Key points

Bone grafting is a common component of foot and ankle surgery that requires successful osseous union. Nonunion, reconstruction, and arthrodesis procedures may pose specific challenges, including bony defects secondary to trauma, malunions, or previous surgery. The average nonunion rate is 10% for ankle arthrodesis and 16% for subtalar joint arthrodesis. Nonunion rates in foot and ankle arthrodesis literature range from 0% to 47% in complex and revision procedures.

In primary arthrodesis of the ankle and hindfoot, the following have been found to increase the risk of nonunion and other noninfectious complications:

• •
  

  Positive smoking history

  
  
• •
  

  Previous attempted fusion

  
  
• •
  

  Presence of avascular bone

  
  
• •
  

  Diabetes mellitus

  
  
• •
  

  Previous solid organ transplantation

  
  
• •
  

  Poor preoperative serum glucose control (>200 mg/dL)

In revision arthrodesis, neuropathy and prior revision attempts have been identified as statistically significant risk factors for nonunion.

Autologous bone grafting (ABG) is the gold standard because of its osteoconductive, osteoinductive, and osteogenic properties. The disadvantages of autograft include limitations in quantity, donor site morbidity, and infections and complications from donor site harvest. For example, one study quoted that 8.8% of patients undergoing autograft procedures have more clinically significant donor site pain (≥20 mm on the visual analogue scale [VAS]) 1 year postoperatively. A recent survey of orthopedic surgeons showed when considering graft for foot and ankle arthrodesis procedures, the strongest factors supporting the use of ABG were clinical or radiographic nonunion, avascular necrosis, smoking history, and evidence of potential for incongruous apposition of bone.

In foot and ankle surgery, many of these clinical scenarios are common, and the surgeon must weigh the advantages and disadvantages when choosing autograft, allograft, and/or orthobiologic bone graft substitute. Orthobiologic bone graft substitutes include cellular bone allograft with mesenchymal stem cells (CBA with MSCs), platelet-derived growth factor (PDGF), platelet-rich plasma (PRP), bone morphogenetic proteins (BMPs), and fetal tissues. This article reviews the various bone grafting and bone substitute options and presents available and recent evidence supporting their use in procedures requiring osseous healing in foot and ankle surgery.

Autologous bone grafting

ABG continues to be the gold standard because of its osteoinductive, osteoconductive, and osteogenic properties. Harvest from the calcaneus or distal tibial metaphysis can provide small amounts of autograft with minimal complexity added to the procedure. Ipsilateral iliac crest bone graft (ICBG), when harvested as a tricortical wedge, has the added benefit of improved osteoconduction or mechanical support, especially in the setting of an opening wedge osteotomy or large bony defects.

A recent logistic regression analysis of 159 foot and ankle studies from 1959 to 2012 concluded that there is a trend toward higher healing rates when using cancellous and structural ABG in foot and ankle surgery compared with allograft, but this was not statistically significant. However, 153 of 159 studies included in this analysis were retrospective case series. Retrospective case series have inherent methodological limitations and have potential for selection bias. In these retrospective case series, surgeons could have elected the use of their preferred graft for procedures they deemed more complex, or those that had an anticipated lower rate of union. By introducing this variable, the results in these studies may not give an accurate picture of treatment efficacy.

Surgeons may choose to use the reamer-irrigator-aspirator (RIA, DePuy Synthes, West Chester, PN, USA) to collect cancellous ABG from the patient’s femoral shaft. Although there is strong evidence supporting RIA use in the orthopedic trauma literature, there is a paucity of data supporting its use in foot and ankle surgery. In 2014, a retrospective study was performed comparing clinical and radiographic outcomes in patients undergoing tibiotalar arthrodesis. When compared with ICBG, use of the RIA showed significantly lower nonunion rates. Furthermore, no patient undergoing RIA had chronic pain at the harvest site compared with 2 in the ICBG group. Length of stay and radiographic fusion were similar in both groups. The RIA may be a viable alternative to ICBG, especially in terms of reducing nonunion rates and donor site morbidity.

Autologous bone grafting

ABG continues to be the gold standard because of its osteoinductive, osteoconductive, and osteogenic properties. Harvest from the calcaneus or distal tibial metaphysis can provide small amounts of autograft with minimal complexity added to the procedure. Ipsilateral iliac crest bone graft (ICBG), when harvested as a tricortical wedge, has the added benefit of improved osteoconduction or mechanical support, especially in the setting of an opening wedge osteotomy or large bony defects.

