Bone Grafting

Bone grafting is used extensively in human and veterinary medicine. It is commonly utilized for treatment of delayed unions or nonunions, filling of bone defects, or joint arthrodesis. It is estimated that 500,000 to 600,000 bone grafting procedures are performed in humans annually in the United States alone [1]. No accurate estimates exist for the number of bone grafts performed in veterinary medicine.

Types of Bone Grafts – Terminology

donor and recipient, on the type of bone harvested, and on the site of the application. Autografts are tissues that are transplanted from one site to another in the same individual. Because the genetic composition of the donor and recipient sites are identical, no immune response occurs. Cancellous bone autografts are the most commonly used bone grafts in veterinary medicine. Allografts are transplanted from one individual to another of the same species. Genetic differences between the donor and recipient may induce an immune reaction based on the recognition of foreign antigens from the graft by the host immune system. Most bone allografts require some form of processing or preservation prior to use. Xenografts are tissues transplanted from one individual to another of a different species. Xenografts often induce a more severe immune response.

Bone grafts are also defined according to the type of bone transplanted. Cancellous bone grafts are composed of the porous, highly cellular trabecular bone found in the long bone epiphyses and metaphyses as well as in the iliac wings and the sternebrae. The trabeculae are covered with a highly reactive endosteal membrane that includes osteoclasts, osteoblasts, and bone lining cells. The cavities between the trabeculae are filled with highly cellular bone marrow. Cortical bone grafts are composed of the dense, compact bone from the diaphyses of long bones. They are mechanically strong, but have a relatively low cellularity. Corticocancellous bone grafts are combinations of the two types of bone, which allow surgeons to take advantage of the benefits of both types of grafts. Ribs and the craniodorsal part of the iliac wings are the main sources of corticocancellous bone grafts in dogs. An allograft used with fresh cancellous bone or bone marrow is termed a composite graft. Vascularized bone autografts or allografts are harvested with their blood supply and microvascular anastomoses are necessary at the time of implantation. Osteochondral grafts include bone and articular cartilage.

A bone graft placed in its normal position (i.e., within bone tissue) is said to have been orthotopically transplanted. If the graft is transplanted to an abnormal anatomic site, it is said to have been heterotopically transplanted.

Function of Bone Grafts

An ideal bone graft would perform four functions: osteogenesis, osteoinduction, osteoconduction, and mechanical support [2,3]. However, depending on the type, the most common bone grafts are able to provide only one or more of these functions (Table 88-1).

Osteogenesis

Osteogenesis is defined as the creation of new bone by bone-forming cells regardless of their origin. This function may be performed by donor cells (osteoblast precursors, osteoblasts, or chondroblasts) that survived the transplantation (fresh autogenous grafts only [2]) or by host cells stimulated by the graft or the grafting process. Cancellous bone has a much greater osteogenic potential than does cortical bone, in part because its copious porosity allows for fast and intense revascularization that contributes to cell survival [4]. Furthermore, cancellous bone is filled with bone marrow, which contains mesenchymal stem cells able to differentiate into osteoprogenitor cell lines. Finally, the trabeculae provide a large surface covered by lining cells and active osteoblasts. The survival of osteogenic cells in a graft depends on the nourishment they receive directly after transfer [5]. Although the percentage of cells surviving transfer of a fresh cancellous autograft is unknown, it has been shown that, in vitro, approximately 65% of the cells may survive up to 3 hours when the graft is maintained at room temperature in a saline solution [6]. However, it has been estimated that the survival rate of osteogenic cells in vivo does not exceed 5% following transplantation [7]. With cancellous bone autografts, early osteogenesis is characterized by active new bone formation in direct contact with the transplanted graft within 1 week following transplantation.4 Similarly, with vascularized cortical grafts, more than 90% of the osteocytes may survive transplantation [8], with new bone formation beginning as early as 24 hours after surgery [9]. In contrast, with nonvascularized cortical grafts, osteogenesis begins with vascular ingrowth, which does not occur until the sixth day after transplantation [10].

