Surgical Reconstruction of Canine Nonunion Fractures Using Bone Morphogenetic Protein-2-loaded Alginate Microbeads and Bone Allografts

Abstract

Background/Aim

Effective treatment of nonunion fractures is challenging as it requires a biological and mechanical environment to promote sufficient osteogenesis. Herein, we present a case series in which we evaluated the clinical efficacy of bone morphogenetic protein-2 (BMP-2)-loaded alginate microbeads and allografts in two dogs with nonunion fractures.

Case Report

A 3-year-old, 2.3-kg, spayed female Pomeranian (Case 1) presented with intermittent lameness of the left forelimb after radial and ulnar fracture repair 8 weeks prior. A 4-year-old, 4.8-kg, spayed female Pomeranian (Case 2) was referred for non-weight-bearing lameness of the left hindlimb due to implant failure following left tibial fracture repair. Both dogs had atrophic bone ends and no bridging calluses at the fracture site on radiographs, and were diagnosed with nonviable nonunion fractures of the radius/ulna and tibia, respectively. The surgical approach involved implant removal, debridement, and fracture gap reconstruction. BMP-2 was loaded into alginate microbeads for a prolonged release with bone allograft chips in both cases. In Case 1, bead grafts were applied directly at the fracture site, while in Case 2, they were implanted inside a frozen cortical bone allograft as a scaffold to fill the large gap. Postoperative radiography revealed excessive callus formation, early radiographic bone union, and cortical bone remodeling, in line with improved lameness scores. At the final follow-up, gait was improved and the desired bone length and shape were achieved in both cases.

Conclusion

Simultaneous use of osteoinductive BMP-2 alginate microbeads and osteoconductive bone allografts yielded functionally and structurally favorable outcomes in canine nonunion fractures, without major complications.

Nonunion fracture is a condition in which the healing process of a fractured area does not progress, regardless of the recovery period after the fracture. The reported incidence rates of nonunion fractures are approximately 4.6% and 4.3% in dogs and cats, respectively (1,2), and the causes are related to a lack of mechanical stability at the fracture site, or an inadequate biological environment for bone healing (3). The classification of nonunion fracture varies depending on the underlying cause. When biological factors are sufficient but mechanical stability is lacking, it is referred to as a viable nonunion. Conversely, when adequate fixation is provided, but biological factors such as blood supply are insufficient, it is termed nonviable nonunion (4,5).

In general, nonunion fractures are more frequently managed with surgical intervention than with conservative approaches. Several factors contribute to the successful healing of nonunion fractures. Recently, the concept of diamonds has emerged as a pivotal framework for addressing these fractures (6). This concept emphasizes the significance of five essential elements in nonunion fracture management: the use of osteoconductive grafts to facilitate bone growth, the application of stem cells responsible for bone differentiation and formation, ensuring essential vascularization at the fracture site for healing, and establishing mechanical stability. Moreover, the incorporation of bioactive growth factors has been shown to be crucial to enhance bone differentiation (7).

Bone morphogenetic protein-2 (BMP-2) is a growth factor known for its osteoinductive properties; it has been shown to promote the differentiation of stem cells into bone tissue. However, BMP-2 has a very short half-life, typically ranging from 7 to 16 min, necessitating the application of high doses to achieve clinical efficacy (8). However, the use of supraphysiological BMP-2 doses has been associated with adverse effects, including heterotopic bone formation and inflammation (9). To minimize these side effects, delivery carriers capable of continuously releasing an appropriate amount of BMP-2 are needed. In clinical practice, bioceramics, such as hydroxyapatite and beta-TCP, as well as materials such as collagen membranes, are employed as BMP-2 delivery vehicles (10,11).

In this study, alginate, a biocompatible material derived from brown algae, was used as the delivery carrier for BMP-2. Sodium alginate solution readily undergoes gelation via cross-linking reactions with cations, such as Ca²⁺ or Ba²⁺ (12). Subsequently, our previous studies confirmed a sustained release pattern of BMP-2 from alginate microbeads for up to 28 days using an in vitro test. Furthermore, when these BMP-2-loaded alginate beads were employed in various bone defect models, including the calvarial and tibial metaphyseal defect models (13,14), they demonstrated significantly higher bone regeneration than the groups treated with collagen sponge or hydroxyapatite. Based on these findings, we applied BMP-2 alginate beads to patients with nonunion fractures and evaluated their clinical efficacy.

