Thank you for the opportunity to respond to the letter regarding our article. RhBMP-2 on an absorbable collagen sponge carrier (rhBMP-2/ACS) has been used successfully to induce spinal fusion. With expanded worldwide clinical use, rare instances of vertebral endplate or vertebral body resorption have been observed after lumbar interbody fusion with various interbody constructs (1,2). This resorptive phenomenon appears as decreased mineral density on computed tomography scans approximately 3 to 6 months after surgery. These resorptive zones have been transient, and fusion subsequently occurred. Animal studies have demonstrated that these localized resorptive zones are a transient accelerated osteoclastic response, induced by rhBMP-2/ACS in a concentration-dose dependent manner, that are followed by osteoblastic activity and de novo bone formation.

Drs. Smoljanovic, Bojanic, and Delimar use the term “substantial shortcomings” in describing the referenced publication, “Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2” (3). They inappropriately speculate that there was some calculated reason for omission of certain radiographic findings reported in the pilot, pivotal, and healing patterns articles (3-5). Dr. Smoljanovic and associates should be aware of several facts. The pilot and pivotal manuscripts reported the prospective Food and Drug Administration (FDA) Investigational Device Exemption (IDE) criteria for assessing interbody fusion, whereas the article assessing healing patterns was based on a retrospective review of the radiographic data obtained in these two studies. The prospective criteria to assess fusion were put in place prior to starting the study and independent and blinded radiologists assessed the films for determining fusion. The radiographic analyses were done outside the clinical research sites. The presence of transient and self-limited zones of active bone remodeling did not represent a failure of fusion. These small, localized, and self-limited zones represent an accelerated healing pattern. These isolated radiographic findings identified only on computed tomography scans in less than 18% of the patients studied had no impact on clinical outcomes or rates of fusion.

Smoljanovic et al. use an unfortunate and equally inaccurate term, “large osteolytic cysts,” which is a mischaracterization of what occurred radiographically and what was described in one of our references (3,6). The localized regions of osteopenia do not represent osteolysis but rather an initial accelerated osteoclastic response that leads to new bone formation. These accelerated bony remodeling zones were small, solitary, localized regions of bone remodeling within the vertebral bodies adjacent to the central opening in one of the two allograft bone dowels. They were not seen on plain radiographs; they were observed only on CT scans. Furthermore, the zones were transient, appeared between 3 and 12 months after surgery, and resolved by 24 months. The zones represent accelerated healing and have no impact on clinical or radiographic fusion outcomes.

Osteolysis is not occurring within these remodeling zones. Osteolysis is an inaccurate and inappropriate term because 1) osteolytic lesions do not resolve within a short period, 2) osteolytic lesions do not actively remodel, and 3) osteolytic lesions do not generate significant new bone without further intervention. The second question raised by Smoljanovic and associates revolves around a misprint within the material and methods section of the pilot study. There was no change in the amount of rhBMP-2/ACS between the pilot and pivotal phases of this trial (3,4). In both studies, 8 to 12 mg of rhBMP-2 was used and was distributed within and between the allograft dowels. The volume of rhBMP-2/ACS used for each bone dowel was determined by the outer diameter of the construct. Importantly, the internal diameter of each dowel varied and depended on the diameter of the donor’s medullary canal. This variability may have led to unanticipated overfilling of some bone dowels and a locally high rhBMP-2 concentration. In addition, the surgical reaming technique for this procedure required a deeper penetration of the endplates than was necessary in the tapered interbody fusion cage study. This increase may have resulted in increased exposure of the rhBMP-2 to cancellous bone that, in turn, may have increased access to mesenchymal stem cells and osteoclasts in trabecular bone.

