Recombinant human bone morphogenetic protein-2 (rhBMP-2) is a powerful osteoinductive protein. When implanted with an absorbable collagen sponge (ACS) carrier, it has been shown to induce bone formation and promote osseous spinal fusion in both animal and human studies1-33. It is available for clinical use, in a variety of doses, as INFUSE Bone Graft (Medtronic, Memphis, Tennessee). INFUSE has been approved by the U.S. Food and Drug Administration (FDA) for single-level anterior lumbar spinal interbody fusion when placed in certain threaded titanium cages.
The success of rhBMP-2 with lumbar interbody fusion has led to the recent expansion of its use to many “off-label” applications within the field of spinal surgery. Recently, numerous peer-reviewed publications have demonstrated rhBMP-2 to be successful when used for anterior cervical discectomy and fusion1-4. However, reports of adverse events related to increased soft-tissue edema adjacent to rhBMP-2 implantation sites in the cervical spine have surfaced34-37. It has been suggested that these adverse effects may be dose-dependent, with the use of higher doses of rhBMP-2 possibly increasing the rate and severity of soft-tissue edema38-40. However, there is little information regarding the magnitude or natural history of this peri-implant edema.
This experimental study was performed to examine the soft-tissue edema induced after implantation of various doses of rhBMP-2 in a rat ectopic model. The objectives were to characterize, with both magnetic resonance imaging (MRI) and histological examination, soft-tissue edema associated with increasing doses of rhBMP-2 implanted on an ACS or injected directly into soft tissues.
This study was reviewed by the Institutional Animal Care and Use Committee at MPI Research (Kalamazoo, Michigan) under study number 1064-024 and was approved by the committee on May 9, 2006. The study was conducted in two parts, with a total of thirty-six Lewis rats undergoing either intramuscular implantation of rhBMP-2 or buffer on an absorbable collagen sponge or intramuscular injection of rhBMP-2 solution. Four implantation sites—the latissimus dorsi and gluteus muscles on both sides—were used for each rat.
Part I
In Part I, sixteen rats received intramuscular implantation of rhBMP-2 in increasing doses (including 0, or buffer only) on an ACS carrier and were killed at either two or seven days. Eight additional buffer/ACS samples (0-μg group) from eight additional rats were obtained from another study arm to supplement the available data for Part I. Therefore, for both the two and the seven-day time point in Part I, eight samples were available for each rhBMP-2 dose and twelve samples were available for the buffer control (see Appendix). This resulted in seventy-two total samples for Part I.
The implants were prepared by soaking a 2.0 × 1.5 × 0.35-cm dry piece of ACS with 0.3 mL of the appropriate rhBMP-2 solution concentration (0, 0.1, 0.43, or 1.5 mg/mL). The rhBMP-2 solution was allowed to soak the ACS for at least fifteen minutes but for no more than 120 minutes prior to implantation into a muscle pouch. These rhBMP-2 concentrations have been shown to be effective at inducing osseous spinal fusion in rat, rabbit and sheep, and nonhuman primate models, respectively, demonstrating that increasing rhBMP-2 concentrations are required to induce bone formation in more advanced animal models. Addition of these rhBMP-2 solution concentrations to create a final implant volume of 0.3 mL resulted in the 0-μg (buffer control), 30-μg (normal), 129-μg (mid), and 450-μg (high) doses to be tested. Previous research has shown that 10 μg of rhBMP-2 implanted on an ACS is sufficient to achieve a 100% rate of fusion in a rat posterolateral spine fusion model41. With 30 μg being the lowest dose examined, and although it was three times higher than the previously effective dose of 10 μg, 30 μg was considered to be the “normal” dose for inducing bone formation in our study. Therefore, the “normal,” “mid,” and “high”-dose labels in the current study were considered to be relative to effective doses for the rat model. Following implantation in the first animal, the treatment placement was rotated in a “Z-pattern” (Fig. 1) (left latissimus, right latissimus, left gluteus, and right gluteus) to allow us to evaluate two implants with each dose (0 μg, 30 μg, 129 μg, and 450 μg) at each implant site at each time point. The additional 0-μg-dose (buffer) control samples utilized in Part I were obtained from another study arm, in which only buffer samples had been implanted and rotated in a similar fashion; therefore, there was no cross-talk from other rhBMP-2 samples.