Case Report

Case presentation. Dog 1, a 3-year-old spayed female Pomeranian weighing 2.3 kg, presented to Seoul National University Veterinary Medical Teaching Hospital (SNU-VMTH) with non-weight-bearing lameness of the left forelimb following a fall. The patient was diagnosed with a left radial and ulnar fracture and underwent fracture repair surgery at our institution. After an 8-week postoperative re-evaluation, the patient exhibited a mild pain response during manipulation of the left radius, but no other significant observations were noted. Radiographic images of the fracture site revealed a decrease in bone diameter and atrophic changes in both the proximal and distal bone segments compared with prior images (Figure 1A). These findings strongly suggested nonviable nonunion of the radial and ulnar fractures, thus prompting the decision to proceed with surgical intervention.

Figure 1. Preoperative radiographic images from case 1 (A) and case 2 (B). Mediolateral and craniocaudal radiographs of the affected limb were taken to assess the fracture site. Both preoperative radiographs revealed atrophic changes in the fracture ends and absence of callus formation. Both cases were consequently diagnosed as nonviable nonunion.

Dog 2, a 4-year-old spayed female Pomeranian weighing 4.8 kg, was initially treated at a local veterinary hospital after falling down the stairs, resulting in a left tibial fracture. Initial treatment involved fracture repair surgery; however, two weeks post-surgery, plate failure led to a recurrent fracture. Subsequently, a second operation was performed at a local hospital, but five months later, the fracture recurred, necessitating referral to our institute (SNU-VMTH) for further management. Orthopedic examination at the initial visit revealed non-weight-bearing lameness and severe instability of the left hind limb. Radiographic evaluation indicated the absence of callus formation at the previous fracture site and presence of a bone defect (Figure 1B). Based on these findings, a diagnosis of nonviable nonunion of the left tibial fracture was established, and nonunion fracture repair surgery was planned.

Revision surgery. Case 1. During the surgical procedures, case 1 was premedicated with cefazolin (33 mg/kg IV) and midazolam (0.1 mg/kg IV). Anesthesia was induced with alfaxalone (2 mg/kg IV) and maintained with isoflurane (1.5-2.0%). Local anesthesia was induced using a brachial plexus block with bupivacaine (6 mg). Throughout the surgery, a continuous rate infusion (CRI) of remifentanil-midazolam-ketamine at a rate of 1 ml/kg/hr was administered for analgesia, followed by a loading dose of remifentanil-ketamine. The left forelimb was extensively shaved and subjected to aseptic disinfection and sterile draping. With the dog in the dorsal recumbent position, a skin incision was made in the left craniomedial region. The antebrachial fascia was incised to expose the proximal radius, and the extensor carpi radialis muscle was laterally retracted. After confirming the previously applied implants, they were removed and tested by bacterial culture. Instability was observed between the bone fragments during manipulation, supporting nonunion of the bone fracture. Subsequently, the sclerotic bone ends were resected using an oscillating saw, and a 0.8 mm K-wire was used to drill the medullary canal until bleeding was induced (Figure 2A and B). After bone fixation with a 1.5 mm, 8-hole dynamic locking plate and screws (DePuy Synthes, Zuchwil, Switzerland) (Figure 2C), we applied a combination of BMP-2-containing alginate microbeads and cancellous bone allograft chips to the fracture site to ensure prolonged delivery of BMP-2 (Figure 2D). As previously described in our studies (13,14), BMP-2-loaded alginate microbeads were created from 1 ml of a 1.2% sodium alginate solution containing 0.5 mg/ml BMP-2 (Novosis; CGBio, Seoul, Republic of Korea) using an encapsulator (B-395 Pro; Büchi Labortechnik AG, Flawil, Switzerland). The beads were then gelled in 100 mM CaCl₂ for 30 min and washed thrice with PBS before use. Cancellous bone allograft chips (3 cc; Veterinary Tissue Bank, Wrexham, UK) were also soaked in a solution containing 0.25 mg BMP-2 for 10 min before application. Following this, Ad-MSC sheets, prepared using a previously-described protocol (15), were applied to promote bone healing and regeneration, while a collagen membrane containing gentamicin (Genta-Coll^® resorb; RESORBA Medical GmbH, Nürnberg, Germany) was placed over the fractured site to prevent graft migration and reduce the risk of infection (Figure 2E). Finally, the antebrachial fascia, subcutaneous tissue, and skin were all closed. A modified Robert Jones bandage was applied to the affected limb; this was replaced daily and removed three days after surgery. The patient was discharged five days after surgery with instructions to restrict activity. Cefadroxil (22 mg/kg), tramadol (2 mg/kg), and streptokinase (0.5 mg/kg) were administered orally twice daily, and pentoxifylline (10 mg/kg) was administered orally three times a day for seven days.