Osteoinduction

Osteoinduction refers to the recruitment of mesenchymal and pluripotent osteoprogenitor cells which later differentiate into bone- or cartilage-forming cells at the implantation site. Differentiation of mesenchymal cells is regulated via growth factors such as bone morphogenetic proteins, transforming growth factor-beta, insulin-like growth factors, fibroblast growth factors, platelet-derived growth factors, tumor necrosis factor, prostaglandin E2, and other cytokines [11]. Although these factors are produced by live cells and are present in the bone matrix, living graft cells are considered more osteoinductive than bone matrix. Because of the trabecular structure, cancellous bone matrix proteins are more readily accessible than are those of cortical bone. Osteoinductive factors may also be released during graft resorption, which in turn contributes to graft incorporation.12 Despite the lack of evidence that inductive proteins and cytokines are active in autologous cancellous bone grafts, cancellous bone is widely believed to be osteoinductive [12].

Osteoconduction

Osteoconduction refers to the three-dimensional process of tissue ingrowth within the graft, which acts as a scaffold. This process involves the invasion of the graft framework by capillaries as well as the migration of host osteoblasts, osteoclasts, and mesenchymal cells [2,13]. In porous cancellous bone grafts, initial fibrovascular invasion occurs quickly and is followed by bone ingrowth, characterized by direct apposition of new bone on the graft trabeculae. During subsequent graft remodeling, necrotic trabeculae are progressively resorbed and replaced by new bone. Although growth factors, collagens, and noncollagenous proteins participate in the regulation of this process, the three-dimensional trabecular structure of cancellous bone is the primary determinant of the speed and completeness of graft incorporation[13]. Histologic differences observed during cancellous versus cortical graft incorporation illustrate this phenomenon. Indeed, dense cortical grafts must undergo a resorptive phase, with removal of necrotic Haversian systems, before fibrovascular ingrowth into resorptive pores or channels can occur [14]. The process of initial bone resorption followed by fibrovascular invasion and appositional new bone formation as seen with cortical bone graft incorporation has been termed “creeping substitution.” The differences between the incorporation of cancellous and cortical bone grafts suggest that the porosity and surface area of the pores directly influence the speed and rate of incorporation.

Mechanical Support

Mechanical or structural support describes the ability of a bone graft to act as a load-sharing space filler in the treatment of large bone defects, and its effectiveness depends on the nature of the graft. Although cancellous bone is an efficient space filler, it does not provide substantial structural support. In contrast, a cortical graft provides excellent early mechanical support. Importantly, the strength of a graft changes over time in response to the biomechanical environment. Local bone resorption and revascularization substantially increase porosity, causing nonvascularized cortical autografts to become weaker at the host-graft interface six weeks after grafting [14-16]. Conversely, vascularized cortical autografts do not undergo resorption and revascularization and, therefore, show superior strength six weeks postoperatively [15]. However, six months following transplantation, the strength of nonvascularized and vascularized cortical grafts is similar [15]. Regardless of the type, cortical grafts are unable to withstand substantial loads, and must be supported by adequate internal fixation during the incorporation process to prevent catastrophic failure.

Biology of Bone Graft Incorporation

The term incorporation is used to describe the biologic interactions between a bone graft and host tissue at the recipient site that result in bone formation with subsequent improvement in mechanical properties [17]. This process includes the host inflammatory reaction to the surgical trauma, the host inflammatory and/or immune reaction to the graft material, and the processes of cell proliferation, migration, differentiation, and revascularization, resulting in new bone formation and union between graft and host bone. The sequence of events leading to graft incorporation is relatively consistent, although some differences exist between incorporation of cancellous and cortical bone as well as vascularized and nonvascularized bone grafts [15,18-20]. The biologic events occurring at the graft site include: (i) hematoma formation; (ii) inflammation, migration, and proliferation of mesenchymal cells, and development of fibrovascular tissue in and around the graft; (iii) vascular invasion of the graft; (iv) resorption of graft surfaces; and (v) bone formation on the graft surface. The incorporation of nonvascularized bone grafts (cancellous or cortical) begins with the formation of a local hematoma. This hematoma is rich in growth factors and other cytokines and allows for the survival of transplanted cells near the trabecular or cortical surfaces up to one week after transplantation [21]. The associated inflammatory response leads to the formation of granulation tissue, which invades the grafted site and contributes to the revascularization of the graft. Revascularization and cell differentiation vary with the type of graft, surgical technique, mechanical environment, and quality of the recipient bed.