Part II
In Part II, twelve rats received intramuscular injections of rhBMP-2 solution, as well as equal doses of rhBMP-2 implanted on an ACS, and were killed at two or seven days. Therefore, for both the two and the seven-day time point in Part II, six samples were available for each rhBMP-2 dose given via both implantation and injection (see Appendix). This resulted in a total of forty-eight samples for Part II (see Appendix). The Institutional Animal Care and Use Committee limited the intramuscular injection volume to 0.1 mL of rhBMP-2 solution. Thus, to enable direct comparison, the volumes of rhBMP-2 on the ACS implants were also limited to 0.1 mL. The implants were prepared by soaking a 2.0 × 1.5 × 0.35-cm dry piece of ACS with 0.1 mL of the appropriate rhBMP-2 solution concentration (0.1 or 1.5 mg/mL). The solution volumes of approximately 0.1 mL led to rhBMP-2 doses of 10 μg (normal) or 150 μg (mid). The rhBMP-2 solution was allowed to soak the ACS for at least fifteen minutes but for no more than 120 minutes prior to implantation into a muscle pouch. For sites receiving intramuscular injection, 0.1 mL of either 0.1 or 1.5 mg/mL rhBMP-2 solution was injected into the appropriate muscle at the time of surgery. This resulted in each intramuscular injection site receiving either 10 or 150 μg of rhBMP-2. It should be noted that surgical incisions were not made prior to injection of the rhBMP-2 into the muscle. As in Part I, a “Z-pattern” treatment rotation scheme was used to try to limit implant-site and possible rhBMP-2-dosage cross-talk influences. For Part II, six injection sites were used for each rhBMP-2 dose and six implantation sites were used for comparison with the same dose of rhBMP-2 implanted on an ACS.
The treatment rotation scheme and a study design overview schematic for Parts I and II are shown in Figure 1.
Surgical Procedure
All surgery was performed in a dedicated surgical area at MPI Research. The surgical sites were clipped, prepared with an iodine scrub alternated with 70% isopropyl alcohol, and painted with iodine solution. Anesthesia was maintained with isoflurane inhalation delivered in oxygen. A dorsal midline incision was made to expose the gluteus superficialis and latissimus dorsi muscles. An off-midline incision was made through each muscle, and four separate muscle pouches were made. After implantation, the muscle pouches and muscle layers were closed with nonabsorbable suture, and the skin was closed with skin staples. At that time, the rats were randomized to either a two or a seven-day survival time. Following surgery, the animals were closely monitored during anesthetic recovery for physiological disturbances including cardiovascular/respiratory depression, hypothermia, and excessive bleeding from the surgical sites. Long-term postoperative monitoring included daily inspection of surgical sites. Buprenorphine was administered (0.01 mg/kg, subcutaneously once a day) the day after surgery. Skin staples were not removed prior to necropsy, except to perform the MRI scans. Following the MRI scans, the wounds were restapled and the animals were submitted for necropsy.
Analysis
The MRI examinations were conducted via a transportable 1.0-T Picker MRI system with a quad knee coil at Northern Biomedical Research (NBR, Muskegon, Michigan). Sterile eye lubricant was placed in both eyes of the animals, and anesthesia was maintained by isoflurane inhalation delivered in oxygen. The staples were removed from the dorsal incision, and the MRI was performed. T2-weighted axial MRIs were acquired at a thickness of 3 mm, recorded, and saved for each animal. Each section was evaluated for hyperintense signals associated with the implant, in the peri-implant area, and unrelated edema. Unrelated edema, such as from the incision, was excluded from the implant-related edema calculations. Each implant area, including both the implant and the associated peri-implant edema, was traced, and the minimum, maximum, mean, and standard deviation of the signal were recorded. All tracings were verified for accuracy by another MRI technician. The area of the traced implant-related edema was then calculated. The area of each tracing was then multiplied by the slice thickness of 3 mm to obtain the volume of the traced area. The volumes were then totaled for each implant site analyzed and converted from cubic millimeters to cubic centimeters for reporting. Statistical comparisons were performed with use of analysis of variance (ANOVA).