A recent logistic regression analysis of 159 foot and ankle studies from 1959 to 2012 concluded that there is a trend toward higher healing rates when using cancellous and structural ABG in foot and ankle surgery compared with allograft, but this was not statistically significant. However, 153 of 159 studies included in this analysis were retrospective case series. Retrospective case series have inherent methodological limitations and have potential for selection bias. In these retrospective case series, surgeons could have elected the use of their preferred graft for procedures they deemed more complex, or those that had an anticipated lower rate of union. By introducing this variable, the results in these studies may not give an accurate picture of treatment efficacy.

Surgeons may choose to use the reamer-irrigator-aspirator (RIA, DePuy Synthes, West Chester, PN, USA) to collect cancellous ABG from the patient’s femoral shaft. Although there is strong evidence supporting RIA use in the orthopedic trauma literature, there is a paucity of data supporting its use in foot and ankle surgery. In 2014, a retrospective study was performed comparing clinical and radiographic outcomes in patients undergoing tibiotalar arthrodesis. When compared with ICBG, use of the RIA showed significantly lower nonunion rates. Furthermore, no patient undergoing RIA had chronic pain at the harvest site compared with 2 in the ICBG group. Length of stay and radiographic fusion were similar in both groups. The RIA may be a viable alternative to ICBG, especially in terms of reducing nonunion rates and donor site morbidity.

Allograft

Allogenic bone grafts (allografts) are harvested from cadavers, avoiding the complications associated with ABG donor site harvest. Although allograft can have osteoconductive properties as does ABG, processing cadaver allograft tissue takes its toll. Although gamma-irradiation and heat sterilization processing are necessary to allow successful transfer of tissue from donor to host, these processes kill live bone cells and cause allograft to lose a significant amount of its osteogenic properties. Processing the graft limits cell viability, which increases osteoblast apoptosis. Graft Processing also destroys other cells that produce cytokines, bone morphogenic proteins, which decreases the osteogenic and osteoinductive properties of the graft. Demineralized bone matrix (DBM) is a form of allograft prepared by acid extraction so it retains BMPs and bone collagens. Therefore, DBM has improved osteoinductive capacity compared with traditional allograft, because it retains more bone morphogenic proteins and bone collagens.

Cellular bone allograft containing mesenchymal stem cells

CBA containing MSCs is a biologic allograft alternative to traditional ABG and other bone graft substitutes. Like ABG, cadaveric CBA with MSCs have osteoconduction, osteoinduction, and osteogenesis properties. Human undifferentiated MSCs and MSCs differentiated into bone, cartilage, and adipose escape a host’s immune system because they express HLA class I, and not HLA class II. Therefore, MSCs from an allogenic or cadaveric source avoids the host’s cell-mediated immune response by avoiding the T cells and lymphocyte cell response. Allograft tissue is harvested to preserve living MSCs or osteoprogenitor cells. Processing and specifications of different CBA with MSCs products vary by company and product.

Jones and colleagues conducted a prospective, multicenter trial of cryopreserved CBA with MSCs (Trinity Evolution; Orthofix, Inc, Lewisville, TX, USA) in patients undergoing ankle and/or hindfoot arthrodesis. Trinity Evolution describes a screening process whereby only 3% of cadaveric donors are used and cryopreserves tissue at −185°C. Orthofix assures a minimum of 250,000 living cells per cubic centimeter, 50,000 of which are MSCs or osteoprogenitor cells. In Trinity Evolution, demineralized cortical bone from the same donor is added to the cancellous bone with cellular component. Jones and colleagues enrolled 103 patients with 171 joint arthrodeses; 76 patients with 129 arthrodeses completed the 1-year follow-up. At their primary endpoint 6 months postoperatively, 63 of 92 (69%) patients and 124 of 153 (81%) joints achieved fusion, assessed by plain radiographs and computed tomographic (CT) scan. At 1 year postoperatively, 54 of 76 (71%) patients and 112 of 129 joints (87%) achieved fusion. Patients experienced a statistically significant improvement in clinical outcomes, demonstrated by improved American Orthopaedic Foot and Ankle Society ankle-hindfoot scores, short form-36 physical component score, and VAS compared with baseline.