Cancellous grafts are revascularized faster than are cortical bone grafts. Histocompatibility and immunogenicity of the graft also play an important role in the rate of vascularization. Incompatible bone grafts induce an immune reaction that delays or inhibits vascularization of the graft [4]; a major mismatch may induce resorption of a cancellous bone allograft and replacement by fibrocartilage [22]. Aseptic and atraumatic surgical techniques improve the success rate of graft incorporation. Fluid exudation, infection, or devascularization of the recipient bed will delay or prevent revascularization of the graft. The mechanical environment of the graft site also has a profound effect on revascularization and cell differentiation. Although instability at the host-graft interface induces local shear strains that may impede vascular invasion of the graft, the effect of the mechanical environment on graft incorporation varies between cancellous and cortical grafts. Cancellous grafts are less susceptible to shear strains owing to their porosity and use as loosely packed fragments. In contrast, the host-graft interface of dense, monolithic cortical bone grafts is highly vulnerable to instability because the proximity of the host and graft fragments induces large local shear strains that prevent vascular penetration of the narrow Haversian canals [17]. Accordingly, cancellous grafts do not require absolute stability, but cortical grafts must be rigidly stabilized. The quality of the tissues at the host site is also a primary factor in the incorporation process. Of particular importance is the vascularity in the graft bed and the abundance and competence of endothelial cell progenitors and connective tissue [17]. Excessive surgical trauma, the presence of necrotic tissue, systemic steroids, infection, and radiation therapy may compromise vascularization of the recipient bed and the pool of local progenitor cells [17,23].

Incorporation of Autografts

Cancellous Autografts

Although the rate of cell survival remains unknown, the majority of transplanted cells probably die as a result of local ischemia or induction of apoptosis. However, bone marrow mesenchymal cells and endothelial cell progenitors are relatively resistant to ischemia and may survive transplantation and may even proliferate in response to changes in oxygen tension, pH, and cytokines [17]. Survival of these cells along with recruitment of host mesenchymal stem cells is the key to successful incorporation of cancellous bone autografts. Following cancellous bone autograft transplantation, vascular invasion of intertrabecular spaces and formation of abundant granulation tissue proceed from the host-graft interface. This process is usually complete within two weeks [4]. With vascular ingrowth, differentiation of osteoprogenitor cells into osteogenic cells (osteoblasts) occurs. Osteoblasts start laying down poorly mineralized new bone (osteoid) on the surface of the necrotic trabeculae [24], while hematopoietic marrow elements accumulate within the graft [10]. By 12 weeks, cancellous bone chip grafts have the radiographic and histologic appearance of mature cancellous bone [4]. Eventually, the necrotic trabecular framework will be resorbed and remodeled via osteoclastic and osteoblastic activity. New bone may be remodeled in response to local stresses. Indeed, when cancellous bone grafts are used for filling of diaphyseal gaps, the remodeling process transforms cancellous bone into cortical bone. Osteogenic activity peaks around 6 to 8 weeks following transplantation, while “corticalization” of the graft in response to mechanical stresses requires up to 24 weeks [24].

Nonvascularized Cortical Autografts

Two major differences exist between cancellous and cortical graft incorporation: (i) the rate of revascularization and (ii) the process of new bone formation. Whereas fresh cancellous autografts are revascularized as early as 1 week after surgery, vascular invasion of cortical autograft is postponed until 7 to 8 weeks [25]. This delay is due to the structure of cortical bone, with vascular penetration of the graft requiring peripheral osteoclastic resorption and vascular infiltration of Volkmann and Haversian canals [14]. In experimental studies, the resorptive activity was confined almost exclusively to the osteons, whereas resorption of interstitial lamellae was uncommon. As the Haversian canals reached a critical size, the resorptive process ceased in favor of new bone apposition [10]. By 8 weeks, only a third of the graft had been replaced with new bone, whereas the incorporation front showed increased porosity and the remaining core of the graft appeared as a necrotic mineralized matrix [14,25]. At 1 year, 40% of the original necrotic bone remained in cortical autografts [14,19]. It is not known whether the failure to repair nonosteonic lamellar bone persists indefinitely. It seems likely that the mixture of necrotic and viable bone in the cortical graft remains unaltered once the anabolic and catabolic stages of repair have been completed. Thus, cancellous grafts tend to be completely replaced by new bone, whereas cortical grafts remain as mixtures of necrotic and viable bone [10]. While consolidation of the host-graft interfaces with woven bone stabilizes the graft and allows vascular penetration into the cortex [26], the delay between initial bone resorption and later bone deposition increases bone porosity at the incorporation front, locally reducing the strength of the graft by half between 6 and 24 weeks [14,25]. As a result, protection of the graft segment via rigid fixation is necessary for at least 6 to 12 months. Furthermore, in order to prevent graft site fractures, implant removal is not recommended.

