). Separate analyses were
Ascertainment of exposure (given the nature of the topic, performed at each time point. We present the results as summary
this was usually via a secure record in which the exposure to plots across all times.
treatment was clearly known) This is called a “2-stage” approach because it is performed in
Outcome of interest not present at start of the study (was 2 stages: First, we estimate effects within trials, and then we
this explicitly checked for [for example, preexisting cancer]?) combine results across trials in a meta-analysis.
Confounders or other factors used to match or controlled
for in analysis
One-Stage Meta-Analyses
Outcome assessment (independent/blind assessment, via se-
We performed 1-stage meta-analyses of the pain and func-
cure medical record, self-report, or no details)
tion outcomes as a comparison to confirm the validity of the
Follow-up (we considered adequacy of duration in relation
2-stage analyses.
to the specific adverse events reported and whether this was sim-
In a 1-stage analysis, all patient data from all trials are com-
ilar across the 2 groups)
bined simultaneously in a single regression model that is stratified
Assessments were checked, with disagreements resolved by
by trial (hence in 1 stage). For ODI, for example, we used a
discussion and/or consultation with a third researcher.
random-effects linear regression model of change in ODI from
Statistical Analysis baseline against treatment received. This model included data at
Our primary statistical method for estimating the efficacy of all time points simultaneously, but with a separate treatment
rhBMP-2 surgery for all the specified outcomes was to use stan- effect estimated at each time point. The model therefore does not
dard 2-stage meta-analytic techniques (10, 11). Individual- assume any particular model for changes in effects over time.
patient data from each trial were analyzed separately, using the However, it does assume the same amount of between-patient
same methods across trials, for all the efficacy outcomes. Separate variation at every time point and the same amount of between-
meta-analyses were performed at each of the specified time study variation in treatment effects (heterogeneity) at every time
points. point. The model was also stratified according to the trial to
We also used a “1-stage” meta-analysis approach in some which each patient belonged (12). This model had the following
analyses, primarily as a sensitivity analysis to confirm the results form:
of 2-stage analyses. These approaches are described in more detail
yijk ⫽ ␣ik ⫹ ␤ikxij ⫹ ⑀ij
below.
All main analyses used a complete-case approach in which ⑀ij⬃N(0, ␴i2)
participants with missing data were excluded from the analysis.
␤ik⬃N(␤k, ␶2)
Estimates of Effect where yijk is the change from baseline in ODI at time k for
Continuously Distributed Outcomes. For the continuously patient j in trial i. xij is a coding for treatment received (1 ⫽
distributed outcomes (ODI, SF-36 PCS, back pain, leg pain), we rhBMP-2, 0 ⫽ ICBG). ␣ik is the baseline change in score in trial
assessed efficacy in terms of the mean difference in outcome i at time k, ␤k is the mean difference between rhBMP-2 and
between the rhBMP-2 and ICBG groups. ICBG surgery at time k (that is, the treatment effect), and ␶2 is
For each patient, the change in the score from baseline to the heterogeneity in the treatment effect across trials.
the time point of interest was calculated. These were then aver- For dichotomous outcomes (successful fusion and adverse
aged for each intervention in each trial and the difference in events), a similar random-effects logistic regression model, also
means within each trial calculated, and these mean differences stratified by trial, was used, with the following form (13):
冉冊
were then combined across trials. This mean difference, along
pijk
with its associated SE, was calculated for each trial. log ⫽ ␣ik ⫹ ␤ikxij
Dichotomous Outcomes. For dichotomous outcomes (suc- 1 ⫺ pijk
cessful fusion, successful return to work, use of pain medication,
␤ik ⬃ N共␤k , ␶2 兲
cancer), we assessed efficacy in terms of the RR for the outcome
between the rhBMP-2 and ICBG groups. where pijk is the probability of successful fusion at time k for
In the 1-stage random-effects meta-analyses, RRs could not patient j in trial i, and so ␤k is the log odds ratio of event (for
be calculated because algorithms did not converge successfully example, successful fusion).