Necropsy examinations were performed on all animals with procedures approved by a veterinary pathologist. The animals were killed with carbon dioxide inhalation followed by exsanguinations via the abdominal vena cava. At the time of necropsy, the animals were examined for external abnormalities including gross soft-tissue swelling and palpable masses. The individual implant/injection sites were removed, fixed in 10% neutral buffered formalin, and trimmed prior to histological processing.
Microscopic examination was carried out on fixed paraffin sections (approximately 5 μm) of the implant/injection sites from all animals at scheduled necropsies. A surface decalcification was performed, and four transverse sections were taken from each implant or injection site. Hematoxylin and eosin and Masson trichrome stain were used to identify newly developed vascular structures. Sections were also qualitatively analyzed for the presence of inflammatory cells, increased vascularity, and bone or cartilage-forming cells.
Source of Funding
External funding was provided by Medtronic, Inc., Memphis, Tennessee, to cover the costs associated with the animals, surgery, and implants.
Part I
Quantitative MRI demonstrated similar average edema volumes in the buffer-control and normal-dose rhBMP-2/ACS groups at two days (mean and standard deviation, 1.081 ± 0.505 cm3 and 1.045 ± 0.299 cm3, respectively). The buffer and normal-dose groups also had similar average edema volume measurements at seven days (0.539 ± 0.405 cm3 and 0.594 ± 0.323 cm3). The edema volumes of both the mid (p < 0.01) and the high (p < 0.05) dose rhBMP-2/ACS groups (1.761 ± 0.456 cm3 and 1.635 ± 0.458 cm3, respectively, at two days and 0.810 ± 0.372 cm3 and 0.766 ± 0.364 cm3 at seven days) were significantly higher than those in the buffer/ACS group at both two and seven days post-implantation. The average edema volumes associated with all dose concentrations and with the buffer/ACS were significantly higher at two days than they were at seven days post-implantation (p < 0.01), and they consistently decreased by approximately half over the five-day time period (Figs. 2, 3, and 4).
Histological assessment demonstrated similar capillary density in all groups at the two-day survival time. However, peri-implant cellular response increased with increasing rhBMP-2 dose (Fig. 5). At seven days, the rhBMP-2/ACS groups demonstrated cartilage formation and ossification as expected, with higher-dose rhBMP-2 sites demonstrating increasing cell response and amounts of ossification at the periphery of the residual ACS implant (Fig. 6). No ossification was identified in the control sites implanted with buffer/ACS. No other observation (capillary density, myofiber degeneration, hemorrhage, or subacute inflammation) differed significantly from the control-implant observations, and the findings were considered primarily related to the surgical procedure and normal tissue response to the physical presence of the implant.
Part II
Quantitative MRI results demonstrated that both doses of rhBMP-2 injected intramuscularly resulted in significantly decreased average edema volumes compared with those following implantation of equal doses of rhBMP-2 on the ACS. This was observed at both two and seven days. Also, similar to what was seen in Part I, the average edema volumes in the rhBMP-2/ACS group were transient, significantly decreasing from two to seven days (Figs. 7 and 8).
The effects of the rhBMP-2/ACS were apparent, with increased fibroplasia at two and seven days as well as increased chondroplasia/ossification and mineralization at seven days, compared with the implant sites that received injections of rhBMP-2 solution. One site injected with the mid dose of rhBMP-2 did demonstrate some mild chondroplasia/ossification. Sites that received rhBMP-2/ACS demonstrated similar capillary density at both doses.
Recombinant human BMP-2 is a powerful osteoinductive agent, demonstrating both chemotactic and angiogenic properties capable of inducing bone formation in both animal and human studies1-33. Available clinically as INFUSE Bone Graft (Medtronic), rhBMP-2 received premarket approval from the FDA in 2002 for single-level anterior lumbar interbody fusion when delivered on an ACS in an LT-Cage Lumbar Tapered Fusion Device (Medtronic)18-42. Subsequently, INFUSE Bone Graft also received premarket approval for use for acute open tibial fractures stabilized with intramedullary nail fixation and for sinus augmentation and alveolar ridge augmentation associated with extraction sockets. With its clinical success as an alternative to autograft, rhBMP-2 use has quickly expanded to other “off-label” applications, including anterior cervical interbody fusion.
