Although the unquestionable value of autologous bone grafting and the analogous value of the reaming by-products in nonunion treatment have been mentioned extensively in the literature, there is ongoing vivid discussion for the treatment of those case scenarios where the fracture nonunion is complicated by other local environment adverse circumstances. The graft expansion with growth factors as the bone morphogenetic proteins (BMPs) offers the possibility to reduce the number of operative procedures, complications, length of hospital stay, and time to union. In this article, we consider the potential clinical scenarios for graft expansion with BMPs.
Despite the ongoing advances in the treatment of fractures and understanding of the fracture repair processes, impaired healing continues to be one of the most debilitating complications of fractures. Up to 10% of the 6.2 million fractures occurring annually in the United States are associated with impaired healing. Many of these cases of impaired fracture healing demonstrate unique characteristics posed not only by the initial trauma sustained with bone defects and impaired vascularity of the area but also as a result of previous treatment modalities. Many of these patients require lengthy treatments associated with both functional and psychosocial impairment. Not less worthy is the economical burden to the patient and the health system.
The standard treatment of most aseptic nonunions is mechanical stabilization with or without biologic stimulation depending on the assessment and classification of the nonunion.
The current gold standard for any given situation requiring bone grafting and especially in situations of fracture nonunion is autologous bone grafting (ABG). Autologous cancellous bone grafting remains a unique biologic method promoting union by stimulating the local biology at the nonunion site. Autologous bone has all three components necessary to promote or enhance bone regeneration: an osteoconductive scaffold, endogenous bioactive molecules, and cells that are able to respond to these signals. Unfortunately, although autogenous bone is considered as the best graft option, significant complications have been reported related to the harvesting site, most often being the anterior iliac crest of the pelvis. Furthermore, the desirable quantity of the required graft at times may be insufficient.
For these reasons, over the years other biologically based strategies have been developed. These include electrical, ultrasound, and shockwave stimulation, a wide range of bone graft substitutes with either osteoconductive or both osteoconductive and osteoinductive properties, and biologic response modifiers that are administered either locally or systemically, including bone morphogenetic proteins (BMPs), platelet-derived growth factors, and parathyroid hormone. These biologic response modifiers, appear to have been used successfully in managing nonunions. In addition to nonunion, the administration of these molecules has been used in many other orthopedic situations, including stabilization of implants, restoration of large segmental bone defects, treatment of osteonecrosis of the femoral head, fusion of joints, cartilage regeneration, augmentation of periprosthetic fractures, and acceleration of fracture healing, especially in patients at high risk of fracture nonunion.
Nonetheless, there are still adverse clinical settings where despite providing the best mechanical environment modification complemented with ABG, failure has occurred. In addition, there are circumstances where the application of growth factors in isolation would not seem enough to promote successful bone healing.
In this study, therefore, we consider in what clinical situations implantation of autologous bone grafting may need enhancement with commercially available growth factors (BMP-2 and BMP-7) to promote successful bone healing.
The use of autologous bone grafting or reaming by-products
Tibia is the most common long bone to sustain a fracture. It has a high risk of developing nonunion because of the compromised soft tissue envelope especially over its anterior medial area. Consequently, it represents the bone with the highest overall incidence of nonunion, and the “nonunion model.”
In the atrophic nonunions, the biologic factor is considered to be mostly the problem, despite the perception that the vascularity at the nonunion site is not compromised. The oligotrophic and even more the atrophic nonunions present insufficient blood supply, or insufficient quantities of bone-forming cells. As a result, augmentation of this poor biologic environment through graft expansion is considered mandatory in achieving union in these difficult nonunion cases. Several reports exist in the literature illustrating the efficacy of autologous iliac crest bone graft (AICBG) in isolation but also in combination with other materials. Overall the success rate with AIGBG is approximately 80% to 90%.
