Abstract
Context.—Orthopedic and spine surgeons are in frequent need of bone for skeletal reconstruction. The amount of autograft is limited, and conventional allograft has some disadvantages, so surgeons are now using increasing amounts of demineralized allograft and a variety of synthetic materials to replace or “extend” autograft.
Objective.—To provide an overview of the composition and histology of the materials most likely to be seen by pathologists today.
Data Sources.—The review is based on published literature and the author's experience with preclinical studies and human biopsies.
Conclusions.—Pathologists are likely to find these skeletal substitute materials in biopsy and resection specimens from patients who have undergone prior treatment, and recognizing a synthetic bone substitute can help explain an otherwise confusing specimen. Pathologists also play an important role in helping define the safety and efficacy of new bioactive materials.
Orthopedic surgeons and neurosurgeons involved in skeletal repair and reconstruction often encounter bone defects that are unlikely to heal if treated by fixation alone. Autograft bone has been used for decades to supplement host repair, but the amount of autograft is limited, and morbidity related to autograft harvesting can be considerable.1 Surgeons commonly use allograft, and they recently have sought synthetic materials that could replace or be mixed with bone graft. Skeletal substitute materials are often needed in spinal fusion, filling voids around failed total joint prostheses, in fracture repair or repair of a nonunion, or after excision of a tumor. These bone graft preparations and graft substitutes may be encountered in subsequent biopsy or resection specimens, and it is important for pathologists to recognize these materials and distinguish them from an unexpected inflammatory response, residual or recurrent tumor, or other foreign material. Histology is key to defining the biocompatibility and bioactive properties of these materials. Pathologists can play an important role in helping recognize complications and defining the most appropriate material for a given clinical application. The purpose of this review is to provide an introduction to the histology of bone graft preparations and graft substitute materials, and to briefly describe the properties of each family of materials that might offer advantages for certain types of skeletal defects.
Several important terms are used to describe the properties of bioactive orthopedic materials. A material is described as osteoconductive if, due to its composition, shape, or surface texture, it promotes bone formation along its surface when it is placed in bone. A material is described as osteoinductive if it induces bone to form in an extraskeletal site, such as within skeletal muscle. A material is osteogenic if it causes bone formation because of the implantation of viable cells, such as autograft or aspirated bone marrow osteoprogenitor cells. All of these terms are relative. For example, a coating of hydroxyapatite (HA) on a titanium total joint implant is more osteoconductive than titanium alone, but titanium is more osteoconductive than cobalt chromium alloy or polymethyl methacrylate bone cement. Understanding these bioactive properties is critical to determining the most appropriate clinical application of a skeletal substitute material.2
Skeletal substitute materials, and orthopedic devices in general, usually show the most bone formation (most favorable histologic appearance) when they are harvested as part of a prospective animal study in which optimum surgery has been combined with an appropriate clinical application without complications. Surgical pathologists occasionally have the opportunity to study clinically successful cases that have been biopsied or retrieved at autopsy,3 but more frequently we see biopsy or resection specimens obtained because of a clinical failure or complication.4 Therefore, the histology commonly seen in human specimens may not be representative of findings that occur when that material is associated with a clinically successful outcome. Nevertheless, even the evaluation of clinically failed cases can provide useful information about safety and efficacy. The descriptions that follow will include examples of preclinical studies to illustrate the findings that are thought to occur in clinically successful human applications, as well as human specimens, most of which have been identified after biopsy or excision ofa clinical failure. This review is not intended to be acomprehensive comparison of available materials, but is instead an introduction to the histologic appearance of skeletal substitute materials most likely to be seen by pathologists as they evaluate biopsies and skeletal resection specimens (Table).
