Nonunion is a potentially devastating outcome after bone fracture. In some cases, multiple surgical procedures are required to achieve bony healing. Approximately 10% of acute long bone fractures have delayed union or progress to nonunion.1–4 The development of nonunion after long bone fracture depends primarily on 2 factors: mechanical stability and the biologic milieu at the time of injury and healing. Inadequate fixation at the fracture site leads to instability, and associated motion strains and shears healing tissues. Biologic factors, such as complex fracture patterns, significant comminution, interfragmentary gap, bone loss, poor soft tissue coverage, and vascular deficiency, as well as patient comorbidites (eg, diabetes) and social habits (eg, tobacco use), also play an important role in the healing process.2,5
The use of bone graft to augment biologic healing is a fairly common practice among orthopedists. Current surgical trends exist for graft implantation during the management of bony defects, fractures with severe comminution, and fracture nonunion.5–7 The innate biology of bone graft plays a crucial role in its efficacy as a surgical adjunct. At the core is the presence of mesenchymal stem cells, bone morphogenetic proteins, and growth factors.3,8 Several bone grafts and bone graft substitutes are available, although the gold standard in use is autogenous bone graft.3,4 Autogenous bone graft is histocompatible with osteogenic, osteoinductive, and osteoconductive potentials, and these properties make it an ideal complement for bone regeneration.9,10 Autogenous bone grafting also allows for revascularization and migration of bone that is forming and resorbing cells into remodeling tissues.11,12 The graft donor site that is most commonly harvested is the iliac crest, but procurement cannot be accomplished without some increase in morbidity (pain, blood loss, increased operative time, and risk of infection).13,14 The incidence of morbidity related to bone graft has been reported in up to 22% of cases, with 12-month donor site pain reported in up to 45% of patients.15–17
Natural variations exist in gene expression levels and the ultimate concentrations of mesenchymal stem cells, growth factors, and bone morphogenetic proteins among bone graft donors and donor sites.18,19 These variations may suggest individual differences in the healing potential of autogenous bone graft.20 Because of these disparities, attention has been turned to the development of synthetic implants and graft adjuvants as potential methods to enhance fracture healing. Studies have shown that synthetic bone morphogenetic proteins can increase fracture healing rates in long bone fractures21 and can be safely used for fracture nonunion22 and spinal fusion.23 Recombinant human bone morphogenetic protein 2 (rhBMP-2; INFUSE, Medtronic, Memphis, Tennessee) has also been shown in a prospective trial to improve healing rates in patients with acute tibia fractures.21 Many surgeons use synthetic implants in an “off-label” fashion for theoretical enhancement of bony healing and potential for progression to bony union. A search of the literature showed that no prospective cohort study has evaluated the effect of rhBMP-2 as an adjunct to autogenous bone graft in the treatment of fracture nonunion. The current study compared patients treated with autogenous bone graft alone with those receiving autogenous bone graft plus rhBMP-2. Within this set of patients, the current study was conducted to determine whether a synthetic bone morphogenetic protein implant works synergistically with autogenous bone graft, leading to enhanced fracture healing and improved outcomes after treatment of long bone nonunion.
Materials and Methods
The research protocol and all procedures followed were in accordance with the ethical standards determined by the institutional review board. Written informed consent was obtained from all participants. Between 2004 and 2012, 288 consecutive patients with long bone nonunion who presented to the offices of 2 orthopedists were enrolled in a prospective research registry. These patients were treated with either surgical revision, revision with autogenous bone graft implant, or autogenous bone graft with synthetic implant adjunct. A chart review was performed to compare patients treated with autogenous bone graft alone with those treated with autogenous bone graft plus rhBMP-2. Patients with known infection at the time of consultation were excluded from the study.
The authors identified 118 patients as meeting the inclusion criteria; 68 had received rhBMP-2 plus autogenous bone graft, whereas 50 had received autogenous bone graft alone. Patients were recruited from the offices of 2 fellowship-trained surgeons who perform procedures at several locations. The availability of a synthetic implant adjunct depended on the location of surgery within the multicenter university hospital system. One hospital within the authors’ academic medical center has ready access to rhBMP-2, whereas another does not. The 2 hospitals draw distinct populations. Therefore, information on the hospital site and the availability of rhBMP-2 was used for treatment group stratification. In cases where the implant was provided, its use as an off-label indication was explained.