The numbers of events in any particularly adverse event Sensitivity Analyses. Some data were missing at some times
category in any trial at any time points were often small, and after surgery because some patients were lost to follow-up or were
some trials had no adverse events in a particular category. In such not available at certain times. To investigate the potential for bias
a situation, where events are rare, 2-stage meta-analyses of RR due to such missing data, we performed a multiple imputation
may be inaccurate because of the corrections required to adjust analysis for the main pain outcomes and for spinal fusion. This
for trials with no events. For all analyses of adverse events, we was achieved by performing a regression of the outcome at each
therefore used only 1-stage meta-analysis. time point against the outcome at the previous time point, strat-
ified by trial, using the complete-case data. Where a participant
Assessment of Heterogeneity
had no recorded outcome at some time point, the predicted out-
We assessed heterogeneity in all 2-stage meta-analyses by
come from the model and its SE were used to impute 10 possible
calculating the I2 statistic for heterogeneity (14).
outcomes, assuming that outcomes were normally distributed.
Exploring Clinical Heterogeneity The resulting 10 complete data sets were analyzed and the results
We investigated how trial-level and participant-level factors averaged to obtain a summary result after imputation.
influenced the effectiveness of rhBMP-2 surgery to try to explain Multiple Testing. We note that we have performed a large
any heterogeneity in the meta-analyses. number of analyses for many outcomes at multiple time points.
For trial-level factors, this was achieved by using subgroup This increases the probability of identifying a significant differ-
analysis. For example, for type of surgery, trials were divided into ence between treatment groups when there is in fact no difference
subgroups according to the surgery type (ALIF, posterior lumbar in any specific analysis to above the nominal 5% significance level
interbody fusion, or posterior lumbar fusion, anterior cervical (that is, that associated with 95% CIs).
discectomy and fusion). Random-effects meta-analyses were con- We have not performed any corrections for multiple testing
ducted within each surgery subgroup category to estimate the because the variation in numbers of outcomes considered and the
effect of rhBMP-2 for each. The results across subgroups were different numbers of time periods involved in some analyses
compared by using tests of heterogeneity (for example, to identify would mean that any correction would be arbitrary and would
any notable differences between surgery types). make comparisons between outcomes and between main analyses
For participant-level factors, we used random-effects 1-stage and sensitivity and subgroup analyses difficult.
meta-analysis methods (12). For example, to investigate the effect
of age, we took the linear regression model used for the 1-stage
analysis described for ODI above and incorporated an age param-
Statistical Software
eter and an interaction term between age and treatment. This
The IPD were supplied by Medtronic in SAS format. All
extended model had the following form:
data management, data checking, and data extraction for the IPD
yijk ⫽ ␣ik ⫹ ␤ikxij ⫹ ␮zij ⫹ ␥xijzij ⫹ ⑀ij were performed by using SAS statistical software (SAS Institute,
Cary, North Carolina).
⑀ij⬃N(0, ␴i2) Statistical analyses were performed by using the R statistical
software package (R Foundation, Vienna, Austria). In particular,
␤ik⬃N(␤k, ␶2)
2-stage meta-analyses were conducted by using the meta library
where zij is the initial age of patient j in trial i, and ␥ is the in R, and 1-stage analyses were performed by using the lme4
interaction between age and treatment, which measures the ex- library for multilevel modeling. Forest plots were produced in R
tent to which age affects the efficacy of rhBMP-2 surgery. This by using an in-house package. All other figures were produced in
model is stratified by trial. R by using the ggplot library.