AUTOGRAFT
Bone graft materials in general include autograft (autogenous bone and/or aspirated cells), allograft, xenograft, synthetic materials, and combinations thereof. The gold standard for bone grafting is autograft harvested from the iliac crest, but surgeons also can make use of chips of autologous bone collected at the time of another procedure. For example, spine surgeons commonly harvest bits of decorticated transverse process, facet joints, and vertebral bodies at the time of a spinal fusion operation. Similarly, bone from a resected femoral head may be harvested and used to fill an acetabular defect at the time of total joint replacement. Segments of rib or fibula, sometimes with intact blood vessels, are used for some procedures that require immediate structural support. Few studies have attempted to compare the quality of autograft bone harvested from different sites. Cancellous autograft is incorporated into new bone relatively rapidly, but larger segments of cortical autograft—for example, a fibular autograft used to reconstruct a necrotic femoral head—may persist for years.2
The Table shows groups of bone graft and skeletal substitute materials organized primarily based on composition. As described below, the various preparations in each group often share mechanical and bioactive properties, but many individual preparations also have unique advantages and limitations. Selected examples from each group will be described and illustrated below.
BONE ALLOGRAFT
Mineralized Allograft
An allograft is tissue harvested from one individual and implanted into another individual of the same species. Several previous reviews have described in detail the histologic features of bone allograft incorporation.2,5–8 Allograft tissue is necrotic, and incorporation with adjacent bone is recognized morphologically by identifying spicules of necrotic bone (without cells in osteocyte lacunae) limited by cement lines from areas of new bone formation. This process of “creeping substitution” mimics the way in which necrotic bone is resorbed and replaced after a fracture or bone infarct. With allograft tissue, resorption and new bone formation occur first beneath periosteum and at the interface between graft and host. While small segments of cancellous allograft are remodeled rapidly, larger segments of cortical bone may remain recognizably necrotic for decades.6
The mineralized allograft preparations in most common use include frozen or freeze-dried chips of cancellous or cortical bone that have been harvested from cadaveric donors and aseptically processed by bone banks or tissue processing centers. Struts of harvested cortex also are frequently used in areas of more significant bone defects,9 and entire segments of bone, such as tibia or femur, are used in some reconstructions after tumor resection.6 Finally, segments of cortical bone can be machined into dowels for use in spine fusion and other applications. The use of fresh allograft is associated with the expected morphologic features of an immune reaction,10 but most donor cells are lysed or removed during the preparation of frozen or freeze-dried allograft preparations. Removal of bone marrow cells is thought to minimize graft rejection, but animal models have shown that persistent cells entrapped in osteocyte lacunae of frozen bone allograft can still sensitize a host.11 Nevertheless, most human studies that have compared clinical results and histology with human leukocyte antigen typing have not shown a clinically important immune reaction related to the use of frozen allograft bone in humans,11,12 and for most clinical applications tissue typing is not performed for bone allografting.
Demineralized Bone Matrix
Classic studies by Marshall Urist13 demonstrated that demineralized fragments of bone could induce bone formation when placed into skeletal muscle, thereby defining the process known as osteoinduction. The process of demineralization apparently releases bone morphogenetic proteins from the bone matrix, allowing these potent factors to induce stromal cells in the adjacent tissue to differentiate along osteoblast lineage. Demineralized bone matrix (DBM) preparations lack cells, so they are expected to be most effective when inserted into host sites that contain adequate vascularity and osteoblast precursor cell populations. When placed into a suitable host site, DBM can promote new bone formation. It is often mixed with autograft or other graft materials and is used to promote spine fusion, enhance fracture healing, and fill skeletal defects.14,15 Many different preparations of DBM now are available (Table), but they have variable osteoinductive properties, presumably based on differences in processing methods, carriers, and factors related to the donor.16,17 Depending on how the bone is machined and processed, different preparations can be composed of granules, strips of interwoven fibers, or puttylike preparations. Histologically, most DBM preparations can be recognized as unmineralized shavings of cortical bone that seem to be necrotic (ie, there are no cells in the osteocyte lacunae; Figure 1). Since the bone is demineralized during its preparation, it can be processed without decalcification in biopsies.Because DBM may be used during reconstruction after excision of a sarcoma, such as an osteosarcoma, the unmineralized spicules of DBM in a subsequent resection specimen should not be misinterpreted as tumor osteoid (Figure 2). In cases of clinical failure, the pieces of residual DBM often are surrounded by fibrous tissue, but because DBM induces endochondral bone formation, areas of cartilage or bone often are associated with the DBM spicules or in the adjacent tissue (Figure 3). DBM is not thought to be significantly antigenic to the host, so it is not usually associated with inflammation.