While a number of studies have shown that rhBMP-2 can induce cervical fusion with limited adverse events1-4, others have demonstrated substantial adverse events related to increased soft-tissue edema adjacent to implantation sites34-37. Some reports have suggested that soft-tissue edema may be rhBMP-2-dose-dependent, with higher doses possibly increasing the rate and severity of soft-tissue edema38-40. The literature contains little information characterizing the magnitude or natural history of peri-implant edema associated with the use of rhBMP-2. The purpose of this study was to examine the soft-tissue edema induced after implantation of rhBMP-2 in various doses in a common rat ectopic model.
Our study demonstrated several important findings regarding rhBMP-2 and its associated peri-implant edema. First, implantation of rhBMP-2 on an ACS resulted in peri-implant soft-tissue edema, which occurred in a dose-dependent manner as observed on T2-weighted MRI scans. Implantation of a normal dose of rhBMP-2 resulted in average edema volumes similar to those associated with implantation of buffer/ACS (the control), at both two and seven days post-implantation. The edema volumes associated with the mid and high doses of rhBMP-2 were significantly higher than those associated with the buffer control at both time points. These data clearly demonstrate a dose-dependent relationship between rhBMP-2 and peri-implant soft-tissue edema in the rat model. This dose dependence was also demonstrated on histological examination of the rhBMP-2 implantation sites, as a heightened cellular response occurred with higher rhBMP-2 doses. No ossification was identified in the control ACS samples as expected, but increasing amounts of bone formation were seen with increasing doses of rhBMP-2 on the ACS.
The edema volume in the high-dose rhBMP-2 group did not significantly increase above that seen with the mid dose of rhBMP-2. This plateau in edema volume may represent a limit to the magnitude of local edema possible with rhBMP-2 related to host factors or physical space limitations in this rat model. For example, once the chemotactic and angiogenic capacity of the host is maximized, the resultant volume of local edema in this animal species may be maximized. Furthermore, the physical space for edema to occur within the muscle pouches and surrounding muscle may limit the amount of soft-tissue edema possible. Whether or not this response limitation occurs in humans requires further investigation.
Notably, the lowest dose of rhBMP-2 used in this study was 10 μg. While this dose has previously proven sufficient for osseous healing in rats, it failed to induce edema above that associated with the control ACS alone. This finding highlights the importance of using appropriate doses of rhBMP-2 in the clinical setting. These data suggest that the ideal rhBMP-2 dose, which maximizes osteoinductivity and minimizes the risk of soft-tissue edema, has not yet been determined for use with anterior cervical fusion in humans.
The second prominent finding in this study was that the peri-implant soft-tissue edema associated with rhBMP-2/ACS implantation was transient. Every dose of rhBMP-2 demonstrated a consistent trend of significantly decreasing edema volumes from day two to seven days post-implantation. To our knowledge, this is the first study to objectively quantify the postoperative soft-tissue edema associated with rhBMP-2 implantation.
The third important finding is the lack of soft-tissue edema after direct intramuscular injection of rhBMP-2 solution. No significant edema was seen after injection of either dose of rhBMP-2 solution in this fashion. Likewise, no substantial chondroplasia or ossification was seen histologically after intramuscular injection of rhBMP-2 at either dose. This finding has important implications with regard to the logistical use of rhBMP-2 for any application. Clearly, the ACS is necessary as a carrier for rhBMP-2 to produce the desired osteoinductive effects. Perhaps more importantly, however, is the likelihood that free rhBMP-2 solution that has dripped or leaked into the soft tissues surrounding an area of proposed spinal fusion produces limited additional inflammatory reaction or increased soft-tissue edema. This is consistent with previous basic-science research demonstrating that rhBMP-2 requires the ACS to induce reproducible bone formation and that rhBMP-2 is quickly metabolized if it is not bound to a carrier42-43.
In conclusion, peri-implant soft-tissue edema occurs in a dose-dependent manner after implantation of rhBMP-2 on an ACS in a rat model. This was demonstrated by increasing soft-tissue edema volumes measured on MRIs as well as heightened cellular responses seen in histological specimens. This edema was transient, however, with all rhBMP-2 groups demonstrating significantly decreased edema at seven days compared with that found at two days post-implantation. At the low dose, rhBMP-2/ACS produced edema equal to that associated with the control ACS implanted alone (i.e., without rhBMP-2). Finally, solutions containing rhBMP-2 that were injected directly into muscles without the ACS carrier resulted in little or no soft-tissue edema.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.