Treatment of congenital pseudarthrosis of the tibia with native bovine BMP: a case report

Bone morphogenetic proteins (BMP) have been shown to induce bone formation and union in long bone defects and nonunions. We report a case of congenital pseudarthrosis of the tibia treated with a composite implant consisting of a biocoral frame, collagen carrier, and native bovine BMP extract. A six-year-old boy had persisting congenital proximal tibial pseudarthrosis despite six prior operations. At surgery, the sclerotic surfaces of both fragments were excised, fixation was performed using Ilizarov's device, and the composite implant and an autograft were applied to the nonunion site. Three months after the operation, radiographs showed union, and at four months, the Ilizarov device was removed. Two years later, the proximal pseudarthrosis remained clinically and radiologically united. It is concluded that BMP may contribute to the healing of congenital tibial pseudarthrosis of the tibia.     

Management of femoral diaphyseal nonunion after nailing with augmentative locked plating and bone graft

Objectives: Exchange nailing (EN) for aseptic femoral shaft nonunion is currently a standard orthopaedic treatment modality. However, according to recent studies there is occasionally a high failure rate when EN is used. In the present study, augmentative locked plating and bone graft was used as an alternative method for treating such cases. The purpose of this study was to report the treatment outcomes of selected femoral diaphyseal nonunions that had initially been treated by nailing. Methods: Thirteen patients with femoral diaphyseal nonunions underwent revision surgery with locked plate and bone graft. The inclusion criteria were: (a) nonisthmic nonunion; (b) atrophic/oligotrophic nonunion. All patients were managed by augmentative locked plating and bone graft. Results: All patients were followed up for a mean of 1.2 years (range, 1–2 years). After revision surgery all nonunions achieved bony union, the average union time being 7.5 months (range, 6–12 months). No major complications were observed in this case series. Conclusion: Augmentative locked plating is an effective and simple technique for treatment of femoral diaphyseal nonunion after nailing in selected cases.     

Augmentative Ilizarov external fixation after failure of diaphyseal union with intramedullary nailing

OBJECTIVE: To investigate the use of the Ilizarov circular fixator and nail retention in treating diaphyseal nonunion following previous intramedullary nailing. DESIGN: Retrospectively reviewed, consecutive series. Mean duration of follow-up after achieving bone union: 19.2 months (range 6 to 33 months). SETTING: A tertiary referral center for nonunion surgery. PATIENTS: Nine patients (two femoral, three tibial, and four humeral nonunions) were included in the study. All patients were referred from other centers after failure to achieve union with intramedullary nailing. Patients who had nonunion with other fixation devices in situ, those with active infection and nonunion following nonoperative treatment, were excluded from the study. The patients had undergone an average of 2.4 operations (range 1 to 5 operations) before application of the Ilizarov fixator. All patients completed the study. INTERVENTION: The circular fixator was used to compress the nonunion site from without, retaining the intramedullary nail in each case. We excluded a patient who had his nonunion site explored followed by bone excision and transport. The mean duration of fixator treatment was 6.2 months (3 to 11 months). MAIN OUTCOME MEASUREMENTS: Clinical and x-ray evidence of bone union, infection, residual deformity, shortening, and assessment of functional outcome. RESULTS: Bone union was achieved in all nine patients using the circular fixator over the nail. The bone results were graded as six excellent, one good, and two fair. All patients reported a reduction in pain and satisfaction with their final outcome. CONCLUSIONS: There is a role for the use of the Ilizarov fixator with nail retention in resistant long bone diaphyseal nonunion in carefully selected patients. This method can achieve high union rates where other treatment methods have failed.     

Resistant nonunions and partial or complete segmental defects of long bones. Treatment with implants of a composite of human bone morphogenetic protein (BMP) and autolyzed, antigen-extracted, allogeneic (AAA) bone.

