Journal of Clinical Neuroscience 16 (2009) 1417–1420

Contents lists available at ScienceDirect

Journal of Clinical Neuroscience

journal homepage: www.elsevier .com/ locate/ jocn

Clinical Study

BMP-7 (OP-1�) Safety in anterior cervical fusion surgery q

John Leach a,b, Richard G. Bittar a,c,d,e,*

a Department of Neurosurgery, The Alfred Hospital, Prahran, Victoria, Australiab Department of Neurosurgery, The Radcliffe Infirmary, Oxford, United Kingdomc Department of Neurosurgery, Royal Melbourne Hospital, Parkville, Victoria, Australiad Department of Surgery, Monash University, Prahran, Victoria, Australiae Department of Surgery, The University of Melbourne, Parkville, Victoria, Australia

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 December 2008Accepted 17 February 2009

Keywords:Adverse effectsAnterior cervical discectomy and fusionBone morphogenetic proteinsBMP7Cervical fusionSafetySpine

0967-5868/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.jocn.2009.02.012

q Dr Bittar was previously an educational consultsupport was provided for this study.

* Corresponding author. Postal address: Precision N517 St Kilda Rd, Melbourne, Victoria 3004, Australia. T3 9821 5386.

E-mail address: drbittar@precisionneurosurgery.co

Bone morphogenetic proteins (BMPs) are increasingly used in spinal fusion surgery. Previous reports ofBMP use in anterior cervical fusion have suggested high rates of complications related to soft tissue swell-ing. We evaluate the safety of using BMP-7 osteogenic protein (OP-1�; Stryker, Kalamazoo, MI, USA) in arelatively contained form and controlled dose. A prospective consecutive cohort of 131 patients under-went anterior cervical discectomy and fusion using interbody cages. In 123 of these patients, BMP-7was also used. The primary outcome measure was the presence (or otherwise) of clinical adverse eventsduring the first 30 days. The secondary outcome was the extent of radiological soft tissue swelling asmeasured on plain radiographs in the early post-operative period. There was no mortality and no reop-eration in this series; however, 2.4% of patients experienced complications of transient brachalgia (1patient), and dysphagia (2 patients). The use of BMPs in spinal fusion is discussed, and the relevant lit-erature reviewed, particularly as it relates to adverse clinical events. We concluded that BMP-7 can beused safely in anterior cervical fusion. The effect of BMP-7 on the rate and timing of fusion, as well asclinical outcome, is yet to be elucidated.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Bone morphogenetic proteins (BMPs) have been developed toenhance surgical bony fusion. They provide an adjunct, andpossibly an alternative, to autograft in patients at high risk fornon-union: those with diabetes mellitus or osteoporosis and thosewho smoke. They have also been used when harvesting of autolo-gous bone is not feasible or desirable.

Whilst the greatest experience with BMPs has been in lumbarfusion surgery, particularly revision postero-lateral lumbar fusion,their use has also been reported in the cervical spine.1–18

Adverse clinical events have been reported with the use ofBMPs in anterior cervical surgery.10–12,16,17 These include soft tis-sue swelling with subsequent dysphagia, dysphonia and airwaycompromise, as well as neurological deficits. The incidence of com-plications may be fairly high, with reports of a 23.2% to 27.5% inci-dence of post-operative swelling and other complicationsfollowing the use of BMP-2 (Infuse�; Medtronic, Minneapolis,

ll rights reserved.

ant for Stryker. No financial

eurosurgery, Suite 5, Level 4,el.: +61 3 9821 5718; fax: +61

m (R.G. Bittar).

MN, USA) in anterior cervical fusion procedures.10,12 We hypothe-sised that these risks may not be as high with the use of an alter-native bone morphogenetic protein, such as BMP-7, and if thissubstance is applied in an appropriate dose and in a relatively con-tained form.

We present a large series of patients undergoing anterior cervi-cal discectomy and fusion with BMP-7 osteogenic protein (OP-1�;Stryker, Kalamazoo MI, USA), in order to enhance fusion and avoidiliac crest bone harvesting, and provide clinical and radiologicalmeasures supporting the safety of such an approach.