A thorough comparison of demographic variables was performed, including age; sex; marital status; employment; body mass index; tobacco, alcohol, or substance use; medical comorbidities; open or closed status of the initial injury; and energy of the mechanism of injury. The patient’s surgical history was evaluated, including the number of previous nonunion surgeries. The long bone type and the atrophic-hypertrophic nature of nonunion were evaluated. Clinical and radiographic assessments were performed at regular intervals (3, 6, and 12 months) after surgery. Surgical outcomes, including progression to healing, time to union, growth status of cultures obtained at the time of surgery, subsequent surgical revision or amputation, and functional assessment scores (using the validated Short Musculoskeletal Functional Assessment questionnaire [SMFA]),24 were compared.
Surgical Technique for Autogenous Graft Harvest and Placement
Standard exposure of the iliac crest was obtained with the patient in the supine position. The external oblique muscle was retracted medially. After subperiosteal elevation of the iliacus, a cortical window in the inner table was made with an osteotome. Approximately 15 to 30 cc of bone graft was harvested from between the inner and outer table of the anterior iliac crest of each patient. Hemostasis was obtained with bone wax. The graft site was closed in layers. The surgeon delivered 1.5 mg/mL of rhBMP-2 as an adjunct to 15 to 30 mL of autogenous bone graft (rhBMP-2 plus autogenous bone graft). After exposure and preparation of the nonunion site, the autogenous bone graft or rhBMP-2 adjunct (along with autogenous bone graft) was placed in and around the nonunion sites before final closure (Figure). No drains were used at the crest donor site.
Intraoperative images of a 57-year-old woman with an atrophic right humerus nonunion showing the fracture nonunion site (A) and following plate fixation and placement of an autogenous iliac crest graft alone (B). Intraoperative image of a 60-year-old woman treated for a left humerus nonunion showing that following plate fixation and iliac crest graft placement, an rhBMP-2–impregnated carrier is applied to the nonunion site (C).
Although participants were drawn from different socioeconomic areas within a large urban region, there were no differences between the rhBMP-2 plus autogenous bone graft and autogenous bone graft–only groups in the demographic characteristics analyzed (Table 1). Mean body mass index; rates of tobacco, alcohol, and substance use; and mean sum of medical comorbidities did not differ. Initial injuries were similar between groups in terms of the percentage of open injuries and the energy of the mechanism of injury (Table 1). Patients in the rhBMP-2 plus autogenous bone graft group had undergone a mean of 1.4 (±1.5) previous nonunion surgeries, and those in the autogenous bone graft–only group had undergone 1.2 (±1.4) previous surgeries (P=.53). Bony site involvement and the atrophic-hypertrophic nature of each fracture nonunion were similar (Table 2). Nonunion fracture sites included the forearm, clavicle, femur, humerus, tibia, and fibula in both groups.
Demographics and Injury
Data on bony healing and progression to union were available for 115 of 118 (97.5%) patients. Rates of healing did not differ between groups and were 98.5% in the patients receiving rhBMP-2 plus autogenous bone graft and 95.9% in those receiving autogenous bone graft alone (P=.57) (Table 3). Mean time to union in the rhBMP-2 plus autogenous bone graft group was 6.6 months (±3.9), whereas it was 5.4 months (±2.7) in the autogenous bone graft–only group (P=.06). The latent healing in the rhBMP-2 group may be associated with an increased reoperation rate when compared with the autogenous bone graft–only group (16.2% vs 8%, respectively, P=.19). The reasons for revision included failure to progress to union in patients with a history of smoking, gross hardware failure, osteolytic changes, and unclear cause.
Healing, Complications, and Functional Outcomes
Surgical complication rates did not differ between treatment groups, although 1 patient in the autogenous bone graft–only group had persistent nonunion and underwent a below-the-knee amputation (Table 3). This patient experienced a 10-year nonunion, endured multiple previous surgeries, experienced multiple unsuccessful attempts at nonunion revision, and electively underwent amputation independent of bone graft–associated complications. The percentage of patients with positive culture findings (17.6% and 18% of the rhBMP-2 and autogenous bone graft–only groups, respectively) did not differ between groups (P=.96). All were treated with 6 weeks of intravenous antibiotics postoperatively. No gross infections were identified during follow-up. There were no complications related to the bone graft harvest site. Follow-up to 12 months was obtained for 100 patients (84.7%). Scores of pain and functional assessment at 12 months also showed no differences between groups (Table 3).