Twenty-five patients with resistant nonunions including partial or complete segmental defects were treated with a composite alloimplant of human bone morphogenetic protein (h-BMP) and autolyzed, antigen-free, allogeneic bone (AAA). The series consisted of 16 females and nine males; average age was 45 years. Preoperative symptoms averaged 30 months (range, five to 83 months); 22 of 25 patients had failed multiple attempts at electrical stimulation. Twenty-three of 25 patients had an average of three prior failed surgical attempts at union (range, one to ten). There were ten segmental defects with an average length of 4 cm (range, 2-9 cm). The composite implant was incorporated as an onlay in 15 extremities and as an inlay graft supported by internal fixation in ten extremities. Seven patients received supplementary autogeneic cancellous bone grafting. Average healing time was six months (range, three to 14 months). Average follow-up time was 21 months (range, five to 82 months). Functional results were rated as excellent, 14; good, five; and fair, five. One failed to unite because of a recurrent infection. Union was obtained in 24 of 25 patients. There were five failures of the original operation that required reoperations; union eventually occurred in four of five extremities by repeat composite grafting and replacement of the failed internal fixation. Bony union between host bone and the composite implant began at an average of eight weeks postoperatively. Present results indicate that h-BMP/AAA composite implants represent adjunctive treatment of difficult nonunions. The h-BMP/AAA composite implants may be implanted in either partial or complete segmental defects of long bones.(ABSTRACT TRUNCATED AT 250 WORDS)     

Surgical treatment of proximal ulna nonunion

We reviewed our experience in the surgical treatment of 12 cases of proximal ulna nonunion. The primary injuries were 2 fracture-dislocations of the olecranon, 6 Monteggia lesions and 3 isolated fractures of the proximal ulna. According to the type of primary injury and its anatomical site, the nonunions were classified into 2 groups, considering that the nonunions nearest to the humerus-ulna joint present a more disabling clinical profile and are more difficult to treat: group A (6 patients — nonunion within 5 cm from the olecranon tip of the olecranon) and group B (6 patients — nonunion between 5 and 10 cm from the olecranon tip of the olecranon). In all cases, after fibrous callus debridement and bone surface remodelling, fixation was performed with plate and screws and homoplastic cortical bone graft (orthogonal or parallel to the plate) and an intercalary bone cylinder when the bone defect was severe. In 3 patients (group A), where the defect was smaller than 1 cm, fixation of the ulna was combined with a resection of the radial neck. Clinical-radiographic healing was achieved in all patients followed for a mean of 27 months. Complications included a case of nonunion due to failure of the intercalary graft with plate breakage. The patient healed after a new surgery performed with same technique. The score, according to the Broberg-Morrey scoring system, was 78 in group A patients and 93 in group B patients. The use of homoplastic cortical bone graft represents an effective technique to improve the mechanical properties of the fixation and supports biological union, even when the bone defect is severe.     

Shortening osteotomy and compression plating for atrophic humeral nonunions: surgical technique

The incidence of nonunion after humeral shaft fractures is low with both operative and nonoperative interventions. However, when a delayed union or nonunion occurs, it can be a very debilitating, leading to limited function and pain. Several methods have been described for the management of humeral diaphyseal nonunions. We present 11 cases managed with a transverse, shortening osteotomy and compression plating with or without bone grafting. Of the 11 cases, 10 had prior surgical treatment of open reduction and internal fixation. Union was achieved in all 11 humeri.     

Bone morphogenetic protein augmentation grafting of resistant femoral nonunions. A preliminary report

Twelve patients with intractable nonunions of the femoral diaphyseal or metaphyseal-diaphyseal shaft were successfully treated by a combination of internal fixation and implants of human bone morphogenetic protein (h-BMP). There was an average of 4.3 surgical procedures per patient attempting union prior to h-BMP implantation. Union was obtained in 11 of 12 patients and in one patient with a repeat stabilization and implantation of h-BMP. Four patients received autogeneic cancellous bone graft and four patients received allogeneic bone grafts. The BMP implant was prepared in the form of an aggregate of h-BMP and bone matrix water-insoluble noncollagenous proteins (h-BMP/iNCP). Fifty to 100 mg of h-BMP/iNCP was either implanted in the fracture gap in ultra thin gelatin capsules, or incorporated in a strip of polylactic/polyglycolic acid copolymer (PLA/PGA) and placed as an onlay across the fracture gap. The average time to union was 4.7 months. Further clinical investigations are planned as a series of matched cases with and without BMP augmentation in order to distinguish h-BMP effects from new or improved methods of fracture fixation combined with autogeneic cancellous bone grafts.     