The biologic properties of the “by-products” of reaming (RBP) have gained special interest very early in the history of reamed intramedullary nailing (IMN), representing an internal autografting procedure during closed reamed nailing. IMN and reaming offers the unique biomechanical advantages of an intramedullary splinting fixation, in association with the osteoinductive stimulus of the “by-products” of reaming. The vascular flow between endosteum and periosteum of the long bones retains nutrition and healing of the nonunion sites even after the temporary destruction of the endosteal blood flow until it is restored. Although it is debatable in the literature whether to perform the IMN procedure openly or closed, it seems that surgeons open the nonunion site in those cases where the existing hardware needs to be removed, in cases with severe malalignment, and in those cases where additional bone graft needs to be added owing to massive bony defects. Reckling and Waters reported favorable results in the series of 33 noninfected tibial nonunions that were treated with a posterolateral approach and cortico-cancellous bone application. On an average of 5 months solid healing was noted in 94% of the patients.
Megas and colleagues treated 50 cases of aseptic tibial nonunions with reamed interlocking nail. On average the reamed IMN was performed 15.6 months post injury. Various primary fixation methods were used and 36% had been open fractures. A closed IMN was attempted in all cases, but in 16 an open procedure was finally performed because of irreducible malalignment or for removal of previous hardware. Autologous cancellous graft was added in three cases because of the extent of the bone deficit. All fractures united in a 6-month period post nailing and the method was advocated as highly effective and safe for aseptic tibial nonunions.
In 1999, Wu and colleagues evaluated 25 cases of tibia shaft aseptic nonunions treated with exchange nailing. Most (88.9%) of the original fractures were closed, stabilized with dynamic nails and developed atrophic nonunions. Exchange nailing was performed without opening the nonunion site and the success rates were significant (96%) in an average period of 16 weeks.
In 2003, Wu presented another series of treating tibial aseptic nonunions with reamed IMN. In this study the original fixation method was plating and in 28 cases with adequate follow-up progressed to union. On average, the nonunions healed in 4.5 months (3.0–7.5) after removal of the original plate fixation, excessive reaming of the medullary canal, and insertion of a Kuntscher nail (13 cases) or locked gerhard küntscher (GK) nail (15 cases). However, the author suggested that whenever a large bony defect is present there should be additional bone grafting from the lilac crest and not performing excessive reaming of the medullary canal.
In the series of Devnani, long-bone fracture nonunions were treated with compression plate fixation and AICBG. Among them the author evaluated the time to union of 10 tibial aseptic nonunions, 8 atrophic and 2 hypertrophic. All of the tibial nonunions of this series united at an average of 19.8 weeks, with a satisfactory functional outcome.
A comparative study of Johnson and Marder for tibial aseptic nonunions treated with IMN was published in 1987. The authors used open IMN techniques and compared the effect on healing rates of bone grafting the nonunion site with the by-products of reaming, or with AICBG. Eleven atrophic and 11 hypertrophic cases were evaluated. Successfully treated were 20 nonunion sites (91.9%), with an average time to union of 12.5 weeks. A statistically significant difference between the atrophic and hypertrophic cases (14.4 vs 10.6 weeks, respectively) was identified. The authors compared their results with those of closed IMN techniques and identified major differences in the time to union in favor of their own open nailing techniques.
Sledge and colleagues have described their experience with static reamed IMN for a period of 6 years. Forty aseptic tibial nonunions were treated with reamed arbeitsgemeinschaft für osteosynthesefragen (AO) or GK nails. The original injury in 18 of them was an open fracture and opening of the fracture site was used in 27 (67.5%) of them for removal of implants, proper realignment, and also for AICBG enhancement. The average time to union was 7.1 months (12–67) and the union rates were 100%. No statistically significant differences were observed between open or closed IMN techniques and the use or not of AICBG for enhancement.