Figure 2. Intraoperative photographs from case 1. (A) Sclerotic fracture ends were confirmed. (B) A K-wire was inserted to open the medullary cavity. (C) A locking plate and screws were applied to stabilize the fractured bone. (D) Cancellous bone chip allografts and BMP-2-loaded alginate microbeads (white dotted box) were placed. (E) Following the application of adipose-derived mesenchymal stem cell sheet, a gentamicin-collagen membrane was wrapped around the fracture site.

Case 2. The anesthetic process was conducted identically to that in case 1, except for the use of femoral and sciatic nerve blocks as local anesthesia. The patient was positioned in the dorsal recumbent position, and the skin was incised on the medial region of the left tibia. After removal of the pre-existing implants, the implants and surrounding granulation tissue were sent for antibiotic susceptibility testing to confirm the absence of infection (Figure 3A). Following meticulous flushing of the fracture site to remove debris and contaminants, the sclerotic bone ends were removed using an oscillating saw (Figure 3B and C). The medullary cavity was drilled using K-wires until bleeding occurred (Figure 3D and E). Subsequently, a precontoured 1.5/2.0 mm locking reconstruction plate and 2.0 mm screws (Jeil Medical Corp., Seoul, Republic of Korea) were temporarily applied on the medial side of the tibia to align the proximal and distal bone fragments. The bone defect between the fragments was confirmed to be 2 cm long. After removing the temporary fixed plate from the proximal fragments, the plate was applied only to the distal fragments. For the preparation of the frozen cortical bone allograft, metacarpal bone shafts were harvested from dogs euthanized for reasons unrelated to the study, wrapped with 0.02 mg/ml gentamicin-soaked gauze, and kept at –80˚C, according to a previously established protocol (16). All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Seoul National University (SNU-220428-3). Prior to use, the cortical bone allograft was thawed in a normal saline solution at 37˚C, followed by osteotomy to adjust the length. Subsequently, the surrounding soft tissue and periosteum were meticulously trimmed, the medullary cavity was debrided using a needle and curette, and the bone graft underwent multiple washing cycles (Figure 3F). The medullary cavity of the prepared cortical bone allograft was filled with alginate microbeads and carefully positioned in the gap (Figure 3G). The allograft and bone segments were stabilized using a bone plate and screws on the medial side (Figure 3H). Orthogonal plating was achieved by applying a 1.2 mm locking plate (Jeil Medical Corp.) to the cranial side (Figure 3I). As in Case 1, Ad-MSC sheets, BMP-2-loaded alginate microbeads (0.5 mg/ml) and cancellous bone allograft chips (0.16 mg/ml) were administered, followed by envelopment of the implantation site with the collagen membrane (Figure 3J and K). Routine closures of the fascia, subcutaneous tissue, and skin were also performed. The dog was discharged three days after surgery with instructions for cage rest. Cefadroxil (22 mg/kg), streptokinase (0.5 mg/kg), and tramadol (4 mg/kg) were orally administered twice daily for 10 days.