Nonunion of the diaphysis of long bones

The treatment of diaphyseal nonunion of long bones is difficult and controversial. We retrospectively reviewed 113 patients with diaphyseal nonunion treated by various modalities, during 15 years at one institution. There were 36 cases of nonunion of the tibia, 23 nonunions of the femur, 21 nonunions of the humerus, 13 nonunions of the radius, 18 nonunions of the ulna and two nonunions of the clavicle. The minimum followup was 24 months (average, 40 months, range 2-15 years). The nonunions were classified as aseptic (84) and septic (29) and additionally classified as hypertrophic (61) and atrophic (52) in order to determine the treatment. The treatment was individualized based on the stability at the nonunion site, need for bone grafting, and control of infection. All fractures healed and every patient in the study regained functional use of the limb without pain or instability and functional range of movements that they had at the time of presentation with nonunion. Residual problems seen in some patients were joint stiffness, limb length discrepancy, and angular deformity. Twenty-six patients required repeat surgery using bone grafting because no satisfactory progress of fracture healing was seen in 4 months. Complications were related to the iliac crest donor site and persistent infection at the nonunion site.     

Percutaneous autologous bone marrow transplantation for nonunion of the femur.

Percutaneous bone marrow injections were performed on 7 nonunions of the femur. There were 6 hypervascular nonunions and one avascular nonunion. Two nonunions presented with active infections. One other patient had a history of infection which had subsided. One nonunion received the injection twice. After the site of nonunion was curetted and the bone surface was scored, 150 ml of bone marrow aspirated from the iliac bone was injected. Complete union occurred in 4 patients within 9 months; all of them were uninfected hypervascular nonunions following intramedullary nail fixation. One nonunion with a bone defect united partially leaving a 1 x 1 cm defect. The two infected femoral nonunions failed to unite. The results show that percutaneous autologous bone marrow injection for femoral nonunions can be considered for uninfected hypervascular nonunions following intramedullary nail fixation. In these cases stimulation of healing processes of fracture leading to consolidation can be expected from bone marrow injection. However, femoral nonunion with an active infection and loss of fixation is considered to be a contraindication for this technique.     

The diaphyseal aseptic tibial nonunions after failed previous treatment options managed with the reamed intramedullary locking nail

BackgroundThe nonunion of open and closed tibial shaft fractures continues to be a common complication of fractures. Tibial nonunions constitute the majority of long bone nonunions seen by orthopaedic surgeons. In this article, we present our approach to the surgical treatment of noninfected tibial shaft nonunions.MethodsBetween 2008 and 2014, 33 patients with aseptic diaphyseal tibial nonunion was treated by reamed intramedullary nailing and were retrospectively reviewed. The initial fracture management consisted of external fixation (27 patients), plate fixation (2 patients) and cast treatment (4 patients). All patients, preoperatively, were evaluated for the signs of the infection, by the same protocol. There were 13 hypertrophic, 16 oligotrophic (atrophic) and 4 defect nonunions registered in our material. The primary goal was to perform a closed intramedullary nailing on antegrade manner. An open procedure was only unavoidable when implants had to be removed or an osteotomy had to be performed to improve the alignment. Functional rehabilitation was encouraged with the assistance of a physiotherapist early postoperative. Patients were examined regularly during followed-up for a minimum of 12 months period for clinical and radiological signs of union, infection, malunion, malalignment, limb shortening, and implant failure.ResultsThe time that elapsed from injury to intramedullary nailing ranged from 9 months to 48 months (mean 17 months).Open intramedullary nailing was unavoidable in 25 cases (75,75%), while closed nailing was performed in 8 patients (24,25%). Osteotomy or resection of the fibula was performed in 78,8% of the cases. All patients were followed up in average period of 2 years postoperative (range 1–4 years), and 31(93,9%) patients achieved a solid union within the first 8 months. Mean union time was 5±0.8 months. Complications included 2 (6,06%) patients, one with deep infection and another case with absence of bone healing. Anatomical alignment has been achieved in the majority of patients, 28 patients (84,8%). The additionally autogenous bone chips were added in 4 patients (12,1%) where cortical defect was greater than 50% of the bone circumference.ConclusionIn conclusion, a reamed intramedullary nail provides optimal conditions for stable fixation, good rotational control, adequate alignment, early weight-bearing and a high union rate of tibial non-unions. Reaming of the medullary canal with preservation of periosteal sleeve create the "breeding ground" for sound healing of tibial shaft nonunions. Additionally cancellous bone grafting is recommended only in the case of defect nonunion.     