This study found that the use of rhBMP-2 as an adjunct to autogenous bone graft implant did not have a desired synergistic effect on the healing of long bone fracture nonunion. The use of autogenous bone graft alone resulted in a 92% single surgery success rate and an ultimate union rate of 96%. The single surgery success rate for rhBMP-2 was 84%, with ultimate progression to union occurring in 98.5% of patients. Furthermore, the addition of rhBMP-2 had no effect on healing time when used in combination with autologous iliac crest bone graft. No evidence of untoward effects or increase in complications was seen with the use of rhBMP-2; therefore, this study found its use as an implant adjunct to be safe.
Research by Urist25 in 1965 showed that bone morphogenetic proteins were essential in fracture repair. Since then, the use of synthetic bone morphogenetic proteins has expanded dramatically. However, use of synthetic graft comes at a cost, especially when coupled with the implantation of autologous bone graft. Cancellous chips (15 mL), which carry only osteoconductive properties, range in price from $211 to $297, depending on the vendor. Demineralized bone matrix, which has osteoconductive and osteoinductive properties, is almost five-fold more expensive than cancellous chips. These 2 options, however, are still more economical than a “small” (5 mL) rhBMP-2 INFUSE sponge, which costs approximately $3500.
Previous clinical studies on the use and success of rhBMP-2 after fracture and management of nonunion are limited. In 1999, the first clinical study of BMP-2 implantation showed its safe use and feasibility for application in the treatment of open tibial fracture.26 In 2002, the results of the BESTT study group (BMP-2 Evaluation in Surgery for Tibial Trauma) showed a significant reduction in secondary interventions, with shorter healing times, in patients with open tibial shaft fracture who were treated with BMP-2.21 Further studies of a subgroup of these patients with Gustilo-Anderson type IIIA and IIIB open injuries found that rhBMP-2 reduced the frequency of bone grafting and other secondary interventions required within this specific cohort.27 To the authors’ knowledge, the use of rhBMP-2 for revision of nonunion has yet to be evaluated by any prospective studies.
Grimshaw et al28 presented results from a retrospective study of 30 patients treated for atrophic nonunion of long bones after baseline open fractures. This study included patients with culture-positive infection at presentation. Patients were treated with rhBMP-2, rhBMP-7, or autogenous bone graft alone as part of their nonunion revision procedure. Rates of union did not differ between the groups (77% of all patients treated with bone morphogenetic proteins vs 75% of those treated with autogenous bone graft alone). Time to union was similar in both groups. The authors concluded that the use of rhBMP offered no clinical advantage over the use of autogenous bone graft alone in the treatment of atrophic nonunion of long bones.
Tressler et al7 retrospectively compared rhBMP-2 plus cancellous chips vs iliac crest bone graft alone in a small series of patients who underwent revision of long bone nonunion. They found no statistical differences in the rate of healing between the treatment groups. However, they saw healing in only 68.4% (13 of 19) of patients with rhBMP-2 plus cancellous chips (vs 85.1% of patients with iliac crest bone graft, P=.09). Based on these results, the authors concluded that rhBMP-2 may provide a suitable alternative to autologous iliac bone graft, and when combined with cancellous chips, it may offer the possible advantages of shorter operative time and reduced intraoperative blood loss compared with iliac crest bone graft harvesting.
The spine literature has reported conflicting results on the use of rhBMP-2 and outcomes after its application. In short, the review by Carragee et al29 of the use of rhBMP-2 in spinal surgery showed increased complication rates, including graft subsidence, infection, retrograde ejaculation, and vertebral osteolysis with poorer global outcomes. This study concluded that the risks and adverse events associated with the use of rhBMP-2 were equivalent to or greater than those associated with iliac crest bone harvesting. Other prospective randomized controlled studies have found that the use of rhBMP-2 for single-level posterior lumbar interbody arthrodesis offers no advantage in clinical outcomes compared with fusion performed with autologous bone graft alone.30,31
The current study has several limitations. In particular, there may be variations in the amount of autologous graft harvested because this information was not uniformly recorded and a dose-volume effect may have confounded the results. The quality of each patient’s bone graft also may vary. No attempt was made to analyze the quality of each individual harvest. The study sample size is moderate and includes a retrospective review of prospectively collected data in 2 patient populations. A large-scale randomized trial with predetermined nonunion types (and fracture patterns) would expand on the findings of this study. Because there was no algorithm for the use of rhBMP-2 during treatment, it is possible that differences in the patient population played a role in the results. However, the authors found no differences in the patients’ sociodemographic characteristics or the details of injury. This is likely related to the population of patients who have orthopedic trauma, regardless of the ultimate follow-up site.