An overview of articles with the proven value of AICBG and RBPs is presented in Table 1 .
|Author Year||Anatomic Site||Method||Graft Used||Results/Union Rates|
|Reckling & Waters 1980||Plaster||33 ICAG||93.9%|
|Johnson & Marder 1987||tibial treatment nonunion fractures||90.9%|
|Sledge et al 1989||100%|
|Wiss et al 1992||All compression plating||39 ICAG||96%|
|Wiss & Stetson 1994||IMN||47 RBP||89%|
|Wu et al 1999||25 closed #s||Reamed IMN||25 RBP||96%|
|Devnani 2001||All compression plating||10 ICAG||100%|
|Friedlaender et al 2001||Locked IMN||62% versus 74%|
|Megas et al 2001||100%|
|Wu 2003||28 RBP||100%|
The use of growth factors
Aiming to overcome the limitations of the autologous bone grafting, bone morphogenetic proteins (BMP-7 and BMP-2) were produced by recombinant DNA technology. They are substances with great osteoinductive properties for the enhancement of bone regeneration in various clinical applications, including the treatment of fracture nonunions. The safety of their administration, combined with the lack of morbidity and the quantity restrictions that characterize autologous bone grafts, have given to this family of molecules a principal role over the other bone graft substitutes.
The initial experimental in vivo and vitro work on BMPs opened the way for the study of Friedlaender and colleagues that started in 1992 and published in 2001. In their milestone work, they evaluated the application of rhBMP-7 (recombinant human bone morphogenic protein-7 or OP1) in tibial nonunions. One hundred and twenty-four tibia aseptic nonunions were enrolled in a multicenter randomized prospective controlled trial. Either rhBMP-7 (in 63 nonunions) or autologous bone graft (in 61 nonunions) was used for enhancement of nonunion healing. The method of fixation was IMN for all cases (locked in 92%). At 9 months postsurgery, 62% of the rhBMP-7 group and 74% of the AICBG group demonstrated radiological union ( P = .158). Overall, the rhBMP-7 administration was safe and proved to be statistically comparable to the gold standard biologic enhancement of autograft. This randomized trial of Friedlaender and colleagues has established BMPs as a bone graft option (see Table 1 ).
The work of Cook (animal model) showed the efficacy of recombinant human osteogenic protein-1 in healing 2-cm segmental defects in nonhuman primates (in ulnae and tibia), whereas the controls filled with AICBG showed only little new bone formation.
This ability of BMPs to regenerate new bone was used in various situations. Following the work of Friedlaender and colleagues the tibia per se is the first model for clinical investigation of potential application of the BMPs. They have been associated with augmenting standard fixation and grafting methods in the acute setting of fractures as well as in established nonunions.
Lately, the use of BMPs enhanced with autologous grafting is emerging in literature in various adverse scenarios ( Table 2 ).
|Authors||Anatomic Site||Graft Used||Results||Comments|
|Vaccaro et al 2003||Posterolateral lumbar fusions – No instrumentation||rhBMP-7 (OP-1) + ICAG||Pilot safety and efficacy study of rhBMP-7 (OP-1)|
|Dimitriou et al 2005||Persistent long-bone atrophic nonunions|
|Ronga et al 2006||Observational, retrospective, nonrandomized (BMP-7 Italian Observational Study [BIOS] Group)|
|Kanakaris et al 2008||BMP-7/OP-1 in all Graft expansion with autograft 25, 36.8%||61 (89.7%)||multicenter registry and database (6 University centers) observational study|
In a multicenter registry and database (six university centers) observational study, Kanakaris and colleagues focused on the application of BMP-7/OP-1. They presented the preliminary results, of a prospective case series of aseptic tibial nonunions. Sixty-eight patients fulfilled the inclusion criteria for this observational study, with a minimum follow-up of 12 months. The median duration of tibial nonunion before BMP-7 application was 23 months (range 9–317). Patients had undergone a median of 2 (0–11) revision procedures before the administration of BMP-7. In 41%, the application of BMP-7 was combined with revision of the fixation at the nonunion site. In 25 cases (36.8%), the BMP-7 was expanded with the use of autologous bone graft (AICBG), out of which 14 (56%) had been previously treated unsuccessfully with ABG. Nonunion healing was verified in 61 (89.7%) of 68 in a median period of 6.5 months (range 3–15).