Anderson and associates performed a retrospective cohort study on 109 consecutive ankle fusions with adjunctive CBA with MSCs compared with proximal tibia ABG. After exclusions, 85 patients remained; 44 patients received CBA with MSCs and 41 patients received proximal tibia bone ABG. The investigators reported that the choice of CBA with MSCs or proximal tibia bone ABG used was determined by patient preference, amount of graft required, and availability of CBA with MSCs. This study did not comment on the specific personal or technical reasons CBA with MSCs or proximal tibia bone ABG was chosen, therefore possibly allowing for selection bias. In addition, the same product of CBA with MSCs was not always used. Anderson and associates published 84% union rate with CBA with MSCs and 95% union rate with ABG, but the difference was not statistically significant ( P = .158). However, the group treated with CBA with MSCs had an average time to fusion of 13.0 ± 2.1 weeks compared with 11.0 ± 2.8 weeks in the ABG group ( P ≤.001). There were no statistically significant differences in complications or patient satisfaction, and no complications occurred related to the proximal tibia ABG donor site.

Protzman and colleagues, Hollawel, Scott and Hyer, and Rush and colleagues each published smaller studies of 25 patients or less examining MSC bone allograft in foot and ankle fusion. These level IV studies demonstrated good and equivocal outcomes. The investigators also published that the CBA with MSCs product was safe and efficacious in foot and ankle arthrodesis.

Bone marrow aspirate

Bone marrow aspirate (BMA) or its concentrate, bone marrow concentrate (BMC), has been used to promote bone repair by delivering pluripotent MSCs to the surgical site while limiting the donor site morbidity associated with ABG. BMA harvest is less invasive than ICBG harvest, and it can be collected under local anesthetic. BMA can enhance bone healing by transplanting stem cells that differentiate into osteoblasts, chondrocytes, and other connective tissue cells.

BMA is often obtained from the ipsilateral iliac crest, long bone metaphyses, or calcaneus. BMA is concentrated to BMC via centrifugation. However, quantitative analysis performed on BMA reveals that some harvest locations have higher concentrations than others. Aspirate obtained from the anterior iliac crest had much higher concentrations of osteoblastic progenitor cells when compared with the aspirate collected from the tibia or calcaneus. Another study comparing the effect of different concentrations of BMC with distal tibia atrophic nonunions found BMC with higher concentrations of progenitor cells were associated with higher rates of bone healing.

Braly and colleagues presented case series evidence showing that BMA is an inexpensive and less invasive therapy for patients with distal tibial metaphyseal nonunions in the setting of stable retained hardware. The patients in this study displayed significant clinical and radiographic improvement from their preintervention state, and 9 of 11 patients attained bony union within 6 months of BMA intervention.

In a case series of zone II and III fifth metatarsal fractures in competitive athletes, Murawski and Kennedy showed that percutaneous screw fixation along with BMC yielded more predictable results in terms of fracture healing. Their mean time to radiographic fracture healing in the BMC treatment group was 5 weeks, and only 2 patients of 26 did not return to their previous level of sporting activity.

BMA has also been useful in the care of diabetic Charcot arthropathy. In 2009, Pinzur presented a case series of 44 patients. Surgical treatment of 46 feet was performed with the use of BMA mixed with PRP. Most of these patients had open wounds with chronic draining osteomyelitis. Forty-two of 46 feet had radiographic evidence of bony fusion at 16 weeks postoperative. These 42 feet also had no drainage or evidence of infection at an average of 26 months after surgery. It is thought that BMA provides the pluripotent cells, and PRP adds cytokines that stimulate healing. PRP will be discussed in further detail later.

Platelet-derived growth factor

PDGF is a polypeptide growth factor that plays a role in embryogenesis, angiogenesis, and osteogenesis. PDGFs are a family of growth factors, classified in groups A, B, C, and D, that forms heterodimers and homodimers. PDGF is released by platelets and macrophages at the sites of injury, and sites of blood clot and early bone healing. These PDGFs and cytokines have chemoattractant and mitogenic effects, thereby increasing the concentration of additional neutrophil and macrophages to sites of healing. As a result, PDGFs help advance the healing cascade in endochondral ossification.