Treatment of femoral diaphyseal nonunions

There have been conflicting reports regarding treatment of femoral diaphyseal nonunions using reamed intramedullary nailing. Although high union rates have been reported using this technique, not all orthopaedic surgeons have experienced the same success. A retrospective review of charts and radiographs of 39 adult patients with nonunions of the femoral diaphysis treated at the authors' institution with reamed intramedullary nailing, compression, and with and without interlocking was done. The time from the index procedure to nonunion repair was 4 to 75 months (average, 19 months). Fifteen patients had 18 procedures between the index operation and nonunion repair. The average followup from nonunion repair to the most recent examination was 22.5 months (range, 3-108 months) with a median of 15 months. At the last followup, the overall union rate was 74% after one procedure and 97% after two or more procedures. There were seven complications including two infections, one pulmonary embolus, one occurrence of a deep venous thrombosis, a hematoma, and one case of malrotation. The data support the use of antegrade reamed nailing as a successful technique for treatment of most femoral diaphyseal nonunions.     

Composite implant composed of hydroxyapatite and bone morphogenetic protein in the healing of a canine ulnar defect

BACKGROUND AND AIMS: Hydroxyapatite (HA) has been considered as a carrier material for bone morphogenetic proteins (BMP). The aim of this study was to evaluate the capacity of a composite implant of HA and native bovine BMP to heal a 2 cm segmental defect in the canine ulna. MATERIALS AND METHODS: A composite HA+BMP implant was compared with plain HA implants and cortical autografts. The fixation was accomplished with an intramedullary Kirschner wire. The bone union was evaluated by X-rays taken at operation and after 3, 6, 9, 12, 16, 25, 35 weeks and by histology and mechanical torsion tests. RESULTS: HA implants were not able to produce complete bone union even with BMP. There was some bridging between the implant and the bone in the defects treated with either plain HA or HA+BMP implant, the bridging being slightly more pronounced with HA+BMP. The autografts showed a significantly better capacity to heal the defect. The HA implant did not resorb markedly during the study. There was no significant difference in mechanical strength between the HA and HA+BMP groups. CONCLUSIONS: HA was not an adequate bone substitute material in this study model, and BMP was not able to enhance sufficiently the poor capacity of HA to heal canine ulnar defects.     

Vascularized free fibular transfer combined with autografting for the management of fracture nonunions associated with radiation therapy

BACKGROUND: The purpose of this study was to evaluate the functional results, rates of union, and complications associated with vascularized free fibular transfer combined with autografting for the treatment of nonunions in previously irradiated bone. METHODS: Seventeen patients who had had eighteen vascularized free fibular transfers combined with autografting for the treatment of nonunion of a fracture in previously irradiated bone were included in this study. There were eleven female patients and six male patients. Eight patients had a bone neoplasm and nine, a soft-tissue neoplasm. The diagnosis was Ewing sarcoma in four patients; lymphoma, malignant fibrous histiocytoma, and rhabdomyosarcoma in two patients each; and cavernous hemangioma, metastatic breast carcinoma, reticulum-cell sarcoma, myxosarcoma, hemangiopericytoma, and fibrosarcoma in one patient each. The remaining patient had a soft-tissue tumor for which the diagnosis was not known. All patients received radiation therapy. The average dose was 5564 centigray. There were no recurrent tumors. The average interval between the radiation therapy and the original fracture was 111 months. The fracture was in the femur in thirteen patients, in the humerus in three, and in the tibia in one. All patients had operative or nonoperative treatment, or both, of the initial fracture, and two had iliac-crest bone-grafting after the initial open reduction and internal fixation procedure. The ages of the patients ranged from thirteen to eighty-two years at the time of the vascularized free fibular transfer. All fibular transfers were applied as onlay grafts because no nonunion was associated with a large segmental defect. Cancellous autogenous bone graft from the iliac crest was used as an additional graft at the proximal and distal junctions of the graft with the bone and at the fracture site in all patients. The average duration of follow-up after the vascularized free fibular transfer was fifty-seven months (range, twenty-eight to 112 months). RESULTS: Sixteen of the eighteen fracture sites united, after an average of 9.4 months (range, three to twenty-four months). Thirteen patients had an excellent result, one had a good result, two had a fair result, and one had a failure of treatment. Four patients had an infection, including one who continued to have a nonunion. The other three patients had union after treatment with antibiotics, debridement, and removal of the hardware. Another patient who had a recalcitrant nonunion eventually required an above-the-knee amputation. CONCLUSIONS: On the basis of this review, we suggest that microvascular fibular transfer combined with autografting is an appropriate treatment option for difficult nonunions associated with previously irradiated bone.     