Ronga and colleagues, in an observational, retrospective, nonrandomized study on the use of BMP-7, reported on treating nonunions in various anatomic sites. The work was performed by the BMP-7 Italian Observational Study (BIOS) Group. The clinical series included 105 patients. Additional grafts were used based on the surgeon’s decision. Radiographic and clinical assessments were performed at progressive time intervals on two groups: BMP-7 + autograft (A) or BMP-7 alone (B). The mean follow-up was 29.2 months. The last assessment showed an overall 88.8% success rate with an average healing time of 7.9 months. In complicated cases, the success of the BMP-7 + autograft group reached 83.3% compared with 76.5% of the standalone BMP-7. If fewer than three operations preceded, healing was 90.6% compared with 87.8% (BMP-7), and if more than 3 operations preceded, the success dropped to 77.8% compared with 75.0% of BMP-7. At 9 months there was overlapping between the unions recorded in the two groups with healing rates of 86.0% and 85.7%, respectively. This is an observational study that illustrates the efficacy of BMP-7 with and without bone grafting for the treatment of long bone nonunions. This type of observation is confirmed by the comparison of the results between the BMP-7 alone group and the BMP-7 + autograft group. The increase in the percentage rate of consolidations at different follow-up shows that the interval between 6 and 9 months is a period that can still be considered useful before declaring failure of the treatment. The overlapping of the results according to the variables “previous complications and number of previous operations” confirms the efficacy of BMP-7.
Dimitriou and colleagues in their recent study evaluated the efficacy and safety of recombinant bone morphogenetic protein-7 (rhBMP-7 or OP-1) as a bone-stimulating agent in the treatment of persistent fracture nonunions. Twenty-five patients with 26 fracture nonunions were treated with rhBMP-7. There were 10 tibial nonunions, 8 femoral, 3 humeral, 3 ulnar, 1 patellar, and 1 clavicular nonunion. The mean follow-up was 15.3 months. The mean number of operations performed before rhBMP-7 application was 3.2, with autologous bone graft and bone marrow injection being used in 10 cases (38.5%). Both clinical and radiological union occurred in 24 (92.3%) cases, within a mean time of 4.2 months and 5.6 months, respectively. No complications or adverse effects from the use of rhBMP-7 were encountered. In 17 (65.4%) cases the BMP application was combined with AICBG. In 16 of these 17 cases nonunion healing was noted. From those who had only rhBMP-7 applied, 8 of 9 healed. Their study supports the view of application of rhBMP-7 combined with autologous bone grafting for the treatment of persistent fracture nonunions.
Vaccaro and colleagues in a pilot safety and efficacy study examined rhBMP-7 (OP-1) as an adjunct to iliac crest autograft in posterolateral lumbar fusions. They combined OP-1 putty with autograft for intertransverse process fusion of the lumbar spine in patients with symptomatic spinal stenosis and degenerative spondylolisthesis following spinal decompression. Twelve patients underwent laminectomy and partial or complete medial facetectomy as required for decompression of the neural elements followed by intertransverse process fusion by placing iliac crest autograft and OP-1 putty between the decorticated transverse processes. No instrumentation was used. Patients were followed clinically using the Oswestry scale and radiographically using static and dynamic radiographs to assess their fusion status. Radiographic outcome was compared with a historical control (autograft alone fusion without instrumentation for the treatment of degenerative spondylolisthesis). The results showed 9 (75%) of the 12 patients obtained at least a 20% improvement in their preoperative Oswestry score, whereas 6 (55%) of 11 patients with radiographic follow-up achieved a solid fusion by the criteria used in this study. Bridging bone on the anteroposterior film was observed in 10 (91%) of the 11 patients. No systemic toxicity, ectopic bone formation, recurrent stenosis, or other adverse events related to the OP-1 putty implant were observed. A successful fusion was observed in slightly over half the patients in this study, using stringent criteria without adjunctive spinal instrumentation. This study did not demonstrate the superiority of OP-1 combined with autograft over an autograft alone historical control, in which the fusion rate was approximately 55%.