The homodimeric form of PDGF subunit B is PDGF-BB. PDGF-BB’s role in formation, repair, and regeneration of bone has been postulated in the mesengenic process, involving MSCs. PDGF-BB is involved in osteoblastic processes by several mechanisms, including the following:

• 1.
  

  Stimulation of vascular endothelial growth factors (VEGF) by pericytes,

  
  
• 2.
  

  Liberating and activating pericytes to functional MSCs,

  
  
• 3.
  

  Mitogenic effects for pericytes and MSCs,

  
  
• 4.
  

  Affecting other factors involved in osteogenesis, including BMPs,

  
  
• 5.
  

  Additional complex communication and coordination functions with PDGF-receptors, MSCs, pericytes, endothelial cells and other signaling pathways.

Another mechanism PDGF-BB indirectly affects osteogenesis is through its role in angiogenesis. PDGF-BB, with the synergistic effect of fibroblastic growth factor-2 (FGF-2), is potently angiogenic and has vascular stabilization effects; this has been demonstrated in the cornea and in hindfoot ischemia of rodents.

Preclinical studies have demonstrated therapeutic value of recombinant human PDGF-BB (rhPDGF-BB) in animals before human use. Initially, rhPDGF-BB benefits were demonstrated in periodontal defects in dogs. Subsequently, rhPDGF-BB studies were performed in osteoporotic fracture models in geriatric rats, fracture healing models in diabetic rats, and distraction osteogenesis models in the femur of rats.

RhPDGF-BB has been studied in several areas of orthopedic surgery, including foot and ankle surgery. RhPDGF-BB is US Food and Drug Administration (FDA) approved for foot and ankle uses and is sold as Regranex (Smith & Nephew, Memphis, TN, USA), which is approved for wound healing in ulcers, Augment Bone Graft (BioMimetic Therapeutics, Franklin, TN, USA), which is approved for foot and ankle arthrodesis, and Augment Injectable Bone Graft (BioMimetic Therapeutics), which is approved for foot and ankle. Augment Injectable Bone Graft is approved for foot and ankle arthrodesis in Canada and is under investigation for this use in the United States.

RhPDGF-BB has been used in combination with β-tricalcium phosphate (β-TCP) carrier. β-TCP is a ceramic bone graft substitute with osteoconductive properties that has been used in orthopedic surgeries before the advent of PDGF. RhPDGF-BB with β-TCP (rhPDGH/β-TCP) is approved for foot and ankle arthrodesis in the United States, Canada, Australia, and New Zealand (Augment Bone Graft; BioMimetic Therapeutics). Before approval of rhPDGF-BB/β-TCP in foot and ankle surgery, rhPDGF-BB/β-TCP was demonstrated in a randomized controlled multicenter trial to be beneficial in periodontal defect surgery. RhPDGF-BB/β-TCP has been compared with ABG in several studies described in later discussion.

Daniels and colleagues performed a prospective, open-label, Canadian multicenter trial using rhPDGF-BB/β-TCP (Augment Bone Graft; BioMimetic Therapeutics) in 60 patients undergoing midfoot, hindfoot, or ankle arthrodesis. This pilot trial followed 60 patients, 59 of which completed the study to 36 weeks (9 months) with clinical and radiographic assessments. Fifty-two of 59 (88%) patients achieved radiographic union during the study. Overall, 54 of 60 (90%) patients and 124 of 130 (95.4%) joints treated with rhPDGF-BB/β-TCP had clinical success within 12 months. Five of 6 of these patients that did not unite had midfoot nonunions, and the remaining 1 of 6 of the patients with nonunion belonged to the ankle/hindfoot group. Daniels and colleagues were able to demonstrate adequate safety, efficacy, and noninferiority of rhPDGF-BB/β-TCP in ankle, hindfoot, and midfoot fusion and set the stage for following randomized controlled trials.