Spicules of a demineralized bone matrix (DBM) preparation in a biopsy obtained to rule out infection at the time of repair of a nonunion. The fracture site had been treated previously with both internal fixation and a commercially available DBM preparation. In this case, the residual DBM shavings are surrounded by fibrous tissue only, with no evidence of new bone formation. The shape of the particles and the empty lacunae without evidence of remodeling are helpful in recognizing this as a DBM material rather than necrotic host bone (hematoxylin-eosin, original magnification ×100). Figure 2. Spicules of demineralized bone matrix (DBM) associated with recurrent chondrosarcoma (CS). This patient had undergone resection of a CS, as well as reconstruction of the proximal tibia with mineralized allograft along with DBM. Evidence of local recurrence 6 weeks later led to re-resection. Spicules of commercial DBM in the specimen should not be misinterpreted as tumor matrix (hematoxylin-eosin, original magnification ×40). Figure 3. The osteoinductive process that is often initiated by demineralized bone matrix (DBM) includes endochondral bone formation. This figure shows areas of new bone formation (NB) and cartilage (Cart) adjacent to DBM that had been used as part of reconstruction of a large skeletal defect (hematoxylin-eosin, original magnification ×12.5). Figure 4. Decalcified section of a block of ProOsteon that had been placed previously into a bone defect in the mandible of a 77-year-old patient. The hydroxyapatite (HA) dissolved, leaving an apparent space that illustrates the porous geometry of this material derived from sea coral. There was no bone apposition in this specimen; instead, the tunnels inside the HA contained only fibrovascular tissue with numerous macrophages (hematoxylin-eosin, original magnification ×40). Figure 5. Decalcified section of a block of ProOsteon that had been used for mandibular augmentation. Apparent empty spaces indicate areas that were occupied by the hydroxyapatite (HA) of the ProOsteon. As seen here, portions of this implant showed good apposition and ingrowth (hematoxylin-eosin, original magnification ×100). Figure 6. Granules of sintered hydroxyapatite (HA) are soluble if sections are decalcified, often leaving a space surrounded by fibrous tissue or bone. This biopsy from a mandibular defect that had been treated previously with granules of HA mixed with autograft shows bone immediately surrounding apparent spaces that had been occupied by HA prior to tissue decalcification (hematoxylin-eosin, original magnification ×100).
Spicules of a demineralized bone matrix (DBM) preparation in a biopsy obtained to rule out infection at the time of repair of a nonunion. The fracture site had been treated previously with both internal fixation and a commercially available DBM preparation. In this case, the residual DBM shavings are surrounded by fibrous tissue only, with no evidence of new bone formation. The shape of the particles and the empty lacunae without evidence of remodeling are helpful in recognizing this as a DBM material rather than necrotic host bone (hematoxylin-eosin, original magnification ×100). Figure 2. Spicules of demineralized bone matrix (DBM) associated with recurrent chondrosarcoma (CS). This patient had undergone resection of a CS, as well as reconstruction of the proximal tibia with mineralized allograft along with DBM. Evidence of local recurrence 6 weeks later led to re-resection. Spicules of commercial DBM in the specimen should not be misinterpreted as tumor matrix (hematoxylin-eosin, original magnification ×40). Figure 3. The osteoinductive process that is often initiated by demineralized bone matrix (DBM) includes endochondral bone formation. This figure shows areas of new bone formation (NB) and cartilage (Cart) adjacent to DBM that had been used as part of reconstruction of a large skeletal defect (hematoxylin-eosin, original magnification ×12.5). Figure 4. Decalcified section of a block of ProOsteon that had been placed previously into a bone defect in the mandible of a 77-year-old patient. The hydroxyapatite (HA) dissolved, leaving an apparent space that illustrates the porous geometry of this material derived from sea coral. There was no bone apposition in this specimen; instead, the tunnels inside the HA contained only fibrovascular tissue with numerous macrophages (hematoxylin-eosin, original magnification ×40). Figure 5. Decalcified section of a block of ProOsteon that had been used for mandibular augmentation. Apparent empty spaces indicate areas that were occupied by the hydroxyapatite (HA) of the ProOsteon. As seen here, portions of this implant showed good apposition and ingrowth (hematoxylin-eosin, original magnification ×100). Figure 6. Granules of sintered hydroxyapatite (HA) are soluble if sections are decalcified, often leaving a space surrounded by fibrous tissue or bone. This biopsy from a mandibular defect that had been treated previously with granules of HA mixed with autograft shows bone immediately surrounding apparent spaces that had been occupied by HA prior to tissue decalcification (hematoxylin-eosin, original magnification ×100).