Appendix Table 1. Details of Publications of Medtronic Trials Appendix Table 1—Continued
Trial Study, Year (Reference) Full Abstract Trial Study, Year (Reference) Full Abstract
Paper Only Paper Only
RCTs (rhBMP-2 vs. LT-CAGE Van Genugten et al, ⻫
ICBG) laparoscopic 2008 (62)
LT-CAGE pilot Boden et al, 2000 (31) ⻫ LT-CAGE Alt et al, 2009 (63) ⻫
LT-CAGE open Dickman, 2001 (56) ⻫ laparoscopic
LT-CAGE open Gornet et al, 2001 (57) ⻫ LT-CAGE Burkus et al, 2009 (38) ⻫
LT-CAGE open Burkus et al, 2002 (33) ⻫ laparoscopic
LT-CAGE open Burkus et al, 2003 (37) ⻫ TELAMON Not published
LT-CAGE open Burkus et al, 2003 (58) ⻫ CRM 2-level pilot Not published
LT-CAGE open Burkus, 2004 (59) ⻫ BCP Mexico pilot Not published
LT-CAGE open Chhabra et al, 2006 (60) ⻫
LT-CAGE open Chhabra et al, 2007 (61) ⻫ ACDF ⫽ anterior cervical discectomy and fusion; ALIF ⫽ anterior lumbar inter-
LT-CAGE open Van Genugten et al, ⻫ body fusion; BCP ⫽ biphasic calcium phosphate; CRM ⫽ compression-resistant
2008 (62) matrix; ICBG ⫽ iliac crest bone graft; PLIF ⫽ posterior lumbar interbody fusion;
LT-CAGE open Alt et al, 2009 (63) ⻫ RCT ⫽ randomized, controlled trial; rhBMP-2 ⫽ recombinant human bone mor-
LT-CAGE open Burkus et al, 2009 (38) ⻫ phogenetic protein-2.
* Terminated after recruitment of 3 patients.
LT-CAGE open Burkus et al, 2011 (64) ⻫
Bone dowel pilot Burkus et al, 2001 (65) ⻫
Bone dowel pilot Burkus, 2004 (59) ⻫
Bone dowel pivotal Burkus et al, 2004 (69) ⻫
Inter Fix PLIF Alexander and Branch, ⻫
2002 (70)
Inter Fix PLIF Burkus et al, 2004 (59) ⻫
Inter Fix PLIF Haid et al, 2004 (71) ⻫
Cornerstone pilot Baskin et al, 2003 (72) ⻫
Mastergraft pilot Bae et al, 2007 (73) ⻫
Mastergraft pilot Dawson et al, 2009 (74) ⻫
BCP U.S. Boden et al, 2002 (32) ⻫
BCP Canada Assiri et al, 2004 (75) ⻫
BCP Canada Alexander et al, 2007 (76) ⻫
BCP Canada Abraham et al, 2008 (77) ⻫
BCP Canada Abraham et al, 2010 (78) ⻫
AMPLIFY Glassman et al, 2005 (79) ⻫
AMPLIFY Dimar et al, 2006 (34) ⻫
AMPLIFY Glassman et al, 2007 (81) ⻫
AMPLIFY McInnis et al, 2010 (82) ⻫
Inter Fix ALIF pilot Not published
Cornerstone ACDF Not published*
pivotal
RCTs (no ICBG group)

Maverick Disc Gornet et al, 2007 (83) ⻫
pivotal
Maverick Disc Gornet et al, 2011 (7) ⻫
pivotal
Single-group rhBMP-2
trials
LT-CAGE Kleeman et al, 2001 (36) ⻫
laparoscopic
LT-CAGE Zdeblick et al, 2001 (84) ⻫
laparoscopic
LT-CAGE Burkus et al, 2003 (37) ⻫
laparoscopic
LT-CAGE Chhabra et al, 2006 (60) ⻫
laparoscopic
LT-CAGE Chhabra et al, 2007 (61) ⻫
laparoscopic

Appendix Table 2. Missing Outcome Data in Medtronic Trials
Outcome Procedure* Missing Observations, n (%)
6 wk 3 mo 6 mo 12 mo 24 mo
ODI score ICBG 18 (3) 14 (2) 29 (5) 42 (7) 79 (13)
rhBMP-2 18 (3) 10 (1) 20 (3) 33 (5) 60 (9)
SF-36 PCS score ICBG 53 (9) 53 (9) 39 (6) 59 (10) 104 (17)
rhBMP-2 91 (13) 88 (13) 54 (8) 64 (9) 100 (14)
Back pain ICBG 36 (6) 33 (5) 45 (7) 59 (10) 95 (16)
rhBMP-2 40 (6) 40 (6) 49 (7) 61 (9) 86 (12)
Leg pain ICBG 36 (6) 33 (5) 45 (7) 59 (10) 95 (16)
rhBMP-2 40 (6) 40 (6) 49 (7) 61 (9) 86 (12)
Spinal fusion ICBG – – 96 (16) 87 (14) 106 (17)
rhBMP-2 – – 98 (14) 89 (13) 112 (16)
ICBG ⫽ iliac crest bone graft; ODI ⫽ Oswestry Disability Index; PCS ⫽ Physical Component Summary; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2;
SF-36 ⫽ Short Form-36.
* 694 patients received rhBMP-2 surgery, and 608 received ICBG surgery.