Cortical bone autograft stabilization using rigid fixation favors its incorporation and influences the biologic events leading to its remodeling [27]. Lack of absolute stability at the host-graft interfaces delays graft incorporation. In a study using a canine ulnar diaphyseal segment model without stable internal fixation, only 22% percent were united at 12 weeks, 54% at 6 months and 66% at 1 year [28]. These results clearly contrast with another experimental study in which a cortical autograft was stabilized using internal fixation. By 3 months, both host-graft interfaces had healed [26]. Similarly another canine study showed that healing of a 4-centimeter femoral cortical autograft stabilized using rigid internal fixation was similar to bone healing under stable conditions [27]. New bone was deposited directly onto the surface of nonviable bone, and revascularization occurred from the recipient site medullary artery or a neonutrient artery traversing the cortical bone and reestablishing a medullary circulation. The graft was incorporated directly into the recipient site without local bone resorption and was used as a scaffold for new bone [27]. Incorporation, not resorption, was the prominent feature of the graft healing. Thus, resorption and replacement are not the fate of all devascularized cortical autografts.

Vascularized Cortical Autografts

A vascularized cortical graft involves bone that has been harvested with its nutrient artery and vein; the transplantation is obtained by transfer to another site with microvascular anastomosis of the two vessels. Incorporation of vascularized cortical autografts is achieved more predictably. Because the blood supply of the grafts is maintained, their incorporation may differ markedly from that of nonvascularized grafts. In particular, when the anastomosis is successful and the graft does not suffer intraoperative ischemia, up to 94% of osteocytes survive the transplantation procedure [8]. Such vascularized cortical grafts heal rapidly at the host-graft interface and their remodeling is similar to that of normal bone. Three months after surgery, vascularized grafts may be indistinguishable from normal bone [25]. Vascularized cortical grafts are stronger than nonvascularized grafts during the initial 6 weeks after transplantation because of the lack of resorption [15]. However, at 6 months, little difference exists in strength between vascularized and nonvascularized cortical autografts in the dog [15]. Although the vascularized cortical graft is not significantly weakened during the initial incorporation phase, appropriate internal fixation must be used until the graft can undergo hypertrophy in response to changes in local mechanical loading, according to Wolff’s law. Although vascularized grafts are not commonly used in veterinary medicine owing to the financial challenges associated with the acquisition and maintenance of operating microscopes and microvascular instrumentation, vascularized grafts are better than non vascularized grafts for large segmental bone defects [12].

Incorporation of Cortical Allografts

Bone allografts have been used in human and veterinary surgery for decades and the histologic events in allograft incorporation have been well described [9,10,16,18,20,26,29-35]. The main differences between cortical autografts and cortical allografts are the slowness of incorporation, the process of revascularization, and the biologic events resulting from the immunogenicity of the graft.

Incorporation of cortical allografts occurs via healing at the host-graft interface, which involves the gradual formation of a callus extending from the host bone to the surface of the allograft without local resorption [20,36]. Union of the host-graft interface is generally obtained within 3 months [26,31,37], although healing times may vary from 10 weeks to 37 weeks [36]. The principal determinants of host-graft union are stability of the construct and contact between host bone and the graft [38]. The process of revascularization and repair of the cortical allografts is controversial. In some studies, incorporation of cortical allografts followed a pattern similar to that seen with cortical autografts, albeit with slower and less extensive vascular penetration and slower new bone formation [5,29,32]. Revascularization and subsequent resorption were almost complete one year postoperatively, although a large amount of necrotic bone remained within the graft [32]. Other studies [20,26,27,33], however, suggest that the pattern of revascularization and repair differs from that observed for cortical autografts involving the deposition of a thin layer of new bone (1 to 2 mm thick) over the necrotic cortex of the graft that remained intact. Whereas the outer bone was supplied by extraosseous vessels, the medullary cavity was avascular [33]. The repair process was slow, with less than 10% of the graft being incorporated at one year and only 20% at 5 1/2 years [20]. Although anecdotal reports have shown that a femoral cortical allograft can remain structurally intact for eight years and that it can assume the mechanical support of limb function for five years after plate removal [33], from a mechanical standpoint, persistence of a large amount of necrotic bone in the allograft is likely to decrease strength in vivo over time. A 50% loss in allograft strength was noted after 10 years and was correlated with an increase in microfractures, probably related to mechanical stresses induced by weight bearing [39]. Because necrotic bone cannot repair itself, these microfractures may ultimately induce catastrophic allograft failure. In human medicine, the risk of these fractures (up to 19%) is significant in the third year after implantation [40-42].