Figure 3. Intraoperative photographs from case 2. (A) After removing the pre-existing implants, (B-E) the sclerotic bone ends were carefully removed with an oscillating saw, followed by insertion of a K-wire into the medullary cavity until bleeding occurred. (F and G) A prepared cortical bone allograft was trimmed to fit the space and filled with BMP-2 alginate microbeads. (H) This construct was then secured on the defect site and fixed with orthogonal locking plates and screws. (I-K) Subsequently, cancellous bone chip allograft and mesenchymal stem cell sheets were applied, and a gentamicin collagen membrane was placed around the fracture site.

Outcome and follow-up. Case 1. The bacterial culture results from the bone fragments and retrieved plates were negative. Clinical evaluation of lameness revealed improvements and mild lameness with partial weight-bearing 1 month after surgery. Two months after surgery, the dog showed no lameness, with normal weight-bearing on walking. Postoperative radiographic images at two weeks showed extensive callus formation around the fracture site and mild soft tissue swelling which subsided by four weeks postoperatively. Two months after surgery, the fracture line became less discernible, and bridging callus formation and cortical bone remodeling were evident, indicating radiographic union. At the three-month follow-up, two screws adjacent to the fracture line were removed to prevent potential stress shielding. Complete healing of the fracture site was maintained in the most recent follow-up images one year after surgery (Figure 4). The radial length, spanning from the radial head to the radiocarpal joint, was 6.47 cm, whereas the contralateral limb measured 6.73 cm, indicating a 3.1% shortening.

Figure 4. Postoperative radiographic images obtained from case 1. Mediolateral and craniocaudal radiographs were used to assess fracture healing. A massive bone callus could be clearly observed at two weeks. Radiographic bone union was confirmed two months after the revision surgery. Screws close to the fracture site were removed at three months. The fracture site is marked with yellow arrows.

Case 2. No bacterial growth was observed in the cultures of the bone fragments or on the removed plate. Regarding postoperative gait evaluation, the patient had preexisting bilateral grade 2 medial patellar luxation, which initially restricted a complete return to normal gait. However, gradual improvement in weight-bearing was observed one month after surgery, and mild lameness and full weight-bearing were observed two months after surgery. This improved gait condition remained consistent even at the 6-month follow-up evaluation. In the postoperative radiographs, substantial callus formation was observed proximally and distally around the cortical bone allograft. At the 2-month follow-up, evidence of connectivity between the allograft and the host bone was observed, indicating radiographic union (Figure 5). Furthermore, the radiographically measured length of the right tibia was 8.90 cm, and that of the left tibia was 8.78 cm, showing a 1.4% decrease in length (Figure 6).

Figure 5. Postoperative radiographic images obtained from case 2. An oblique view was used to assess the bone graft site. Bone callus formation was clearly observed at two weeks. Radiographic evidence of bone union was confirmed two months after revision surgery. The boundaries between the cortical bone allograft and the host bone are indicated by yellow arrows.

Figure 6. Limb length discrepancy with the contralateral limb was observed in both cases. Craniocaudal radiographic images were used to measure the bilateral limb length. (A) The left radial length measured 6.47 cm, while the right radial length was 6.73 cm, indicating a 3.1% reduction. (B) The length of the right tibia was 8.90 cm and that of the left tibia was 8.78 cm, representing a 1.4% decrease after bone healing.

Discussion

Herein, we present two cases of nonunion fractures that were successfully treated using a combination of osteoconduction, osteoinduction, osteogenesis, and fracture stabilization, satisfying the diamond concept. This is the first study to report the use of BMP-2-loaded alginate microbeads to successfully treat nonunion fractures in dogs. Radiographic bone union was confirmed on radiographs 56 days post-surgery in each case, which was notably shorter than the average duration of 77.6 and 208 days reported in retrospective studies of nonunion fractures treated with BMPs (1,17). Moreover, considering that other patients with nonunions treated at SNU-VMTH with similar graft materials, except BMP-2 alginate microbeads, achieved bone union within 3-6 months (Supplementary Table I), the continuous delivery of BMP-2 may induce rapid bone healing in these cases.