Fracture nonunion is a challenging complication after relatively common injuries. Many independent variables can account for fracture nonunion, and it is impossible to control for subtle differences that are inherent to the treatment of each fracture type. Future studies would require multicenter organization and randomization to control for preexisting factors related to fracture nonunion. Ultimately, this study found rhBMP-2 to be a safe adjuvant to autologous iliac crest bone graft. However, rhBMP-2 did not provide a synergistic effect when used together with autogenous bone graft. Given the high cost of this product, the use of rhBMP-2 may not be warranted in the management of fracture nonunion when autologous iliac crest bone graft is available and implanted. Further conclusions about the benefits of the use of rhBMP-2 in the absence of autogenous bone graft cannot be made based on the findings of this study.
- Blick SS, Brumback RJ, Lakatos R, Poka A, Burgess AR. Early prophylactic bone grafting of high-energy tibial fractures. Clin Orthop Relat Res. 1989; 240:21–41.
- Lynch JR, Taitsman LA, Barei DP, Nork SE. Femoral nonunion: risk factors and treatment options. J Am Acad Orthop Surg. 2008; 16:88–97.
- Schmidmaier G, Schwabe P, Wildemann B, Haas N. Use of bone morphogenetic proteins for treatment of non-unions and future perspectives. Injury. 2007; 38(suppl 4):S35–S41. doi:10.1016/S0020-1383(08)70007-X [CrossRef]
- Sen MK, Miclau T. Autologous iliac crest bone graft: should it still be the gold standard for treating nonunions?Injury. 2007; 38(suppl 1):S75–S80. doi:10.1016/j.injury.2007.02.012 [CrossRef]
- Borrelli J Jr, Prickett WD, Ricci WM. Treatment of nonunions and osseous defects with bone graft and calcium sulfate. Clin Orthop Relat Res. 2003; 411:245–254. doi:10.1097/01.blo.0000069893.31220.6f [CrossRef]
- Rajan GP, Fornaro J, Trentz O, Zellweger R. Cancellous allograft versus autologous bone grafting for repair of comminuted distal radius fractures: a prospective, randomized trial. J Trauma. 2006; 60:1322–1329. doi:10.1097/01.ta.0000195977.18035.40 [CrossRef]
- Tressler MA, Richards JE, Sofianos D, Comrie FK, Kregor PJ, Obremskey WT. Bone morphogenetic protein-2 compared to autologous iliac crest bone graft in the treatment of long bone nonunion. Orthopedics. 2011; 34:e877–e884.
- Tsiridis E, Upadhyay N, Giannoudis P. Molecular aspects of fracture healing: which are the important molecules?Injury. 2007; 38(suppl 1):S11–S25. doi:10.1016/j.injury.2007.02.006 [CrossRef]
- Bauer TW, Muschler GF. Bone graft materials: an overview of the basic science. Clin Orthop Relat Res. 2000; 371:10–27. doi:10.1097/00003086-200002000-00003 [CrossRef]
- Samartzis D, Shen FH, Goldberg EJ, An HS. Is autograft the gold standard in achieving radiographic fusion in one-level anterior cervical discectomy and fusion with rigid anterior plate fixation?Spine (Phila Pa 1976). 2005; 30:1756–1761. doi:10.1097/01.brs.0000172148.86756.ce [CrossRef]
- Khan SN, Cammisa FP Jr, Sandhu HS, Diwan AD, Girardi FP, Lane JM. The biology of bone grafting. J Am Acad Orthop Surg. 2005; 13:77–86.
- Stevenson S, Emery SE, Goldberg VM. Factors affecting bone graft incorporation. Clin Orthop Relat Res. 1996; 324:66–74. doi:10.1097/00003086-199603000-00009 [CrossRef]
- Arrington ED, Smith WJ, Chambers HG, Bucknell AL, Davino NA. Complications of iliac crest bone graft harvesting. Clin Orthop. 1996; 329:300–309. doi:10.1097/00003086-199608000-00037 [CrossRef]
- Ebraheim NA, Elgafy H, Xu R. Bone-graft harvesting from iliac and fibular donor sites: techniques and complications. J Am Acad Orthop Surg. 2001; 9:210–218.