The follow-up radiographic analysis (5) was done in a retrospective manner from films gathered in the prospective IDE reported in the Journal of Bone and Joint Surgery in 2005 (4). The follow-up study was conducted in response to other reports with similar observations after the market introduction of INFUSE Bone Graft (Medtronic Spinal and Biologics, Memphis, TN) and completion of the initial FDA study. The advantage of using these data to better understand the possible mechanism and cause of these defects is that this study was conducted in the context of an IDE trial in which the amount of rhBMP-2 and sponge was determined prospectively and used consistently throughout the patient series. We know now, in retrospect, one confounding variable in this trial may be the variability in the internal volume of the bone dowels because the central canal was not machined.

In a presentation at the 2007 Annual Meeting of the North American Spine Society and the subsequently published article (6), it was reported that overfilling a defect or hyperconcentrating the rhBMP-2 solution resulted in short-term radiolucencies when it was implanted into the cancellous bone within the distal femur of sheep. The histological evaluation in animals treated in that manner exhibited an accelerated level of bone turnover whereby, in the early time points, the neighboring bone tissue was observed to be demineralized. Normal healing without any change in the surrounding tissue was observed in defects that were treated appropriately (i.e., not overfilled nor hyperconcentrated). With longer- term follow-up, these radiolucent zones, like those reported by Burkus et al. (5), did undergo mineralization and appear to be healing.

Meisel et al. (7) demonstrated a similar finding in humans. In 16 of 17 patients, Meisel administered 12 mg of rhBMP-2 equally divided within two PEEK Telamon cages. In terms of volume, they delivered 4 cc of INFUSE Bone Graft in each interbody construct with an internal volume ranging from 0.75 to 1.3 cc, depending on their size. Therefore, they were significantly overfilling the device relative to its internal volume in a manner similar to that evaluated in the sheep study referenced earlier. They reported that all patients attained a radiographic fusion by 6-month follow-up despite observing the transient resorption surrounding the cage at the 3-month time point. In a personal communication, Meisel cites the sheep study mentioned above to help put these findings into perspective.

Dr. Smoljanovic and associates ask, ”We would like to know whether any of the investigators at the numerous clinical sites at which the patients were enrolled (5 plus 13) in the study over the 3-year period (May 1998 to March 2001) considered discontinuing and redesigning the study after the unexpected large vertebral resorptions were observed 6 months after the surgery in their patients”. Again, the facts are as follows: 1) The accelerated remodeling occurred in only 18% of the patients; 2) the remodeling zones were small and localized to a small (4.0 mm) area adjacent to only one of the four openings of the bone dowels; 3) the remodeling zones were seen only on thin-cut CT scans; 4) the remodeling zones were transient and all went on to heal by 24 months; 5) the remodeling had no impact on clinical or radiographic outcomes. We are at a loss as to why Smoljanovic et al. would suggest that a study would be halted or altered based on an isolated and transient radiographic finding that did not impact clinical or fusion outcomes. It was noted in the radiographic study (5) that there were no differences in fusion rates, Oswestry Disability Index scores, Short Form-36 Physical Component Summary scores, or re-operations associated with the radiographic appearance of remodeling observed at 6 months. In fact, the study that is the subject of the article currently under discussion demonstrated that superior outcomes were achievable in some of these parameters when compared with patients treated with iliac crest bone graft.

This phenomenon was not observed in the pivotal study for rhBMP-2/ACS in combination with the LT Cage fusion devices and, therefore, was not mentioned in the original product labeling. The reaming techniques of the endplates and exposure of cancellous bone were significantly different in these two studies. Since completion of the studies mentioned earlier and out of an abundance of caution, the package insert has been updated (2008) to advise surgeons of the risk of inappropriate use described as overfilling a defect or hyperconcentrating the rhBMP-2 solution when preparing the INFUSE Bone Graft.

References

1. McClellan JW, Mulconrey DS, Forbes RJ, Fullmer N. Vertebral bone resorption after transforaminal lumbar interbody fusion with bone morphogenetic protein (rhBMP-2). J Spinal Disord Tech. 2006;19:483–6.