Treatment of femoral and tibial diaphyseal nonunions using reamed intramedullary nailing without bone graft

INTRODUCTION: Intramedullary nailing has long been used successfully in the treatment of aseptic nonunions of the femur and tibia. However, recently the efficacy of reamed intramedullary nailing in the treatment of nonunions of the femur has been questioned by some publications reporting unfavourable results. The purpose of this study is to evaluate the treatment results of femoral and tibial diaphyseal nonunions with intramedullary nailing. PATIENTS AND METHODS: We retrospectively reviewed thirty-two patients with femoral or tibial diaphyseal nonunions who were treated with reamed intramedullary nailing between May 2002 and April 2006. Fixation status at the time of treatment were nail in twenty-eight patients (12 femurs, 16 tibiae), plate in three cases (2 femurs, 1 tibia), no implant in one femur. We used a dynamically locked, reamed intramedullary nailing. Only in bone defects greater than 50% of the cortical diameter and more than 2 cm in length was open bone grafting performed. RESULTS: Solid union was achieved in 93% (fourteen of fifteen) of femoral nonunions and 94% (sixteen of seventeen) of tibial nonunions. CONCLUSIONS: Our protocol with a dynamically locked, reamed nailing with the use of an oval hole and no open bone grafting for a defect less than 50% of the diameter and immediate weight bearing was successful in the treatment of femoral and tibial diaphyseal nonunions.     

Operative treatment of olecranon nonunion.

Records of five patients treated surgically for nonunion of the olecranon were reviewed. Four of the five fractures leading to nonunion were comminuted or oblique. Three nonunions occurred after tension band wiring, one nonunion occurred after open reduction internal fixation with a semitubular plate, and one nonunion occurred after treatment with a cast. The median interval from fracture to treatment of nonunion was 8 months. All nonunions were treated surgically. Four patients were treated with a tension band plate technique. All nonunions united at a median of 3 months. The median follow-up period was 36 months (range, 12-48 months).     

Management of proximal scaphoid bone pseudarthroses and fractures with the mini-Herbert screw via a dorsal approach

32 patients, aged 16 to 49 years, were treated by osteosynthesis using the Herbert mini screw. The indications were five fractures (type B3, Herbert classification), two delayed unions (type C) and 25 nonunions (type D1 to D3, Filan and Herbert classification). Six patients received no bone grafts, 19 received cancellous bone graft from the radius. An interpositional iliac crest bone graft was used in four, and a vascularized bone graft from the distal radius in three cases respectively. The average postoperative immobilisation in a forearm splint was nine weeks. 26 patients could be recruited for clinical follow-up at an average of 14,5 months. The radiological results were assessed in 30 cases (94%). Bony consolidation was achieved in 26 cases (100% of the fractures, 84% of the nonunions). In three cases a loosening of the screw, and in three further cases a dislocation into the radiocarpal joint were observed. A humpback deformity was present in four cases. Three patients showed a persisting nonunion, one patient a fibrous union. Early degenerative changes of the radiocarpal joint were observed in six cases. The clinical follow up examination showed an average grip strength of 91% (JAMAR II), 94% for the three finger, and 95% for the pinch grip compared to the contralateral side. The mean postoperative pain score on the visual analog scale was one for resting conditions, eleven for motion and 33 under stress. The range of motion was 79% of the opposite side for extension/flexion and 83% for radial/ulnar deviation. The average DASH-score reached 15 points. The Herbert mini screw has proven to be a reliable implant for reconstruction of proximal pole fractures and nonunions of the scaphoid.     