2. Methods

A total of 131 patients underwent an anterior cervical fusionover 30 months. These procedures were performed by a single sur-geon (RB), and ranged from one to four levels (Table 1). Fully in-formed consent was obtained before surgery.

All patients underwent an anterior interbody fusion using poly-etheretherketone (PEEK), carbon fibre, or trabecular metal cages,with or without an anterior cervical plate. Neither autograft norallograft was used. Of the 131 patients, 123 patients had a combi-nation of BMP-7 (0.5–1 unit; 1.75–3.5 mg BMP-7 and 0.5–1.0 g col-lagen, the dose depending on the number of levels fused) andtricalcium phosphate placed within the cage; eight patients hadtricalcium phosphate alone. No BMP-7 or tricalcium phosphate

Table 1Number of levels of ACDFs performed

No. of levels No. of patients

1 942 333 34 1

ACDF = anterior cervical discectomy and fusion.

1418 J. Leach, R.G. Bittar / Journal of Clinical Neuroscience 16 (2009) 1417–1420

was placed outside the cage. Wound drains were not used in eithergroup.

Patients were evaluated prospectively for early adverse out-comes (within the first 30 days), including respiratory embarrass-ment, need for reintubation, dysphagia requiring furtherinvestigation or treatment, reoperation, readmission, andmortality.

Patients who underwent cervical spine plain radiographs 24hours to 72 hours post-operatively had these films examined retro-spectively by a blinded, disinterested third party. Prevertebral soft-tissue swelling at the level of the C6 vertebral body was comparedbetween individuals receiving BMP-7 and those who did not. Theseresults were compared to a historical post-operative anterior cervi-cal fusion cohort (who did not receive BMP-7) from theliterature.17

3. Results

There was no mortality in this series and no patients requiredre-exploration or drainage of haematoma. Three patients (2.4%)in the group receiving BMP-7 experienced post-operativecomplications:

1. Recurrent brachialgia. This patient experienced recurrent armpain 72 hours post-operatively and was re-admitted. This painsettled completely with non-steroidal anti-inflammatory medi-cations and did not recur. His long-term clinical outcome hasbeen satisfactory.

2. Sudden onset of dysphagia and dysphonia 8 dayspost-operatively. There was no objective evidence of airwaycompromise or soft tissue swelling. The patient lived in a ruralarea and was intubated for transport to a metropolitan medicalfacility. She was extubated several hours later. A CT scan of herneck did not reveal soft tissue swelling or haematoma in excessof what would be expected post-operatively, and the individualwas reviewed by an otolaryngologist who found no evidence ofswelling or laryngeal nerve dysfunction. These symptomsresolved completely within 24 hours, and corticosteroids werenot required. It was concluded that there was no organic basisfor her presentation, and that there was a significant psycholog-ical contribution to her symptoms.

3. Moderate dysphagia which persisted beyond 3 months post-operatively, and resolved completely by 12 months. Due tothe progressively improving nature of this symptom, furtherinvestigation was not undertaken. No specific treatment wasrequired.

Prevertebral soft tissue measurements at C6 24 hours to 72hours post-operatively were carried out on 20 patients receivingBMP-7 and in seven non-BMP patients. The mean prevertebralsoft tissue measurement in the BMP-7 group was 20.9 mm(16–27 mm). This compared with 18.7 mm (15–25 mm) in thenon-BMP-7 group (p = 0.05), and 18 mm in the historical controlgroup.

4. Discussion

4.1. Basic science

BMPs are low molecular weight glycoproteins which belong tothe transforming growth factor-beta (TGF-ß) superfamily. BMPsare growth factors which act on the osteogenic protein receptorsignalling pathways. They promote both osteoblast proliferationand the differentiation of pluripotent mesenchymal cells intoosteoblasts.

BMPs have been isolated and purified from growth factors thatare present physiologically and that are active in post-natal bonegrowth. Animal studies have demonstrated the ability of BMPs toinduce bone formation and heal skeletal defects, including in thepresence of nicotine.18,19 BMPs are soluble and, in order to be effec-tive for local bone fusion, they must be placed on an appropriatecarrier that will release the compound in a controlled fashion.