Immunogenicity of bone allografts has been demonstrated in experimental studies. The immunologic response of the host is predominantly a cell-mediated response to cell surface antigens carried by cells in the allograft, i.e., major histocompatability complex (MHC) class I and class II antigens. MHC class I antigens are found on all nucleated cells of the body and MHC class II antigens are expressed on the surface of macrophage-myeloid lineage cells, as well as osteoblasts [43,44]. Accordingly, because bone marrow cells represent a significant source of MHC I and MHC II antigens, removal of bone marrow in allografts is likely to decrease their immunogenicity [45]. Although both collagen and matrix can induce immune responses, they are considered relatively weak when compared with cell-mediated immunogenicity.44 Evidence also exists that all types of allografts induce the production of graft-specific antibodies that can be detected as early as three weeks after transplantation [46,47]. Whereas the immune response depends on the MHC class I and class II disparities [47], its duration and intensity also depend on graft treatment and size, with frozen mismatched allografts inducing a weaker response than fresh allografts [46].

When the allografts are relatively small, the antidonor antibodies are detectable for a shorter period of time than when massive allografts are used [47]. Several studies on experimental animals have shown that the process of incorporation of bone allografts is negatively influenced by the degree of mismatch between major histocompatibility complex antigens [29,30,35]. However, even if antibodies are formed against bone allografts, no clear evidence exists that they are directly involved in the rejection process [10]. Despite the experimental evidence that allografts can induce a host immune reaction, the clinical significance of this reaction is unclear. In biopsy specimens obtained 9 to 78 months after transplantation of large frozen cortical allografts in humans, no clear relationship was demonstrated between the extent of graft incorporation and the degree of histocompatibility between graft and host [48].

Although most cortical allografts heal via intramembranous bone formation, clinical and experimental studies have shown that incorporation and remodeling may vary extensively. Incorporation may involve complete revascularization, resorption, and repair [29,32], or leave a large amount of necrotic bone, with biologic attachment of soft tissues to the allograft surface and good functional results for several years [20,33]. Similarly, remodeling may occur in response to mechanical stresses, according to Wolff’s law [9,49].

Incorporation of Processed Cortical Allografts

Allografts are often treated to decrease their immunogenicity and for long-term preservation. Such processing methods influence the degree of biologic reaction and/or mechanical properties. Freezing reduces the effects of histocompatibility mismatch and the biologic activity of the graft, most likely because of cellular destruction [35]. Likewise, freeze-drying reduces the graft immunogenicity [16], but also substantially alters its material properties, making it more brittle [50] and susceptible to fractures [37]. Ethylene oxide (EO) sterilization allows storage of allografts for long periods of time and reduces the risk of infection from graft contamination. However, EO-sterilized bone develops structural alterations during storage at room temperature that decrease its compressive strength and screw pullout load [51]. Conversely, storage at -20°C for one year does not significantly modify a graft’s resistance to compressive, bending, and torsional loads [52]. Although EO is an effective sterilization agent, it may have a deleterious effect on graft incorporation [53-55].

Gamma irradiation is the most widely used method in human medicine. However, because irradiation alters collagen crosslinks [56] and destroys the bone matrix fibrillar network,5 this process affects the material properties of the graft in a dose-dependent fashion [50]. Specifically, irradiation reduces graft strength and energy-absorbing capacity prior to failure [57,58]. However, experimental studies in rats have shown that low-dose irradiation followed by soaking in 70% ethanol and deep freezing does not compromise the natural course of graft incorporation [59]. Autoclaving induces protein denaturation, which significantly decreases the graft osteoinductive and osteoconductive properties [11]. Although experimental studies in dogs showed that autoclaved cortical autografts can be incorporated [60], the process markedly delays incorporation [5]. In addition, a 43% complication rate has been associated with the use of autoclaved allografts in human medicine [61]. In an attempt to increase cortical graft revascularization, microperforations of the graft have been performed. Although a perforation pattern involving an array of 300 μm laser-made holes placed 3 mm apart did not significantly affect compression strength and flexural deformity [62], perforation of cortical bone has been shown to substantially improve the amount of newly formed bone and enhance allograft incorporation [63].