Appendix Figure 1. Reduction in pain scores from preoperative results, by treatment received.

Treatment
ICBG
rhBMP-2
10 10
Mean Change in Pain Score

0 0
–10 –10
–20 –20
–30 –30
0 5 10 15 20 25 0 5 10 15 20 25
10 10
0 0
–10 –10
–20 –20
–30 –30
0 5 10 15 20 25 0 5 10 15 20 25
ICBG ⫽ iliac crest bone graft; PCS ⫽ Physical Component Summary; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2; SF-36 ⫽ Short
Form-36.

Appendix Figure 2. Results of 1-stage meta-analyses of pain and function outcomes at 6 wk and 3, 6, 12, and 24 mo after surgery.
2.5 2.5

(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

0.0 0.0
–2.5 –2.5
–5.0 –5.0
–7.5 –7.5
0 5 10 15 20 25 0 5 10 15 20 25
2.5 2.5

(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

0.0 0.0
–2.5 –2.5
–5.0 –5.0
–7.5 –7.5
0 5 10 15 20 25 0 5 10 15 20 25
ICBG ⫽ iliac crest bone graft; PCS ⫽ Physical Component Summary; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2; SF-36 ⫽ Short
Form-36.

Appendix Figure 3. Subgroup analyses for Oswestry Disability Index (top) and successful fusion (bottom), by surgical approach.
Oswestry Disability Index
Trial (Reference) MD (95% CI)
Anterior lumbar
LT-CAGE pilot (21) –10.61 (–24.28 to 3.07)
LT-CAGE open (22) 1.39 (–3.49 to 6.26)
Bone dowel pilot (23) –16.03 (–26.18 to –5.88)
Bone dowel pivotal (24) –4.10 (–14.76 to 6.56)
Inter Fix ALIF pilot (27) –8.11 (–18.55 to 2.32)
Combined –6.61 (–13.74 to 0.51)
Cervical
Cornerstone pilot (25) –16.58 (–32.15 to –1.01)
Combined –16.58 (–32.15 to –1.01)
Posterior lumbar
Inter Fix PLIF (15) –0.98 (–10.99 to 9.03)
Mastergraft pilot (26) –8.57 (–19.87 to 2.74)
BCP U.S. (28) 2.70 (–21.20 to 26.60)
BCP Canada (29) 0.23 (–5.09 to 5.55)
AMPLIFY (30) –1.25 (–4.71 to 2.20)
Glassman et al (18) –2.00 (–8.36 to 4.36)
Combined –1.34 (–3.81 to 1.14)
Overall combined –3.48 (–6.47 to –0.49)
–30 –20 –10 0 10 20
Successful Fusion
Trial (Reference) RR (95% CI)
Anterior lumbar
LT-CAGE open (22) 1.06 (0.99 to 1.15)
Bone dowel pilot (23) 1.46 (1.08 to 1.98)
Bone dowel pivotal (24) 1.20 (1.00 to 1.44)
Inter Fix ALIF pilot (27) 0.86 (0.66 to 1.12)
Combined 1.11 (0.95 to 1.29)
Posterior lumbar
Inter Fix PLIF (15) 1.19 (0.94 to 1.49)
Mastergraft pilot (26) 1.35 (1.00 to 1.84)
BCP U.S. (28) 0.70 (0.47 to 1.05)
BCP Canada (29) 1.37 (1.20 to 1.58)
AMPLIFY (30) 1.07 (1.01 to 1.14)
Combined 1.15 (0.97 to 1.36)
Overall combined 1.14 (1.03 to 1.25)
0.50 0.75 1.00 1.50 2.00
ICBG better rhBMP-2 better
ALIF ⫽ anterior lumbar interbody fusion; BCP ⫽ biphasic calcium phosphate; ICBG ⫽ iliac crest bone graft; MD ⫽ mean difference; PLIF ⫽ posterior
lumbar interbody fusion; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2; RR ⫽ relative risk.