4.2. BMPs and spinal fusion

The initial popular use of BMPs in surgery was as an alternativeto allograft in non-union of fractured long bones such as the tibiaand humerus.18,19 Spinal surgeons have been seeking an effectivealternative to autograft that eliminates donor-site morbidity andreliably leads to fusion even in high-risk patients. Although spinalimplants have evolved and improved, the result of spinal fusionsurgery is usually dependent on the occurrence of bony fusion.Without solid fusion, there is a high risk of hardware fatigue andbreakage.20 BMPs have been investigated in the thoracolumbarspine and, more recently, the cervical spine. Currently, BMP-7(OP-1�) is approved by the United States Food and Drug Adminis-tration (FDA) for use in revision posterolateral lumbar fusions andBMP-2 is FDA approved for use in anterior lumbar interbody fu-sion. Other uses of BMPs in spine surgery are currently off-label.

4.2.1. BMP-2 and anterior lumbar interbody fusionBurkus evaluated the use of BMP-2 for anterior lumbar inter-

body fusion using two tapered threaded cages, randomising 279patients to receive either BMP-2 (4.2–8.4 mg) or autogenous iliaccrest bone graft.21 Clinical outcomes in the two groups were com-parable, there was a non-significant higher rate of fusion in theBMP-2 group (94.5% versus 88.7% at 24 months), and BMP-2 usewas associated with reduced operative time and blood loss. Theauthors concluded that BMP-2 eliminated the need for iliac crestautograft.

4.2.2. BMPs and posterolateral lumbar fusionNumerous clinical studies have reported the use of BMPs in

both non-instrumented and instrumented posterolateral lumbarfusion.1–6,22–24 Whilst the use of BMPs has been shown to reduceoperative time and blood loss, and BMPs have been shown to besafe, fusion rates have not been shown to be significantly betterthan control groups using iliac crest autograft.

Vacarro et al. studied the use of BMP-7 (OP-1�) innon-instrumented lumbar posterolateral fusion for degenerativespondylolisthesis in patients of standard risk for non-union.22

Thirty-six patients were randomised to receive either BMP-7(3.5 mg per side) or iliac crest autograft. There was a non-signifi-cant difference in fusion rates (55% vs. 40%) as assessed by plainfilms, but a trend to increased fusion rates in the OP-1 group wasobserved. The patient follow-up was incomplete, and the value ofthe study was further limited by the use of plain radiographs ratherthan fine-cut CT scans to assess radiographic fusion.

Another study by the same author evaluated BMP-7 as anadjunct to iliac crest autograft in 12 patients undergoing

J. Leach, R.G. Bittar / Journal of Clinical Neuroscience 16 (2009) 1417–1420 1419

non-instrumented posterolateral lumbar fusions.23 Only nine ofthe 12 patients had 24-month follow-up. The fusion rate was only50% (similar to historical fusion rates without BMPs) with 70%showing some bridging bone. No toxicity was reported.

Johnsson et al. performed radiostereometric assessment of ver-tebral movements on 20 patients who had been randomised to re-ceive either BMP-7 or iliac crest autograft for non-instrumentedL5/S1 lumbar fusion, with no significant difference.1

Kanayama et al. randomised 19 patients undergoing lumbardecompression and pedicle screw instrumentation for spinal ste-nosis and degenerative spondylolisthesis to receive either BMP-7(3.5 mg per side) or iliac crest autograft mixed with 5 g of hydroxy-apatite/tricalcium phosphate.2 Radiological fusion judged by plainradiography and CT scan was observed in seven of nine in the BMP-7 group and nine of 10 of the control group.

Dimar et al. compared BMP-2 to iliac crest bone graft in 98 pa-tients undergoing single-level instrumented posterior lumbar fu-sions.3 The dose of BMP-2 used was higher than usuallyrecommended (20 mg per side vs. 6 mg per side). The fusion rateas judged by post-operative CT scans was higher in the BMP-2group (88%) than the autograft group (73%). In addition, reducedoperative time and reduced blood loss was observed in the BMP-2 group.

Haid et al. reported a group of 67 patients undergoing single le-vel posterior lumbar interbody fusion with stand-alone titaniumcages who were randomised to receive either BMP-2 and collagenor autologous bone.4 Although there was a trend to higher fusionrate in the BMP-2 group (92.3% vs. 77.8%), this difference was notsignificant.