Incorporation of Osteochondral Grafts

Osteochondral grafts include shell grafts and massive grafts. Shell grafts include articular cartilage and 2 to eight mm of subchondral bone, and are generally used for joint resurfacing. Shell autografts are incorporated quickly and completely when fixation is adequate [11]. Bone healing is obtained within 3 months [64-66]. In fresh osteochondral autografts, articular cartilage remains viable. In an experimental study of mosaicplasty (multiple autologous osteochondral plugs implanted in articular cartilage defects) in dogs, host-graft cartilage interfaces healed with fibrocartilage; a continuity between graft and host cartilage was obtained as early as 16 weeks and maintained until 1 year [67]. Massive allogeneic osteochondral grafts are used to reconstruct joints in limb-sparing procedures after tumor resection. These grafts are composed of cortical bone, epiphyseal and metaphyseal cancellous bone, and articular cartilage. The pattern of bone incorporation is similar to that of bone allografts [29]. Although the cartilage of frozen osteochondral allografts undergoes significant cellular and structural alteration including chondrocyte death, thinning, and fibrillation [68], it may provide adequate function for up to six years [55] before osteoarthritis becomes a problem [42]. Soft tissue including ligaments, tendons, and fascia may be firmly reattached to the surface of the allograft by a seam of appositional bone laid onto the graft [55].

Use of Bone Grafts

The three principal indications for bone grafts are enhancement of bone healing, replacement of bone lost through trauma or surgical resection, and in joint surgery.

Cancellous Grafts

Cancellous bone autografts are the most commonly used bone grafts in veterinary orthopedics, particularly to enhance healing in the treatment of comminuted fractures of long bones. Biologic fracture stabilization can be achieved using: (i) an “open-but-do-not-touch” approach that permits viewing of the fracture fragments, realignment of the bone, and cancellous bone grafting [69,70]. or (ii) a closed reduction that involves alignment without surgically exposing the fracture site. Although any non reconstructible fracture that requires open reduction and internal fixation may benefit from a cancellous autograft, use of minimally invasive techniques of fracture treatment, i.e., closed reduction without cancellous bone grafting, is more successful than open approaches [71]. In our experience, comminuted fresh fractures of long bones of dogs and cats treated with a minimally invasive approach and the use of a plate-rod construct without any cancellous graft heal successfully. Therefore, systematic use of cancellous bone autografts to enhance bone healing in comminuted fractures of long bones is questionable.

Treatment of delayed union or non-union fractures includes the use of rigid fixation often associated with cancellous bone grafting. Although most nonunions are biologically viable and are adequately treated by rigid internal fixation with compression of bone fragments, fresh autologous cancellous bone packed into the nonunion site and around bone ends should be considered as part of the procedure. In biologically inactive nonunions, cancellous bone grafting is essential [72]. Another major indication for cancellous bone autografts is filling of bone defects after surgical excision of benign bone tumors or bone cysts, after debridement of open fractures, and following sequestrectomy in osteomyelitis. Regardless of the condition, the bone ends and the surrounding soft tissues must provide revascularization of the cancellous graft, and adequate stabilization must be achieved. If the soft tissue bed is poorly vascularized, delayed grafting is preferable to immediate grafting after debridement and wound lavage [11].

Treatment of osteomyelitis requires excision of infected soft tissue and bone, which may result in significant bone loss, requiring filling of the defect with a bone graft or bone graft substitute. Because cortical grafts are likely to become infected and bone substitutes may act as foreign bodies, only fresh autologous cancellous bone should be used to fill such defects. However, because autologous cancellous bone undergoes significant resorption at the center of the graft [73], cancellous bone autografts are usually not recommended for the treatment of large segmental bone defects [73,74].