HA BLOCKS AND GRANULES
Many synthetic combinations of calcium and phosphate are osteoconductive. The rate at which these materials disappear, either by dissolution at neutral pH or at the acid pH produced by osteoclasts, is influenced by the calcium– phosphate ratio, crystallinity, and surface area. Among the first of these materials available for clinical use was ProOsteon (Interpore Cross International, Irvine, Calif).18 This material is derived from the conversion of the calcium carbonate of sea coral into a highly crystalline HA. Different species of sea coral have different shapes, and one of the first materials available commercially was derived from the Gonioptera genus, containing pores 500 to 600 μm in diameter in one direction, with interconnecting pores approximately 200 to 250 μm in diameter. These 3-dimensionally continuous pores provide the potential for extensive bone ingrowth if the HA block is placed into a suitable host site. Porous HA is brittle and has poor tensile properties, but its mechanical properties improve if bone apposition and ingrowth occur. ProOsteon was first approved by the US Food and Drug Administration in 1992 for use in filling bony voids, and it has been used to enhance repair of selected fractures,19 to augment bone in maxillofacial surgery, and as a graft extender for spine fusion, among other applications. It is available in blocks that can be machined into different shapes or as granules. As a highly crystalline material, ProOsteon HA lasts in vivo for a very long time (years). In biopsy or resection specimens, it may appear grossly as pieces of mineralized material, sometimes thought grossly to be bone or bone cement. The HA itself dissolves if the tissue sample is submitted for decalcification. In that case, the spaces occupied by the HA seem “empty” on microscope slides (Figure 4), although areas of bone ingrowth or apposition often can be recognized (Figure 5). A more recent formulation is incompletely converted from calcium carbonate to HA. Once the thin shell of HA is penetrated by osteoclasts, the inner core of calcium carbonate dissolves more rapidly. This material has less compressive strength than the original ProOsteon but is thought to be resorbed more rapidly in vivo.
Besides ProOsteon, surgeons also have access to granules of HA that have been made by precipitation and sintering. These granules also are osteoconductive, are very slow to dissolve in vivo, and are used usually as a “bone graft extender” (Figure 6). Although HA is not soluble at neutral pH, it can be dissolved by the acidic environment created by osteoclasts. This process is often, but not always, coupled with new bone formation. The slow dissolution of an HA coating on an implant or bone graft substitute can be associated with the release of small (less than 10 μm) crystals of HA that are then taken up by macrophages. Therefore, we expect to see occasional “foamy” macrophages around HA materials, but other types of inflammatory reactions do not occur unless there is a coexisting infection.
BLOCKS OR GRANULES OF SOLUBLE CALCIUM-SALTS
Sintered HA as described above is osteoconductive, but it is relatively insoluble at neutral pH. Its slow rate of dissolution is considered by some surgeons to be a disadvantage in certain clinical applications. Other preparations of calcium phosphate, such as tricalcium phosphate (TCP), biphasic calcium phosphate, and calcium sulfate, have varying degrees of solubility at neutral pH, based in part on the crystalline structure, surface area, and proportion of magnesium, zinc, strontium, or carbonate substitutions.20
Calcium sulfate is especially soluble in vivo, serving as a calcium source for mineral formation. Calcium sulfate preparations have been reported to be effective in filling unloaded skeletal defects,21 and one preparation (Osteoset, Wright Medical Products, Nashville, Tenn) was cleared by the Food and Drug Administration in 1996. Osteoset and other calcium sulfate preparations are now in common use for filling defects22,23 and as a carrier for antibiotics.24 If it has not already dissolved in the patient, calcium sulfate partially dissolves during routine decalcification of biopsy specimens, sometimes resulting in a void or an amorphous material surrounded by fibrous tissue or bone. Its appearance in microscope slides depends on how long it has been in vivo. Days or a few weeks after implantation, calcium sulfate has an appearance that suggests a precipitated or crystallized calcium compound, often with little bone apposition (Figure 7). Biopsies at later times show either normal bone, a cyst with fluid or fibrous tissue, or trabeculae of bone with entrapped residual mineral that is not birefringent with polarized light (Figure 8). Calcium sulfate preparations from different manufacturers have slightly different formulations, and some may be associated with an inflammatory reaction during the phase of rapid dissolution. Cysts, wound drainage, and cellulitis have been reported after calcium sulfate use.25,26 It may not be possible to tell histologically whether inflammation associated with calcium sulfate is due to infection or is just a transient inflammatory reaction related to the material itself.