Appendix Figure 4. MD in operating time (minutes) between rhBMP-2 and ICBG in the Medtronic trials.
Trial (Reference) MD (95% CI)
LT-CAGE pilot (21) 0.00 (0.00 to 0.00)

LT-CAGE open (22) –22.35 (–31.08 to –13.63)
Bone dowel pilot (23) –7.50 (–32.10 to 17.10)
Bone dowel pivotal (24) –29.31 (–45.35 to –13.28)
Inter Fix PLIF (15) –24.88 (–50.86 to 1.11)
Cornerstone pilot (25) –0.17 (–38.73 to 38.39)
Inter Fix ALIF pilot (27) –15.72 (–36.06 to 4.61)
BCP U.S. (28) –61.42 (–109.26 to –13.58)
BCP Canada (29) –7.00 (–28.06 to 14.07)
AMPLIFY (30) –21.28 (–31.91 to –10.66)
Mastergraft pilot (26) 0.00 (0.00 to 0.00)
Combined –20.81 (–26.32 to –15.30)
–100 –75 –50 –25 0 25 50
Details on the Medtronic trials are available in Appendix Table 1. ALIF ⫽ anterior lumbar interbody fusion; BCP ⫽ biphasic calcium phosphate;
ICBG ⫽ iliac crest bone graft; MD ⫽ mean difference; PLIF ⫽ posterior lumbar interbody fusion; rhBMP-2 ⫽ recombinant human bone morpho-
genetic protein-2.

Appendix Figure 5. Meta-analysis of use of 4 types of pain medication in Medtronic trials.
A. Muscle relaxant B. Nonnarcotic

2 2
Relative Risk for Drug Use

(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

1.5 1.5
1 1
0.75 0.75
0.5 0.5
0 5 10 15 20 25 0 5 10 15 20 25
C. Strong narcotic D. Weak narcotic

2 2

1.5 1.5
(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

1 1
0.75 0.75
0.5 0.5
0 5 10 15 20 25 0 5 10 15 20 25
ICBG ⫽ iliac crest bone graft; rhBMP-2 ⫽ recombinant human bone morphogenetic protein-2.

Appendix Figure 6. Relationship between relative risk for fusion and mean difference in pain outcomes across trials 24 mo after
surgery.

10 10
5 5
Mean Difference in
Mean Difference in
Pain Outcome
Pain Outcome
0 0
–5 –5
–10 –10
–15 –15
0.75 1.00 1.25 1.50 0.75 1.00 1.25 1.50

Relative Risk for Successful Fusion Relative Risk for Successful Fusion

10 10
5 5
Mean Difference in
Mean Difference in
Pain Outcome
Pain Outcome
0 0
–5 –5
–10 –10
–15 –15
0.75 1.00 1.25 1.50 0.75 1.00 1.25 1.50

Relative Risk for Successful Fusion Relative Risk for Successful Fusion
PCS ⫽ Physical Component Summary; SF-36 ⫽ Short Form-36.

Appendix Figure 7. Meta-analysis of all adverse events in 11 Medtronic trials.
Adverse Event ICBG, n rhBMP-2, n OR (95% CI)
Arm and neck pain 12 13 1.02 (0.46–2.27)

Arthritis/bursitis 16 24 1.42 (0.75–2.69)
Back and leg pain 207 268 1.20 (0.99–1.46)
Cardiovascular 99 93 0.86 (0.65–1.16)
Death 10 6 0.56 (0.20–1.55)
Dural injury 12 13 0.94 (0.42–2.10)
Dysphagia 2 3 0.62 (0.10–3.96)
Gastrointestinal 115 117 0.90 (0.69–1.17)
Implant 10 22 1.93 (0.91–4.11)
Infection 90 104 1.06 (0.79–1.42)
Neurologic 119 152 1.15 (0.90–1.48)
Other pain 63 97 1.39 (1.00–1.92)
Respiratory 27 34 1.11 (0.66–1.85)
Retrograde ejaculation 1 5 4.76 (0.55–40.95)
Spinal 63 70 0.99 (0.70–1.41)
Trauma 137 153 1.01 (0.80–1.29)
Urogenital 63 81 1.19 (0.85–1.66)
Vascular 4 6 1.35 (0.37–4.86)
Vertebral fracture 1 3 2.79 (0.29–26.89)
Wound complication 4 6 1.50 (0.42–5.36)
0.25 0.50 1.00 2.00 5.00 10.00
More common with ICBG More common with rhBMP-2
Details on the Medtronic trials are available in Appendix Table 1. ICBG ⫽ iliac crest bone graft; OR ⫽ odds ratio; rhBMP-2 ⫽ recombinant human
bone morphogenetic protein-2.