2. Pradhan BB, Bae HW, Dawson EG, Patel VV, Delamarter RB. Graft resorption with the use of bone morphogenetic protein: lessons from anterior lumbar interbody fusion using femoral ring allografts and recombinant human bone morphogenetic protein-2. Spine. 2006;31:E277–84.

3. Burkus JK, Transfeldt EE, Kitchel SH, Watkins RG, Balderston RA. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine. 2002;27:2396-408.

4. Burkus JK, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery. J Bone Joint Surg Am. 2005;87:1205–12.

5. Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine. 2006;31:775–81.

6. Toth JM, Boden SD, Burkus JK, Badura JM, Peckham SM, McKay WF. Short-term osteoclastic activity induced by locally high concentrations of recombinant human bone morphogenetic protein–2 in a cancellous bone environment. Spine. 2009;34:539-50.

7. Meisel HJ, Beier A, Hoell T, et al. Transient reduced mineral density associated with BMP-enhanced spinal fusion [abstract]. In: Proceedings of the North American Spine Society, 20th Annual Meeting; Spine J; 2005. p 55S–6S.

To the Editor:

In the paper “Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery”, Burkus et al. (1) presented clinical and radiographic outcomes in 79 patients who received recombinant human bone morphogenetic protein-2 soaked into an absorbable collagen sponge (rhBMP-2/ACS) in combination with threaded cortical allograft dowels and compared those results with the outcomes in 52 patients who received autograft in a stand alone anterior lumbar interbody fusion (LIF). It was a prospective, randomized, multicenter United States Food and Drug Administration (FDA)-approved investigational device exemption study conducted in two sequential phases. In the pilot phase (2), 46 patients were enrolled at 5 clinical sites, and in the pivotal phase (1), 85 patients were enrolled at 13 clinical sites. The patients were enrolled over a three-year period (May 1998 to March 2001). The study protocols for both phases were identical. Stability and radiolucent lines were assessed on plain radiographs with use of anteroposterior, lateral, and flexion-extension views. In addition, thin-slice (1-mm overlapping) computed tomography (CT) scans with coronal and sagittal plane reconstructions were used to assess bridging bone and allograft incorporation. The radiographs and CT scans were reviewed by two independent radiologists in a blinded fashion to critically assess fusion at 6, 12, and 24 months. A third independent radiologist was used to adjudicate conflicting findings.

However, we believe there are substantial shortcomings in the two studies (1,2). First, the authors did not report any transient resorption of trabecular bone within vertebral bodies after the rhBMP-2/ACS application in the pilot phase (2). Furthermore, they claimed that because of early incorporation of the allograft into the vertebral end plates in the rhBMP/ACS group, radiolucent lines were not seen after surgery in the investigational group, and so the study proceeded to the pivotal phase. However, although large osteolytic cysts were discovered within the vertebral bodies several years later (Figure 10) (3), there was no explanation offered by the authors.

Secondly, although the authors claimed that the study protocols for both phases were identical (1), it was found that the dose of rhBMP-2 used per level of interbody fusion was decreased from 12 - 18 mg in the pilot phase to 8.4 - 12 mg in the pivotal phase (4).

Although the dose of rhBMP-2 was decreased by 30%, 18% of the resorptions occurred among the patients from the rhBMP-2/ACS group at the end of the second phase (1). Our first analysis, performed almost three years ago, revealed that the resorptions occurred in cases when additional rhBMP-2/ACS (other than those placed within the interbody spacers) was placed adjacent to the interbody implant in direct contact with the vertebral endplates (5). Subsequently, a more specific analysis revealed that because of the imperfection of the carrier, i.e. initial burst release of rhBMPs from collagen sponge (6), the size of the contact area between the rhBMP-2/ACS and the trabecular bone of vertebral bodies is the most important factor associated with the incidence of the resorptions (7) (and T. Smoljanovic, et al., unpublished data, 2008). In the studies of Burkus et al. (1,2) the larger contact area was created by placing of an additional rhBMP-2/ACS between the bone dowels, the central portion of which was filled by single rhBMP-2/ACS. Another way of creating the larger contact area between the rhBMP-2/ACS and vertebral endplates in the other studies, which reported the resorptions, was achieved due to the design of the interbody spacers, i.e. by their large apertures as in case of femoral ring allografts (FRA) or poly-ethyl-ether-ketone (PEEK) cages. To be fair, it should be mentioned that there are several studies in which the larger contact surface between the rhBMP-2/ACS and trabecular bone of vertebral bodies existed, and which had CT control within early follow up, but in which no resorption was reported (T. Smoljanovic, et al., unpublished data, 2008).