Biopsy of a draining cyst in the calcaneus of a 14-year-old boy who had been treated 2 weeks prior with pellets of calcium sulfate (Osteoset). The amorphous mineral does not have a specific histologic appearance and can only be inferred as residual calcium sulfate based on the history (hematoxylin-eosin, original magnification ×100). Figure 8. Residual calcium sulfate (CaS) 12 weeks after insertion of Osteoset pellets in a surgically created defect in a sheep. In this case, most of the CaS had dissolved and been replaced by a combination of fibrous tissue and bone. A small amount of residual mineral is surrounded by bone (hematoxylin-eosin, original magnification ×100). Figure 9. A, Biopsy of a recurrent odontogenic keratocyst (K) containing granules of tricalcium phosphate (Collagraft TCP) that had been inserted during a previous curettage procedure. There are areas of new bone associated with the periphery of several of the granules in this field (hematoxylin-eosin, original magnification ×40). B, Higher magnification of other TCP granules from the same specimen. These granules do not show bone formation but are instead surrounded by only fibrous tissue. The residual material is not birefringent with polarized light (hematoxylin-eosin, original magnification ×200). Figure 10. Vitoss porous tricalcium phosphate (TCP) 6 weeks after insertion into a bone defect in a rabbit. The TCP is the slightly gray, granular-appearing mineral that is completely surrounded by bone. Vitoss is mostly porous, so it may be difficult to identify with certainty in routinely prepared specimens without the clinical history of prior use (hematoxylin-eosin, original magnification ×250). Figure 11. Dense network of mineralized collagen fibers represents Healos that had been placed adjacent to the spine at the time of an operation intended to induce intervertebral body fusion. A biopsy was obtained 1 year later at the time of an operation for failed fusion. Experimental studies suggest that Healos can be associated with new bone formation, especially when mixed with bone marrow stromal cells or bone morphogenetic proteins, but in this example it was not associated with bone formation (hematoxylin-eosin, original magnification ×400).
Biopsy of a draining cyst in the calcaneus of a 14-year-old boy who had been treated 2 weeks prior with pellets of calcium sulfate (Osteoset). The amorphous mineral does not have a specific histologic appearance and can only be inferred as residual calcium sulfate based on the history (hematoxylin-eosin, original magnification ×100). Figure 8. Residual calcium sulfate (CaS) 12 weeks after insertion of Osteoset pellets in a surgically created defect in a sheep. In this case, most of the CaS had dissolved and been replaced by a combination of fibrous tissue and bone. A small amount of residual mineral is surrounded by bone (hematoxylin-eosin, original magnification ×100). Figure 9. A, Biopsy of a recurrent odontogenic keratocyst (K) containing granules of tricalcium phosphate (Collagraft TCP) that had been inserted during a previous curettage procedure. There are areas of new bone associated with the periphery of several of the granules in this field (hematoxylin-eosin, original magnification ×40). B, Higher magnification of other TCP granules from the same specimen. These granules do not show bone formation but are instead surrounded by only fibrous tissue. The residual material is not birefringent with polarized light (hematoxylin-eosin, original magnification ×200). Figure 10. Vitoss porous tricalcium phosphate (TCP) 6 weeks after insertion into a bone defect in a rabbit. The TCP is the slightly gray, granular-appearing mineral that is completely surrounded by bone. Vitoss is mostly porous, so it may be difficult to identify with certainty in routinely prepared specimens without the clinical history of prior use (hematoxylin-eosin, original magnification ×250). Figure 11. Dense network of mineralized collagen fibers represents Healos that had been placed adjacent to the spine at the time of an operation intended to induce intervertebral body fusion. A biopsy was obtained 1 year later at the time of an operation for failed fusion. Experimental studies suggest that Healos can be associated with new bone formation, especially when mixed with bone marrow stromal cells or bone morphogenetic proteins, but in this example it was not associated with bone formation (hematoxylin-eosin, original magnification ×400).