Appendix Figure 8. Meta-analyses of adverse events.
A. Implant-related B. Infection
Odds Ratio for Adverse Events

16 16
(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

8 8
4 4
2 2
1 1
0.5 0.5
0.125 0.125
0 5 10 15 20 25 0 5 10 15 20 25
C. Neurologic D. Pain

16 16
8 8
(rhBMP-2 vs. ICBG)
(rhBMP-2 vs. ICBG)

4 4
2 2
1 1
0.5 0.5
0.125 0.125
0 5 10 15 20 25 0 5 10 15 20 25

Appendix Table 3. Details of Published Studies Reporting on the Safety of rhBMP-2 Surgery
Study, Year (Reference) Comparator Surgical Patients, n Adverse Events Reported

Approach
rhBMP-2 Comparator
Buttermann, 2008 (85) ICBG Cervical 30 36 Wound complications
Crawford et al, 2009 (86) ICBG Cervical 41 36 Wound complications
Comer et al, 2012 (39) ICBG Anterior lumbar 239 233 Retrograde ejaculation, infection, and urine
retention
Pradhan et al, 2006 (87) ICBG Anterior lumbar 9 27 –
Anand et al, 2006 (88) ICBG Posterior lumbar 85 15 Neurologic events, hardware failure, and
leg pain/radiculitis
Glassman et al, 2007 (89) ICBG Posterior lumbar 91 35 Hardware failure
Lee et al, 2010 (90) ICBG Posterior lumbar 86 41 Infection and wound complications
Mannion et al, 2011 (91) ICBG Posterior lumbar 17 19 Heterotopic bone formation
Mummaneni et al, 2004 (92) ICBG Posterior lumbar 25 19 Neurologic events
Rihn et al, 2009 (93) ICBG Posterior lumbar 86 33 Heterotopic bone formation, osteolysis,
infection, hardware failure, wound
complication, leg pain/radiculitis, and
inflammatory cyst formation
Singh et al, 2006 (94) ICBG Posterior lumbar 41 11 Heterotopic bone formation
Lu et al, 2010 (95) Allograft Cervical 100 50 Dysphagia
Slosar et al, 2007 (96) Allograft Anterior lumbar 45 30 Heterotopic bone formation and wound
complications
Vaidya et al, 2007 (97) Allograft Various 36 41 Hardware failure and dysphagia
Mindea et al, 2009 (98) Allograft/autograft Posterior lumbar 35 8 Leg pain/radiculitis
Gray and Rampersaud, 2010 (99) Autologous bone Posterior lumbar 82 39 Heterotopic bone formation, osteolysis,
hardware failure, and leg pain/radiculitis
Joseph and Rampersaud, Autologous bone Posterior lumbar 23 10 Heterotopic bone formation
2007 (100)
Longley et al, 2009 (101) No rhBMP-2 Cervical 84 84 Dysphagia
Smucker et al, 2006 (102) No rhBMP-2 Cervical 69 165 Dysphagia
Vaidya et al, 2007 (103) No rhBMP-2 Cervical 22 24 Dysphagia
Williams et al, 2011 (104) No rhBMP-2 Cervical 4532 652 Dysphagia and wound complications
Xu et al, 2011 (105) No rhBMP-2 Cervical 48 156 Heterotopic bone formation, infection,
hardware failure, dysphagia, recurrent
laryngeal neck palsy, and wound
complications
Yaremchuk et al, 2010 (106) No rhBMP-2 Cervical 260 515 Dysphagia
Rowan et al, 2011 (107) No rhBMP-2 Posterior lumbar 64 40 Leg pain/radiculitis
Williams et al, 2011 (104) No rhBMP-2 Posterior lumbar 6049 18 267 Wound complications and inflammatory
cyst formation
Cahill et al, 2009 (2) No rhBMP-2 Various 17 626 53 026 Dysphagia and wound complications
Williams et al, 2011 (104) No rhBMP-2 Various 11 933 43 929 Wound complications
Gerszten et al, 2011 (108) Bone marrow Anterior lumbar 45 54 Heterotopic bone formation, neurologic
aspirate events, hardware failure, wound
complications, and vascular events
Taghavi et al, 2010 (109) Bone marrow Posterior lumbar 24 38 Hardware failure
aspirate
Pimenta et al, 2011 (110) Silicated calcium Posterior lumbar 15 15 Heterotopic bone formation
phosphate
Hiremath et al, 2009 (111) Mixed Cervical 16 67 Neurologic events, hardware failure,
dysphagia, recurrent neck palsy, and
wound complications