We speculated why the authors (2), who did not report resorptions in the pilot phase (3), then presented the CT scans with such large resorptions. One might speculate that there simply was no CT scan without resorptions within early follow up among the patients in the rhBMP-2/ACS group. There are reports in the literature in which the resorptions were found in each case where CT scanning was performed within the early follow up in patients who underwent LIF assisted with rhBMP-2/ACS (8,9), but as only some patients were scanned (usually those who were symptomatic (8)), it was not possible to raise any questions about the incidence of resorptions in the work of Burkus et al. (1,2) until recently when a paper by Meisel et al. (10) was published. Meisel et al. (10) performed CT scanning at 3 months after the surgery in all patients (N=17) who underwent posterior LIF assisted with rhBMP-2/ACS within the PEEK cage and with additional posterior fixation. All those patients, in the words of the authors, were asymptomatic. The resorptions surrounding the PEEK cage were found in 100% of the patients. Although the dose of rhBMP-2 clinically used per level of interbody fusion was found to have no influence on development of the resorptions (T. Smoljanovic, et al., unpublished data, 2008), we should mention that Meisel et al. (10) used from 6 to 12 mg of rhBMP-2 per level.

Finally, we come to clinical consequences of the resorptions. Mentioning fusion assessment outcomes, Burkus et al. (1) stated that in the autograft group, no patient had a fracture or migration or extrusion of the allograft implant. But, the authors did not mention migration or extrusion of the allograft implant (neither subsidence of interbody spacer nor loss of correction) presenting the results of rhBMP-2/ACS group. It was only mentioned that all radiolucent areas had resolved by 24 months after surgery. However, the analysis of reported resorptions revealed that patients in whom the resorptions developed were often faced with spacer subsidence, loss of correction, graft migration and the failure of spinal interbody fusion even with additional stabilization of fused levels (7) (and T. Smoljanovic, et al., unpublished data, 2008). However, there is another study, in which additional stabilization was not used as in the study of Burkus et al. (1,2), and in which observed resorptions had no clinical impact (11).

Although Burkus et al. (1,2) did not report clinical significance of the resorptions, we would like to know whether any of the investigators at the numerous clinical sites at which the patients were enrolled (5 plus 13) in the study over the three-year period (May 1998 to March 2001) considered discontinuing and redesigning the study after the unexpected large vertebral resorptions were observed 6 months after the surgery in their patients. (There are examples where enrollment of patients was stopped immediately after preliminary results of the resorptions become available (12,13), or where the use of rhBMP-2/ACS was abandoned in practice after completion of the studies due to the side effects, high cost, and the availability of a suitable alternative (9,14)). We also question why there was no mention of resorptions among potential adverse effects of the device in the premarket approval application (PMA) of the InFUSE Bone Graft/LT-CAGE Lumbar Tapered Fusion Device to the Center for Devices and Radiological Health (CDHR) of the FDA (15)? We have previously posed these questions and they remain unanswered (3,4,16,17). We hope that this letter will stimulate discussion in the wider medical scientific community to resolve them.

The authors did not receive any outside funding or grants in support of their research for or preparation of this work. Neither they nor a member of their immediate families received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, division, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immediate families, are affiliated or associated.