Besides calcium sulfate, a number of other soluble materials are available, usually to extend autograft. Granules of β-TCP are osteoconductive, and they undergo dissolution more rapidly than highly crystalline HA but more slowly than calcium sulfate. Several different preparations of TCP may be encountered in biopsies or resection specimens (Table). One preparation is composed of HA-TCP granules mixed with bovine collagen along with granules of HA (Collagraft, Zimmer Corp, Warsaw, Ind). It is often mixed with autograft, and it is approved for use in fractures that also have been treated with internal or external fixation, as well as for osseous defects that do not bear significant load.27–29 The TCP dissolves during decalcification of biopsies but can sometimes be inferred by a space surrounded by bone or fibrous tissue (Figure 9, A and B).
A more recent formulation of β-TCP (Vitoss, Orthovita, Malvern, Pa) is 3-dimensionally macroporous, containing spaces into which bone ingrowth can occur. It also has a microporosity that is thought to promote diffusion of nutrients and transmission of fluid pressures. It does not have significant compressive strength by itself, but it is being used as an osteoconductive filler as well as a bone graft extender and a carrier of aspirated bone marrow cells. Vitoss also dissolves with decalcification and is difficult to recognize with certainty in most routine specimens. The presence of a synthetic calcium phosphate bone graft extender can be suggested in biopsy specimens by residual mineral or spaces in a shape that suggest a synthetic material (Figure 10). β-TCP is extremely biocompatible, so acute inflammation seen in a biopsy specimen should be interpreted as a probable coexisting infection rather than a reaction to the material itself.
Another osteoconductive material that is sometimes mixed with aspirated cell is Healos (DePuy Johnson and Johnson, Raynam, Mass). Healos is composed of bovine collagen fibers onto which a poorly crystalline calcium phosphate with a composition similar to HA has been precipitated. Healos by itself is osteoconductive, but the addition of cells30 or a bone morphogenetic protein31 may create a material with additional osteogenic or osteoinductive properties. Healos is thought to dissolve or be reabsorbed within weeks in vivo, but if it has been placed into a relatively avascular site (such as a nonunion) it can persist for months. Healos is difficult to recognize with certainty in microscope slides, but its presence can be suggested if one finds a fibrillar network of what seem to be mineralized fibers in the appropriate clinical setting (Figure 11).
BONE MORPHOGENETIC PROTEINS
As noted above, early studies by Urist et al13,32–34 led to the recognition that bone mineral contains proteins that influence osteoblast growth and differentiation. Subsequent studies led to the purification of numerous bone morphogenetic proteins that have structural homology with the transforming growth factor β molecule. Bone morphogenetic proteins (BMPs) are now recognized to be important growth and differentiation factors that influence many aspects of tissue growth besides just bone. Among these molecules, BMPs 2, 4, 6, and 7 (also known as osteogenic protein 1 [OP-1]) are thought to be the most important BMPs for bone formation.35,36 Although early studies tested BMPs purified from large volumes of demineralized bone, genetic engineering technology has led to the production and commercialization of human recombinant BMP-2 (Genetics Institute, Cambridge, Mass) and BMP-7 (or OP-1, Stryker Biotech, Hopkinton, Mass), and additional molecules, such as recombinant GDF-5, are expected to become available in the near future. There are now literally thousands of publications that describe the biology and testing of BMPs in preclinical models, as well as in human clinical trials to treat fracture nonunions, spine fusion, and other applications.37–40
While recombinant BMP molecules are extremely potent, they are difficult to use clinically in a powder or solution. Their handling properties and biologic activity are enhanced when BMPs are delivered with carrier materials, but the best carriers for various surgical applications have not yet been determined. Two formulations in relatively widespread use are BMP-2 adsorbed to a bovine collagen sponge (Infuse, Genetics Institute) and BMP-7 with a carrier of bovine type I collagen, sometimes mixed with carboxymethyl cellulose (OP-1 Device, Stryker Biotech). Many studies are underway to try to identify better carriers for selected indications.41,42 Infuse and the OP-1 Device are now being used in the United States to promote spine fusion43 and to treat nonunions.44 Both BMPs have been used in Europe and Australia for a variety of other indications.