Review: Safety and Effectiveness of Recombinant Human Bone Morphogenetic Protein-2 For Spinal Fusion

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Review: Safety and Effectiveness of Recombinant Human Bone Morphogenetic Protein-2 For Spinal Fusion

Uploaded by

Copyright:

Available Formats

Annals of Internal Medicine Review

Safety and Effectiveness of Recombinant Human Bone Morphogenetic

R ecombinant human bone morphogenetic protein-2

Downloaded from https://annals.org by guest on 04/06/2019

tion (rhBMP-2 concentration, 1.5 g/L) and unlicensed Statistical Methods

Downloaded from https://annals.org by guest on 04/06/2019

Figure 1. Summary of evidence search and selection.

References identified in database search (n = 6807)

Excluded during de-duplication (n = 4073)

Screening of titles and abstracts (n = 2837)

Materials provided by Medtronic

Effectiveness analysis* Safety analyses

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Figure 2. Meta-analyses of pain outcomes at 6 wk and 3, 6, 12, and 24 mo after surgery.

A. Back pain B. Leg pain

Mean Difference in Outcome

(rhBMP-2 vs. ICBG)

Time Since Surgery, mo Time Since Surgery, mo

C. Oswestry Disability Index D. SF-36 PCS

Mean Difference in Outcome

(rhBMP-2 vs. ICBG)

Time Since Surgery, mo Time Since Surgery, mo

Downloaded from https://annals.org by guest on 04/06/2019

Figure 3. Forest plot of RR for successful fusion 24 mo after surgery.

Trial (Reference) Fused/Total, n/N RR (95% CI)

LT-CAGE open (22) 123/130 106/119 1.06 (0.99–1.15)

0.50 0.75 1.00 1.50 2.00

ICBG better rhBMP-2 better

Downloaded from https://annals.org by guest on 04/06/2019

Figure 4. Meta-analysis of adverse events at or shortly after surgery in 11 Medtronic trials.

Adverse Event ICBG, n rhBMP-2, n OR (95% CI)

Arm and neck pain 2 1 0.65 (0.06–7.24)

0.25 0.50 1.00 2.00 5.00 10.00

More common with ICBG More common with rhBMP-2

Downloaded from https://annals.org by guest on 04/06/2019

Figure 5. Forest plot of cancer incidence in the Medtronic trials.

Trial (Reference) Cancer/Total, n/N RR (95% CI)

LT-CAGE open (22) 2/143 1/136 1.92 (0.18–20.89)

0.1 1.0 5.0 10.0 20.0

rhBMP-2 better ICBG better

Downloaded from https://annals.org by guest on 04/06/2019

Author, Year (Reference) Comparator Surgery Events/Total, n/N RR (95% CI)

Heterotopic bone formation

0.25 1.00 5.00 10.00 20.00 40.00

More common with comparator More common with rhBMP-2

Downloaded from https://annals.org by guest on 04/06/2019

groups, however, so whether this increased risk is genuine .acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum⫽M12

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Quality Assessment and Risk of Bias

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

Downloaded from https://annals.org by guest on 04/06/2019

RCTs (no ICBG group)

Downloaded from https://annals.org by guest on 04/06/2019

Outcome Procedure* Missing Observations, n (%)

Downloaded from https://annals.org by guest on 04/06/2019

A. Back pain B. Leg pain