- Robertson PA, Wray AC. Natural history of posterior iliac crest bone graft donation for spinal surgery: a prospective analysis of morbidity. Spine (Phila Pa 1976). 2001; 26:1473–146. doi:10.1097/00007632-200107010-00018 [CrossRef]
- Younger EM, Chapman MW. Morbidity at bone graft donor sites. J Orthop Trauma. 1989; 3:192–195. doi:10.1097/00005131-198909000-00002 [CrossRef]
- Goulet JA, Senunas LE, DeSilva GL, Greenfield ML. Autogenous iliac crest bone graft: complications and functional assessment. Clin Orthop Relat Res. 1997; (339):76–81. doi:10.1097/00003086-199706000-00011 [CrossRef]
- D’Ippolito G, Schiller PC, Ricordi C, Roos BA, Howard GA. Age-related osteogenic potential of mesenchymal stromal stem cells from human vertebral bone marrow. J Bone Miner Res. 1999; 14:1115–1122. doi:10.1359/jbmr.19126.96.36.1995 [CrossRef]
- Takemoto RC, Fajardo M, Kirsch T, Egol KA. Quantitative assessment of the bone morphogenetic protein expression from alternate bone graft harvesting sites. J Orthop Trauma. 2010; 24:564–566. doi:10.1097/BOT.0b013e3181ed29a6 [CrossRef]
- Stolzing A, Jones E, McGonagle D, Scutt A. Age-related changes in human bone marrow-derived mesenchymal stem cells: consequences for cell therapies. Mech Ageing Dev. 2008; 129:163–173. doi:10.1016/j.mad.2007.12.002 [CrossRef]
- Govender S, Csimma C, Genant HK, et al. Recombinant human bone morphogenetic protein-2 for treatment of open tibial fractures a prospective, controlled, randomized study of four hundred and fifty patients. J Bone Joint Surg Am. 2002; 84:2123–2134.
- Bong MR, Capla EL, Egol KA, et al. Osteogenic protein-1 (bone morphogenic protein-7) combined with various adjuncts in the treatment of humeral diaphyseal nonunions. Bull Hosp Jt Dis. 2005; 63:20–23.
- Carlisle E, Fischgrund JS. Bone morphogenetic proteins for spinal fusion. Spine J. 2005; 5(6 suppl):240S–249S. doi:10.1016/j.spinee.2005.02.014 [CrossRef]
- Swiontkowski MF, Engelberg R, Martin DP, Agel J. Short musculoskeletal function assessment questionnaire: validity, reliability, and responsiveness. J Bone Joint Surg Am. 1999; 81:1245–1260.
- Urist MR. Bone: formation by autoinduction. Science. 1965; 150:893–899. doi:10.1126/science.150.3698.893 [CrossRef]
- Bax BE, Wozney JM, Ashhurst DE. Bone morphogenetic protein-2 increases the rate of callus formation after fracture of the rabbit tibia. Calcif Tissue Int. 1999; 65:83–89. doi:10.1007/s002239900662 [CrossRef]
- Swiontkowski MF, Aro HT, Donell S, et al. Recombinant human bone morphogenetic protein-2 in open tibial fractures: a subgroup analysis of data combined from two prospective randomized studies. J Bone Joint Surg Am. 2006; 88:1258–1265. doi:10.2106/JBJS.E.00499 [CrossRef]
- Grimshaw CS, Cannada LK, Watson JT. Atrophic non-union after open fracture: are bone morphogenetic proteins the answer to the problem? Presented at the Mid-America Orthopaedic Association Annual Meeting. , April 2011. , Tucson, Arizona. ; and the Annual Meeting of the American Academy of Orthopaedic Surgeons. , February 2012. , San Francisco, California. .
- Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. Spine J. 2011; 11:471–491. doi:10.1016/j.spinee.2011.04.023 [CrossRef]
- Chutkan NB. The effect of recombinant human bone morphogenetic protein-2 in single-level posterior lumbar interbody arthrodesis. Orthopedics. 2013;36:712–713.
- Michielsen J, Sys J, Rigaux A, Bertrand C. The effect of recombinant human bone morphogenetic protein-2 in single-level posterior lumbar interbody arthrodesis. J Bone Joint Surg Am. 2013; 95:873–880. doi:10.2106/JBJS.L.00137 [CrossRef]
Demographics and Injury
|Age, mean years (SD)
|Female sex, %
|Mean body mass index (SD)
|Tobacco use, %
|Alcohol use, %
|Substance use, %
|Medical comorbidities, mean sum (SD)
|Initial open injury, %
| Low energy, %
| High energy, %
|Previous nonunion surgeries, mean (SD)
| Forearm, %
| Clavicle, %
| Femur, %
| Humerus, %
| Tibia, %
| Foot-ankle, %
Healing, Complications, and Functional Outcomes
|Mean time to union, months (SD)
|Positive culture obtained during surgery, %
|Pain at 12 months (range, 0–10) (SD)
|Mean SMFA at 12 months (SD)
| Arm and hand