Mummaneni et al. and Villavicencio et al., in separate studies,reported the use of BMP-2 in transforaminal lumbar interbody fu-sions.5,6 In the first study, there was no difference in the fusion ratebetween those receiving BMP-2 and iliac crest autograft. In the sec-ond, a 100% fusion rate was reported with BMP-2 but there was nocontrol group.

4.2.3. Other posterior spinal fusion studiesGovender et al. reported a prospective series of nine patients

undergoing spinal fusion who were at high risk of non-union.7

There were four posterior lumbar fusions and five posterior cervi-cal (craniocervical and C1/2) fusions including three intraduralprocedures. The 3.5 mg of BMP-7 (OP-1�) with collagen carrierand 2.5 mL of saline was mixed with autologous bone harvested lo-cally or from the iliac crest. There were no adverse clinical eventsreported. The follow-up period was short (mean 5.22 months,range 1–15 months) and assessment of fusion was incomplete.

Furlan et al. reported a series of 30 patients at high risk for non-union who underwent 16 lumbar and 14 cervical procedures.8

BMP-7 (OP-1�) with a bovine collagen carrier was reconstitutedin 2.5 mL of saline to form putty, which was mixed with localautologous bone before being implanted in the posterolateral gut-ter. There was one patient with heterotropic soft tissue calcifica-tion but no other adverse events reported. Fusion rate in thishigh risk group was 80%, although this did not translate into signif-icantly improved Oswestry Disability Index scores at 1 year norimproved physical health scores at 2 years.

In a retrospect, Glassman found that, for patients who weresmokers, BMP-2 use in single-level instrumented posterior lumbarfusion was associated with higher rates of fusion than autologousbone graft (94.1% vs. 76.2%).9

4.2.4. BMPs and anterior cervical fusionThe use of BMPs in anterior cervical spine surgery does appear

to confer an extremely high rate of fusion, notwithstanding thatthe rate of fusion with these surgical procedures is usually high(84.9–97.1% for single-level disc disease).10–12,24

Baskin et al. reported a randomised controlled anterior cervicalfusion study using BMP-2.10 Thirty-three patients with degenera-tive disc disease were randomised to receive either a fibular allo-graft with BMP-2 (0.4 mL of a 1.5 mg/mL solution on a collagensponge) or iliac crest autograft, with both groups receiving an ante-rior cervical plate. There was 100% fusion in both control and BMP-2 groups at 12 months. There were two instances of anterior boneformation at adjacent segments in the BMP-2 group but also onepatient in the control group.

Tumialan and Rodts reported a retrospective series of 200 pa-tients who underwent ACDF using PEEK cages filled with BMP-2and collagen sponge.11 With a mean follow-up of 16.7 monthsthere was 100% radiographic fusion.

4.2.5. BMPs and reported adverse eventsThe main issue surrounding the use of BMPs in anterior cervical

spine surgery to date has been its safety. Several studies have raisedserious concerns about potential side effects, some of which may belife-threatening. Shields et al. reported a 23% complication rate in aretrospective series of 151 patients undergoing anterior cervicaldiscectomy or vertebrectomy and fusion.12 In that study, up to2.1 mg per level of BMP-2 soaked in a sponge was used within thecage. A higher dose was used in the vertebrectomy patients and addi-tional BMP-2 was frequently placed lateral and anterior to the cageor graft. The BMP-2 dose was more than three times that used inthe Baskin et al. study.10 Reported complications included haema-toma (10%) and prolonged hospital stay for dysphagia or dyspnoea(8.6%). In Tumialan’s series11, there was a significant dysphagia rateof 7% and a 2% re-operation rate for haematoma or seroma. The rateof complications in our series was not only of lower incidence butalso less severe in degree, when compared to previous studies.