In 2011, DiGiovanni and colleagues performed a prospective randomized controlled multicenter trial to assess the feasibility of rhPDGF-BB/β-TCP with ABG. This feasibility study enrolled 20 adult patients undergoing ankle or hindfoot fusion at a 2:1 ratio of rhPDGF-BB/β-TCP to ABG. Fourteen patients received rhPDGF-BB/β-TCP (Augment Bone Graft; BioMimetic Therapeutics) with rigid internal fixation compared with 6 patients who received ABG with rigid fixation of their ankle or hindfoot arthrodesis. Nineteen of 20 patients completed the 36-week study. At 36 weeks, 10 of 13 (77%) patients treated with rhPDGF-BB/β-TCP had achieved osseous union, determined by greater than 50% osseous bridging on CT scan. In contrast, 3 of 6 (50%) patients receiving ABG achieved osseous union. Two patients from the rhPDGF-BB/β-TCP group had a nonunion requiring revision. In addition, DiGiovanni and colleagues documented an increase in operative time by 26 minutes in the ABG group.

Following the feasibility trial, DiGiovanni and associates performed a prospective randomized controlled multicenter trial comparing rhPDGF-BB/β-TCP (Augment Bone Graft; BioMimetic Therapeutics) to ABG in patients receiving hindfoot or ankle arthrodesis. This study enrolled 434 patients in 37 centers in the United States and Canada, and 397 patients remained after disqualifying criteria (exclusions, medical issues). Again, the group used the 2:1 ratio, enrolling 260 patients (394 joints) into the rhPDGF-BB/β-TCP group and 137 patients (203 joints) in the ABG group. Following the groups for 52 weeks, DiGiovanni and associates demonstrated comparable fusion rates clinically and radiographically for the most part. Some differences were found in the all-joints analysis (n = 597); the CT-determined fusion rate at 24 week was 66.5% in the rhPDGF-BB/β-TCP group and 62.6% in the ABG group ( P <.001). As well, 3-aspect radiographic union at 52 weeks also demonstrated a difference of 48.5% union in the rhPDGF-BB/β-TCP group and 44.3% in the ABG group ( P <.001). Clinical healing and pain at the arthrodesis sites were comparable in both groups. ABG patients had more clinically significant donor site pain (≥20 mm on the VAS) at 24 and 52 weeks, respectively. In addition, 1 patient was hospitalized for donor site infection, and 1 patient had donor site cellulitis.

Notably, the investigators reported decreased radiographic union rates in patients receiving greater volumes of rhPDGF-BB/β-TCP graft. When separated into volumes of rhPDGF-BB/β-TCP used (1–3 cc, 4–6 cc, or 7–9 cc), patients receiving 7 to 9 cc had decreased radiographic healing rates, which were comparable to union rates when ABG was used. In a subsequent analysis, DiGiovanni and colleagues determined that adequate graft fill determined by thin-cut CT scans at 9 weeks is an important factor in successful union of ankle and hindfoot fusions.

In 2015, Daniels and associates performed a prospective randomized controlled multicenter study on rhPDGF-BB/β-TCP with type 1 bovine collagen matrix (rhPDGF-BB/β-TCP-collagen) (Augment Injectable Bone Graft; BioMimetic Therapeutics, Inc, now Wright Medical Technologies, Franklin, TN, USA). The injectable rhPDGF-BB in β-TCP-collagen carrier with cannula allows an option for easier handling. This study of 75 patients requiring ankle or hindfoot fusion enrolled 63 patients into the rhPDGF-BB/β-TCP-collagen group and 12 patients into the ABG control group. Daniels and associates used 142 historical control patients from another study by DiGiovanni and colleagues. The investigators documented a clinically significant improved CT-determined fusion rate in the rhPDGF-BB/β-TCP-collagen group of 84.1% compared with 67.0% in the ABG group at 24 weeks ( P <.001). Radiographic results generally either favored fusion rates in the rhPDGF-BB/β-TCP-collagen group compared with ABG or trended toward favoring the rhPDGF-BB/β-TCP-collagen group but did not reach statistical significance. In addition, Daniels and associates established that rhPDGF-BB/β-TCP-collagen patients achieved union more quickly, at 14.3 ± 8.9 weeks compared with 19.7 ± 11.5 weeks when ABG was used ( P <.001). Both groups had equivocal clinical union rates of 87.3% and 88.3% at 52 weeks with rhPDGF-BB/TCP-collagen and ABG, respectively. There was no statistically significant difference in therapeutic failures (nonunions, delayed union, further surgery). The rhPDGF-BB/transforming growth factor-β (TGF-β)-collagen group experienced improved clinical success rate (weight-bearing VAS and no secondary procedures), and no donor site pain and comorbidities.