Of course, the BMP molecules themselves cannot be recognized histologically, but if biopsies are obtained before the carriers disappear, then the presence of a BMP preparation can be suggested based on morphologic findings in biopsies. The bovine collagen of the OP-1 Device has a histologic appearance similar to that of a demineralized bone matrix material (Figure 12), whereas the collagen sponge used with BMP-2 for the Infuse device is a tight network of collagen fibers. Occasionally, patients develop transient chronic inflammation related to the bovine collagen of the OP-1 Device, but otherwise we do not expect to see any inflammation associated with the use of either BMP-2 or BMP-7.
Although recombinant bone morphogenetic proteins (BMPs) are thought to induce bone formation in most cases, sometimes a subsequent biopsy shows only fibrous tissue and residual carrier. This biopsy was obtained adjacent to a transverse process in the spine where the OP-1 Device (osteogenic protein 1 [OP-1]) had previously been used to promote posterolateral spine fusion. The bovine collagen carrier in this case is histologically similar to a demineralized bone matrix product and is associated with chronic inflammation (hematoxylin-eosin, original magnification ×100). Figure 13. Biopsy obtained at the time of hardware removal, approximately 1 year after a calcaneal fracture had been treated with Norian SRS along with a fracture fixation plate. Residual calcium phosphate cement is surrounded mostly by bone, but some of the cement has been resorbed and replaced by lamellar bone in a manner similar to the way normal bone remodels (undecalcified section, Giemsa stain, original magnification ×100). Figure 14. Biopsy obtained during surgical fixation of a periprosthetic fracture that occurred 2 years after fixation of a prior pathologic fracture related to recurrent malignant lymphoma. The biopsy shows spaces that had been occupied by a 3-dimensionally porous hydroxyapatite (ProOsteon) (HA), along with numerous particles of demineralized bone matrix (DBM). Recurrent malignant lymphoma (L) is also present (hematoxylin-eosin, original magnification ×12.5).
Although recombinant bone morphogenetic proteins (BMPs) are thought to induce bone formation in most cases, sometimes a subsequent biopsy shows only fibrous tissue and residual carrier. This biopsy was obtained adjacent to a transverse process in the spine where the OP-1 Device (osteogenic protein 1 [OP-1]) had previously been used to promote posterolateral spine fusion. The bovine collagen carrier in this case is histologically similar to a demineralized bone matrix product and is associated with chronic inflammation (hematoxylin-eosin, original magnification ×100). Figure 13. Biopsy obtained at the time of hardware removal, approximately 1 year after a calcaneal fracture had been treated with Norian SRS along with a fracture fixation plate. Residual calcium phosphate cement is surrounded mostly by bone, but some of the cement has been resorbed and replaced by lamellar bone in a manner similar to the way normal bone remodels (undecalcified section, Giemsa stain, original magnification ×100). Figure 14. Biopsy obtained during surgical fixation of a periprosthetic fracture that occurred 2 years after fixation of a prior pathologic fracture related to recurrent malignant lymphoma. The biopsy shows spaces that had been occupied by a 3-dimensionally porous hydroxyapatite (ProOsteon) (HA), along with numerous particles of demineralized bone matrix (DBM). Recurrent malignant lymphoma (L) is also present (hematoxylin-eosin, original magnification ×12.5).