Bone resorption is another potentially important issue. BMPspromote osteoblast differentiation and proliferation but they alsohave a stimulatory effect on osteoclasts. This osteoclast effect oc-curs early and, if high doses BMPs are used, bone resorption ratherthan bone growth may occur. McClellan evaluated 26 of 198 pa-tients who had undergone a transforaminal lumbar interbody fu-sion (TLIF) with BMP-2.13 There was a 70% incidence of boneresorption, including a 30% incidence of severe bone resorption de-fined as an area of resorption of >75% of the graft area or >1 cm by1 cm. The dose of BMP-2 was not controlled in their series. Therelationship of osteolysis to fusion rates was not reported. Vaidyaet al. reported significant graft subsidence when BMP-2 was usedwith allograft in anterior and lumbar interbody fusions.14 In cervi-cal interbody fusions using BMP-2, cage migration, loosening, andsubsidence were reported, in association with high rates offusion.25

Joseph and Rampersaud reported increased heterotropic boneformation in patients undergoing posterior minimal access inter-body fusion with BMP-2 (20.8%) compared to those withoutBMP-2 (8.3%).15 Despite the bone growth being in the canal or fora-men it was clinically silent in this series. The fusion rate was signif-icantly higher in the BMP-2 group (91.3% vs. 50%). The BMP-2 dosewas controlled at 4.2 mg/level and was placed only anterior to orwithin the cage.

An increased rate of post-operative radiculitis has been re-ported following the use of BMP-2 in patients undergoing TLIF.Sanfilippo et al. reported a 23% incidence of persistent radiculitisin patients undergoing TLIF with BMP-2 vs. 3% in those undergoingthe same procedure without BMP-2.26 In their study, BMP-2 wasplaced both within and around the cage. It is hypothesised thatplacement of BMP-2 around the cage and near the nerve rootmay have led to inflammatory effects on soft tissue around the rootwith subsequent radiculitis.

In July 2008, the FDA issued a public health notificationregarding complications associated with BMP use in cervical spine

1420 J. Leach, R.G. Bittar / Journal of Clinical Neuroscience 16 (2009) 1417–1420

fusion.16 At least 38 reports of swelling of soft tissue structures inthe neck had been reported including airway compromise requir-ing intubation or tracheostomy. The complications generally oc-curred between day 2 and 14, and the overwhelming majoritywere associated with the use of BMP-2 rather than BMP-7. TheFDA recommendation was that BMPs were either not used in ante-rior cervical fusion or used in the context of a clinical trial.

4.2.6. Safe use of BMP in anterior cervical surgeryOur series demonstrates that BMP-7 can be used safely in ante-

rior cervical fusion surgery. Although there was a slight increase inpost-operative prevertebral swelling on radiological evaluation,this effect was not significant clinically. There was no incidenceof re-operation. In contrast to earlier studies, the dose of BMP usedwas controlled (1.75–3.5 mg) and it was used in a contained fash-ion within the cage rather than being placed around the cagewhere it may more easily spread to the soft tissues anterior tothe surgical site.

5. Conclusion

This study demonstrates that BMP-7 can be used safely in ante-rior cervical fusion surgery. A slight increase in post-operative pre-vertebral swelling in such patients is not clinically significant. Theeffect of BMP-7 on the rate and timing of fusion, as well as clinicaloutcome, is yet to be elucidated.

References

1. Johnsson R, Stromqvist B, Aspenberg P. Randomised radiostereometric studycomparing OP-1 (BMP-7) and autograft bone in human non-instrumentedposterolateral lumbar fusion. Spine 2002;27:2654–61.

2. Kanayama M, Hashimoto T, Shigenobu K, et al. A prospective randomised studyof posterolateral lumbar fusion using osteogenic protein-1 (OP-1) versus localautograft with ceramic bone substitute. Spine 2006;31:1067–74.

3. Dimar JR, Glassman SD, Burkus KJ, et al. Clinical outcomes and fusion success at2 years of single-level instrumented posterolateral fusions with recombinanthuman bone morphogenetic protein-2/compression resistant matrix versusiliac crest bone graft. Spine 2006;31:2534–9.

4. Haid RW, Branch CL, Alexander JT, et al. Posterior lumbar interbody fusion usingrecombinant human bone morphogenetic protein type 2 with cylindricalinterbody cages. Spine J 2004;4:527–38.

5. Mummaneni PV, Pan J, Haid RW, et al. Contribution of recombinant humanbone morphogenetic protein-2 to the rapid creation of interbody fusion whenused in transforaminal lumbar interbody fusion: a preliminary report. JNeurosurg Spine 2004;1:19–23.