These results are consistent with the findings of previous studies on alginate microbeads, which showed a slow initial burst of BMP-2 release for 14 days, followed by a plateau of low-level release. Therefore, it has been hypothesized that BMP-2 may also act during the reparative stage, where the soft callus mineralizes and transforms into a hard callus to induce osteoblast differentiation and accelerate bone regeneration (18). Furthermore, accelerated bone union coincided with improvements in the clinical symptoms in both cases. This could be attributed to the substantial callus formation within the initial two weeks post-treatment, which in turn facilitates early stabilization of the fracture site, enabling immediate weight bearing and mobilization (19). Consequently, it has been postulated that the prolonged release of BMP-2 may concurrently promote the bone healing process and improve clinical symptoms.

Limb length discrepancy is a common complication of nonunion fracture surgery. Previous studies have shown that a discrepancy of 20% or more can alter the direction of joint reaction force, leading to gait disturbance (20,21). Therefore, cortical allografts are typically recommended for cases where the gap is expected to be greater than 20% of the total bone length. In the present study, the two cases exhibited a different gap size. Therefore, we only applied the cortical allograft in case 2, which had a larger fracture gap of 22% of the contralateral limb length. A 2-cm cortical bone allograft was implanted to fill the gap, and cancellous bone chips were applied around it. Although bone graft failures, such as allograft fracture, infection, or graft-host bone nonunion (22,23), can occur with nonunion fracture treatment using allografts, no adverse effects associated with bone grafts were observed in either case. Further, the post-operative limb length discrepancy in both cases was minimal at 96.1% and 98.6% of the contralateral limb length, respectively. This suggests that the application of appropriate bone grafts and successful osseointegration can minimize the risk of complications and gait disturbances caused by postoperative mechanical problems.

Regarding the clinical effectiveness of sustained release of BMP-2, neither of our cases developed significant complications such as excessive or ectopic bone formation, which are known to be associated with high-dose BMP-2 (24). Although mild soft tissue swelling at the fracture site was observed in both animals at 2 weeks, it spontaneously resolved in the four-week radiographic images, without requiring any supportive treatment. When calculated based on the volume, the concentrations of BMP-2 applied in each case were 0.25 mg/ml and 0.33 mg/ml in Cases 1 and 2, both of which were lower than the species-specific effective dose of BMP-2 (0.75 mg/ml) reported in spinal fusion in dogs (25). In addition, considering the release pattern of BMP-2 from the alginate microbeads, as measured in previous studies (13), the actual amount of BMP-2 administered was even lower. Therefore, sustained delivery of BMP-2 can induce appropriate bone regeneration while minimizing side effects.

Additionally, these dogs exhibited no immunogenic response related to the implanted alginate microbeads, in line with prior in vitro cytotoxicity tests, confirming the biocompatibility of the alginate microbeads (13). While the findings from this case report, which encompassed two dogs with up to one-year follow-up, are promising, the study’s limitations include the small number of cases and the absence of histological analysis of the regenerated bone tissue due to its application to patients. Furthermore, the assessment of osteogenesis was confined to radiographic evaluation. Therefore, it is imperative to conduct extensive long-term studies in more dogs to establish the clinical efficacy and safety of BMP-2-loaded alginate microbeads.

Conclusion

Overall, the combined application of bone allografts, BMP-2-loaded alginate microbeads, and Ad-MSCs with adequate internal fixation resulted in structurally and functionally favorable bone union in the two dogs with nonunion fractures. The controlled release of BMP-2 promoted effective bone growth and allowed for early weight-bearing while preventing potential complications associated with BMP-2 overdose and promoting a balance between bone formation and resorption at bone remodeling sites. These findings indicate that the use of bone allografts and prolonged BMP-2 delivery through alginate microbeads could be a viable therapeutic strategy for nonunion fractures.

Supplementary Material

Available at: https://doi.org/10.6084/m9.figshare.24772884.v1

Funding

This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) and funded by the Ministry of Education (No. 2018R1D1A1B07047451) and the Korean Government (MSIT) (No. 2023R1A2C1003001).

Conflicts of Interest

The Authors declare no conflicts of interest in relation to this study.

Authors’ Contributions

S.L. and B.-J.K. conceived the study; S.L. collected and analyzed the data and wrote the manuscript; B.-J.K. reviewed and revised the manuscript. All the Authors have read and approved the final manuscript.