INJECTABLE CEMENTS
Most of the materials mentioned above have essentially no mechanical strength at the time of implantation, so they must be protected by metal hardware or used in areas of the skeleton that are not subjected to significant load. However, some clinical applications benefit from the use of a material that has its own mechanical strength shortly after insertion. This is especially true for applications in which the material would be loaded primarily in compression, such as tibial plateau fractures, calcaneus fractures, and vertebral augmentation (vertebroplasty or kyphoplasty).45–49 Polymethyl methacrylate cement (PMMA), the same material used in total joint arthroplasty, has good compressive strength and has been especially useful for vertebral augmentation.50,51 PMMA dissolves during routine microscope slide preparation, but its presence in tissue can be inferred sometimes by recognizing voids that contain small particles of barium sulfate or alumina that have been mixed with the PMMA to make it radio-opaque. Many pathologists have seen residual PMMA in membranes around failed total joint prostheses. Injected into a recent osteoporotic compression fracture or vertebral metastasis, PMMA reduces pain and seems to minimize further vertebral collapse, but it is exothermic as it cures, potentially causing necrosis of adjacent bone. It is not osteoconductive, and particles of PMMA are known to induce a macrophage reaction that can result in bone resorption, so alternative cements are being developed to enhance poor-quality cancellous bone. These cements have variable biologic properties, strength, and durability. Norian SRS (Synthes Corp, Paoli, Pa), BoneSource (Stryker Orthopaedics, Mahwah, NJ), and Callos (Skeletal Kinetics, Cupertino, Calif) are calcium phosphate cements that are mixed in the operating room, injected into poor-quality bone, and cure to develop compressive strength about midway between that of cancellous and cortical bone.52 They are extremely osteoconductive and are very slowly resorbed by osteoclasts and replaced by bone similar to the way normal bone remodels.47 Similar cement preparations with slightly different handling properties are available for filling cranial defects. A different calcium phosphate cement (α-BSM, ETEX Corp, Cambridge, Mass), has relatively poor compressive strength by itself, but a modification of the cement is being used as a carrier for BMP-2.42,53 Several silica-based cements are available for dental procedures, and a silica-containing bioactive cement (Cortoss, Orthovita) is available in Europe and is under evaluation in the United States for both enhancing fixation of screws in poor-quality bone and for vertebral augmentation.
Undecalcified preparations allow easy visualization of calcium phosphate cements (Figure 13). The cements dissolve during decalcification of routinely processed biopsies, but adsorbed proteins sometimes allow recognition of where the cement had been surrounded by bone. Fragmented particles of cement are taken up and dissolved by macrophages, but otherwise these injectable cements are extremely biocompatible and are not associated with inflammation.
CONCLUSIONS
Bone graft and bone graft substitute or extender materials are now in widespread use and are present in many biopsy and resection specimens that are submitted to pathology laboratories. Although their recognition may not influence the immediate care of an individual patient, pathologists have an important role in helping document the safety and efficacy of these new biomaterials, and recognizing a bone graft material can help explain an otherwise complex specimen (Figure 14). Some of the synthetic materials dissolve during routine decalcification, so the specific identification of any given material can be difficult. However, an apparently empty space surrounded by bone or fibrous tissue in a biopsy obtained from a nonunion, site of previous tumor resection, or spine fusion is likely to have been occupied by one of the bone graft substitute/ extender materials mentioned above. Unmineralized (demineralized) spicules of a DBM material or of the OP-1 Device should not be misinterpreted as tumor osteoid when encountered at the site of a prior tumor, and the lack of cells in the lacunae of a DBM material should not be misinterpreted as ischemic necrosis of host bone. With the exception of a mild inflammatory reaction to bovine collagen in some patients and inflammation sometimes seen during the first few weeks after calcium sulfate insertion, these biomaterials are not usually associated with significant inflammation, so acute inflammation associated with one of these materials should raise the suspicion of ongoing infection.
Acknowledgments
This study was supported in part by Core Center for Musculoskeletal Disorders award AR 050953 from the National Institutes of Health.
References
The author or his research laboratory has received funding related to research involving skeletal substitute material from the following sources: Stryker Orthopaedics, Stryker Biotech, DePuy Spine Corporation, Medtronic Sofamor Danek Corporation, Orthovita Corporation, Osteotech Corporation, Musculoskeletal Transplant Foundation, Norian Corporation (Synthes Corporation), and Skeletal Kinetics Corporation.
Author notes
Reprints: Thomas W. Bauer, MD, PhD, Departments of Pathology, Orthopaedic Surgery, The Spine Center, and the Center for Orthopaedic Research, Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH 44106 ([email protected])