Osteogenic protein-1 (bone morphogenic protein-7) combined with various adjuncts in the treatment of humeral diaphyseal nonunions

A prospective study was conducted to determine the efficacy of using recombinant BMP-7 (rhOP-1) as an adjuvant in the treatment of diaphyseal humeral nonunions. Twenty-three consecutive patients with atrophic humeral diaphyseal nonunions were treated at seven separate institutions. All nonunions were fixed with either a compression plate or an intramedullary nail in conjunction with various bone grafting techniques. Recombinant OP-1 was delivered to the fracture site in a Type I collagen carrier at the time of fixation. All fractures went on to eventual union. There were no serious complications and no adverse reactions to the rhOP-I implant. Our study suggests that rhOP-1 may be a safe and effective adjuvant for the treatment of humeral diaphyseal nonunions.     

Bone morphogenetic protein but not transforming growth factor-beta enhances bone formation in canine diaphyseal nonunions implanted with a biodegradable composite polymer

BACKGROUND: The purpose of the present study was to create an effective bone-graft substitute for the treatment of a diaphyseal nonunion. METHODS: A standardized nonunion was established in the midportion of the radial diaphysis in thirty mongrel dogs by creating a three-millimeter segmental bone defect (at least 2 percent of the total length of the bone). The nonunion was treated with implantation of a carrier comprised of poly(DL-lactic acid) and polyglycolic acid copolymer (50:50 polylactic acid-polyglycolic acid [PLG50]) containing canine purified bone morphogenetic protein (BMP) or recombinant human transforming growth factor-beta (TGF-beta1), or both, or the carrier without BMP or TGF-beta1. Five groups, consisting of six dogs each, were treated with implantation of the carrier alone, implantation of the carrier with fifteen milligrams of BMP, implantation of the carrier with 1.5 milligrams of BMP, implantation of the carrier with fifteen milligrams of BMP and ten nanograms of TGF-beta1, or implantation of the carrier with ten nanograms of TGF-beta1. At twelve weeks after implantation, the radii were examined radiographically and the sites of nonunion were examined histomorphometrically. RESULTS: We found that implantation of the polylactic acid-polyglycolic acid carrier alone or in combination with ten nanograms of TGF-beta1 failed to induce significant radiographic or histomorphometric evidence of healing at the site of the nonunion. The radii treated with the carrier enriched with either 1.5 or fifteen milligrams of BMP showed significantly increased periosteal and endosteal bone formation on histomorphometric (p < 0.05) and radiographic (p < 0.02) analysis. CONCLUSIONS: Bone formation in a persistent osseous defect that is similar to an ununited diaphyseal fracture is increased when species-specific BMP incorporated into a polylactic acid-polyglycolic acid carrier is implanted at the site of the nonunion. TGF-beta1 at a dose of ten nanograms per implant did not induce a similar degree of bone formation or potentiate the effect of BMP in this model. CLINICAL RELEVANCE: The biodegradable implant containing BMP that was used in the present study to treat diaphyseal nonunion is an effective bone-graft substitute.     

Augmentation of fracture healing by hydroxyapatite/collagen paste and bone morphogenetic protein‐2 evaluated using a rat femur osteotomy model

In fracture treatment, biological bone union generally depends on the bone's natural fracture healing capacity, even in surgically treated cases. Hydroxyapatite/collagen composite (HAp/Col) has high osteoconductivity and stimulates osteogenic progenitors. Furthermore, it has the potent capacity to adsorb bone morphogenetic proteins (BMPs). In this study, we prepared an injectable HAp/Col paste and evaluated its augmentation of bone union. Furthermore, the effect of HAp/Col paste combined with BMP‐2 was also evaluated. We used a rat femur osteotomy model with a defect size of 1 mm. Male Wistar rats were assigned to one of the following four groups; a control group without any implant, a HAp/Col implant group, a group that received an absorbable collagen sponge (ACS) implant impregnated with BMP‐2 (1 μg), and a group that received a HAp/Col implant impregnated with BMP‐2 implant. Micro‐CT analysis, three‐point bending tests, and histological evaluation were performed. Bone union was achieved in two of eight cases in the HAp/Col group, five of eight cases in the ACS + BMP‐2 group, and all cases in the HAp/Col + BMP‐2 group at 8 weeks post‐surgery. The control group did not achieve bone union. In addition, in the HAp/Col + BMP‐2 group, the biomechanical strength of the fused femurs was comparable to that of the contralateral intact femur; the ratio of the mechanical load at the breaking point of the osteotomy side relative to that of the contralateral side was 1.00 ± 0.151 (SD). These results indicate that HAp/Col paste with or without BMP‐2 augments bone union. © 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 36:129–137, 2018.