References

1. Burkus JK, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery. J Bone Joint Surg Am. 2005;87:1205-12.

2. Burkus JK, Transfeldt EE, Kitchel SH, Watkins RG, Balderston RA. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine. 2002;27:2396-408.

3. Smoljanovic T, Pecina M. Re: Burkus JK, Transfeldt EE, Kitchel SH, et al. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine 2002;27:2396-408. Spine. 2008;33:224.

4. Smoljanovic T, Pecina M. Unexplained decreasing of rhBMP-2 dose. J Bone Joint Surg Am. Electronically published letter. 23rd October 2007. Available from: http://www.ejbjs.org/cgi/eletters/87/6/1205#5117.

5. Smoljanovic T, Grgurevic L, Jelic M, Kreszinger M, Haspl M, Maticic D, Vukicevic S, Pecina M. Regeneration of the skeleton by recombinant human bone morphogenetic proteins. Coll Antropol. 2007;31:923-32.

6. Seeherman H, Wozney JM. Delivery of bone morphogenetic proteins for orthopedic tissue regeneration. Cytokine Growth Factor Rev. 2005;16:329-45.

7. Smoljanovic T, Bicanic G, Bojanic I. Update of comprehensive review of the safety profile of bone morphogenetic protein in spine surgery. Neurosurgery. 2009; In press.

8. Lewandrowski KU, Nanson C, Calderon R. Vertebral osteolysis after posterior interbody lumbar fusion with recombinant human bone morphogenetic protein 2: a report of five cases. Spine J. 2007;7:609-14.

9. Vaidya R, Weir R, Sethi A, Meisterling S, Hakeos W, Wybo CD. Interbody fusion with allograft and rhBMP-2 leads to consistent fusion but early subsidence. J Bone Joint Surg Br. 2007;89:342-5.

10. Meisel HJ, Schnöring M, Hohaus C, Minkus Y, Beier A, Ganey T, Mansmann U. Posterior lumbar interbody fusion using rhBMP-2. Eur Spine J. 2008;17:1735-44.

11. Kuklo TR, Rosner MK, Polly DW Jr. Computerized tomography evaluation of a resorbable implant after transforaminal lumbar interbody fusion. Neurosurg Focus. 2004;16:E10.

12. Laursen M, Hoy K, Hansen ES, Gelineck J, Christensen FB, Bunger CE. Recombinant bone morphogenetic protein-7 as an intracorporal bone growth stimulator in unstable thoracolumbar burst fractures in humans: preliminary results. Eur Spine J. 1999;8:485-90.

13. Pradhan BB, Bae HW, Dawson EG, Patel VV, Delamarter RB. Graft resorption with the use of bone morphogenetic protein: lessons from anterior lumbar interbody fusion using femoral ring allografts and recombinant human bone morphogenetic protein-2. Spine. 2006;31:E277-84.

14. Vaidya R, Carp J, Sethi A, Bartol S, Craig J, Les CM. Complications of anterior cervical discectomy and fusion using recombinant human bone morphogenetic protein-2. Eur Spine J. 2007;16:1257-65.

15. U.S. Food and Drug Administration, Department of Health and Human Services, Center for Devices and Radiological Health: Infuse Bone Graft /LT-Cage Lumbar Tapered Fusion Device—P000058. Available from: www.fda.gov/cdrh/pdf/P000058.html, Accessed January 12, 2007.

16. Smoljanovic T, Bojanic I, Dapic T. Significance of Early CT Evaluation after the Lumbar Interbody Fusions Assisted with rhBMP-2. Am J Neuroradiol. 2009; In press.

17. Smoljanovic T, Pecina M. Re: Burkus JK, Sandhu HS, Gornet MF. Influence of rhBMP-2 on the healing patterns associated with allograft interbody constructs in comparison with autograft. Spine 2006;31:775-81. Spine. 2008;33:226.