6. Villavicencio AT, Burneikiene S, Nelson EL, et al. Safety of transforaminallumbar interbody fusion and recombinant human bone morphogenetic protein-2. J Neurosurg Spine 2005;3:436–43.

7. Govender PV, Rampersaud YR, Rickards L, et al. Use of osteogenic protein-1 inspinal fusion: literature review and preliminary results in a prospective seriesof high risk cases. Neurosurg Focus 2002;13:1–6.

8. Furlan JC, Perrin RG, Govender PV, et al. Use of osteogenic protein-1 in patientsat high risk for spinal pseudarthrosis: a prospective cohort study assessingsafety, health-related quality of life, and radiographic fusion. J Neurosurg Spine2007;7:486–95.

9. Glassman SD, Dimar JR, Burkus K, et al. The efficacy of rhBMP-2 forposterolateral lumbar fusion in smokers. Spine 2007;32:1693–8.

10. Baskin DS, Ryan P, Sonntag V. A prospective randomised controlled cervicalfusion study using recombinant human bone morphogenetic protein-2 withthe Cornerstone-SR allograft ring and the Atlantis anterior cervical plate. Spine2003;28:1219–25.

11. Tumialan LM, Rodts GE. Adverse swelling associated with use of rh-BMP-2 inanterior cervical discectomy and fusion. Spine J 2007;7:235–9.

12. Shields LBE, Raque GH, Glassman SD, et al. Adverse effects associated withhigh-dose recombinant human bone morphogenetic protein-2 use in anteriorcervical fusion. Spine 2006;31:542–7.

13. McClellan JW. Vertebral bone resorption after transforaminal lumbar interbodyfusion with bone morphogenetic protein (rhBMP-2). J Spinal Disord Tech 2006;19:483–6.

14. Vaidya R, Weir R, Sethi A. Interbody fusion with allograft and rhBMP-2 leads toconsistent fusion but early subsidence. J Bone Joint Surg 2007;89:342–5.

15. Joseph V, Rampersaud YR. Heterotopic bone formation with the use of rhBMP2in posterior minimal access interbody fusion. Spine 2007;32:2885–90.

16. United States Food and Drug Administration. FDA Public Health Notification:life-threatening complications associated with recombinant human bonemorphogenetic protein in cervical spine fusion. http://www.fda.gov/cdrh/safety/070108-rhbmp.html.

17. Suk KS, Kim KT, Lee SH, et al. Prevertebral soft tissue swelling after anteriorcervical discectomy and fusion with plate fixation. Int Orthop 2006;30:290–4.

18. Gautschi O, Frey S, Zellweger Z. Bone morphogenetic proteins in clinicalapplications. ANZ J Surg 2007;77:626–31.

19. Lane J. Bone morphogenic protein science and studies. J Orthop Trauma2005;19:S17–22.

20. Steffee AD, Brantigan JW. The variable screw placement spinal fixation system:report of a prospective study of 250 patients enrolled in FDA clinical trials.Spine 1993;18:1160–72.

21. Burkus JK, Gornet MF, Dickman CA, et al. Anterior lumbar interbody fusionusing rhBMP-2 with tapered interbody cages. J Spinal Disord Tech2002;15:337–49.

22. Vaccaro AR, Anderson DG, Patel T, et al. Comparison of OP-1 putty (rhBMP-7) toiliac crest autograft for posterolateral lumbar arthrodesis. Spine2005;30:2709–16.

23. Vaccaro AR, Patel T, Fischgrund J, et al. A 2-year follow-up pilot studyevaluating the safety and efficacy of OP-1 putty (rhBMP-7) as an adjunct to iliaccrest autograft in posterolateral lumbar fusions. Eur Spine J 2005;14:623–9.

24. Fraser JF, Hartl R. Anterior approaches to fusion of the cervical spine: ametaanalysis of fusion rates. J Neurosurg Spine 2007;6:298–303.

25. Vaidya R, Sethi A, Bartol S, et al. Complications in the use of rhBMP-2 in PEEKcages for interbody spinal fusions. J Spinal Disord Tech 2008;21:557–62.

26. Sanfilippo JA, Lee JY, Rihn J, et al. BMP-2 causes increased postoperativeradiculitis following TLIF. Spine J 2007;7:5S–6S.