Masquelet et al., [73] described a new procedure for the treatment of segmental defects using cancellous autografts. The first stage consisted of inserting a cement spacer into the defect. This spacer has a mechanical role, preventing invasion of the recipient site with fibrous tissue, which may preclude further bone healing. It also has a biologic role, inducing formation of a richly vascularized pseudosynovial membrane around the cement plug. In the second stage, 1 month later, the membrane is incised and the cement spacer replaced with cancellous bone autograft, after which the membrane is sutured over the graft. This technique has been successfully used experimentally to treat large (3 cm) femoral diaphyseal and periosteal defects in sheep. In the group in which a cancellous graft was placed inside the induced membrane, the defects healed with restoration of the normal bone diameter. In contrast, graft resorption occurred in all animals when the autologous cancellous bone was placed after removal of the membrane.73 This membrane, therefore, actively contributes to the revascularization of the bone graft, acting as an in situ delivery system for growth factors (VEGF, TGF-β) and osteoinductive factors (BMP-2) [73,75] In people, this technique allowed reconstruction of bone defects up to 25 cm in length [75].

Joint arthrodesis requires stable fixation and use of cancellous bone graft (Fig. 88-1). Several clinical reports and experimental studies have demonstrated the usefulness of autogenous cancellous graft in arthrodesis. Indeed, the use of autogenous cancellous graft has been shown to enhance the rate of new bone formation and joint fusion when compared with non grafted joints, which allowed earlier removal of external coaptation devices [76].

To optimize the incorporation and survival rate of graft osteogenic cells, fresh cancellous bone graft should be placed in the recipient site as soon as possible after being harvested. For example, storage of cancellous bone for three hours, using methods similar to those used for organ preservation, resulted in a 20% decrease in the number of viable cells [6]. Cancellous bone may be applied to the surgical site with or without compression. Compression of cancellous bone autograft does not enhance osteogenic capability [77] and may reduce the osteoconductive potential of the graft [78]. Although a larger quantity of autologous cancellous bone is often considered beneficial, overfilling does not enhance early osteogenesis and must be avoided if it necessitates the use of multiple harvest sites [79]. Small graft fragments (approximately 3 mm by 1 mm) have been shown to have the best rate of revascularization and the greatest rate of survival of osteogenic cells [80].

Figure 88-1. Partial carpal arthrodesis with a cancellous bone autograft. A: Day 0; B: Follow up at 2 months showing graft incorporation and remodeling with complete bone fusion.

Corticocancellous Grafts

Corticocancellous grafts obtained from the iliac crest or a rib may be used as either blocks or small chips. A block of corticocancellous graft may be used as a cortical graft in relatively small segmental defects (Fig. 88-2). Whereas cortical bone provides mechanical support and volume expansion, cancellous bone enhances graft incorporation. Accordingly, corticocancellous grafts are often useful in neurosurgery to induce spinal fusion (e.g., in Wobbler syndrome) or when relatively large amounts of bone graft are required [81]. Incorporation of corticocancellous grafts is similar to that for cancellous bone grafts.

Cortical Grafts

Although cortical allografts have been widely used in the treatment of long-bone comminuted fractures in dogs and cats [2,11,24,26,27,31-33,36,37,51,82,83] the recent increased reliance on biologic fracture-repair techniques have made the use of cortical grafts nearly obsolete. Cortical autografts may be used in opening wedge osteotomies to limit limb segment shortening in the treatment of angular deformities. A small piece of cortical bone is packed within the bone defect to provide mechanical support. With rigid fixation, incorporation of such a cortical autograft is obtained within 2 to 3 months. Cortical allografts, associated with autogenous fresh cancellous bone packed around the host-graft interfaces, are mainly used in limb-sparing surgeries for the treatment of bone tumors. Here, too, the use of distraction osteogenesis via bone transport or use of metal prostheses may contribute to the reduction of clinical use of cortical allografts.

Osteochondral Grafts

Neither massive osteochondral allografts nor mosaicplasty are routinely used in small animal surgery. Although massive osteochondral allografts could potentially be used in limb-salvage procedures, logistical difficulties in obtaining anatomically appropriate grafts to achieve good joint function makes the use of cortical allografts with arthrodesis a more effective option. Recession sulcoplasty (block or wedge) in the treatment of patellar luxation is the osteochondral autograft most commonly used in small animal surgery. With such an osteochondral autograft, healing between the grafted sulcus and the femoral condyle is generally complete by 10 weeks. The subchondral trabecular bone is slightly thickened and the articular cartilage of the graft is viable and appears histologically normal [64].

Figure 88-2. Use of a corticocancellous bone graft (iliac wing) for the treatment of a nonunion of the ulna. A: preoperative radiograph; B: postoperative radiograph; C: follow-up at 6 months.