Comminuted, displaced fractures of the tibial plateau; calcaneal fractures; intertrochanteric fractures of the proximal part of the femur; and Colles fractures may be difficult to treat. Specifically, a stable, satisfactory reduction may be difficult to achieve and to maintain with use of conventional methods because of the fragmentation of cortical and cancellous bone. The treatment of these fractures may be facilitated by the use of an injectable biomaterial that solidifies in situ, provides immediate mechanical stability, helps to maintain a satisfactory reduction, and is gradually resorbed and simultaneously replaced by normal host bone.
Historically, autogenous grafts, allografts, and a variety of biomaterials have been used for the repair of osseous defects and the augmentation of compromised bone1-17,23,24,26-28,32-37. However, problems related to the availability of graft material, donor-site morbidity, immunogenicity, biomechanical integrity, and long-term compatibility have limited the success of these methods11,13,14,21,28,30.
Recent developments in the formulation of calcium phosphate cements may make these materials useful for fulfilling these clinical needs. These materials are osteoconductive, they can have substantial compressive strength, and they can be resorbed in vivo1,4,6,17,24,28,29. As bone-mineral resorption occurs within an osteoclast-mediated acidic microenvironment, the material should have characteristics that render the apatite soluble at an acidic pH. Carbonated apatite of low crystalline order has been shown to have the greatest solubility at an acidic pH18,22,24,29,38,39. Synthetic carbonated apatites are classified according to the mode of carbonate substitution. Substitution of a carbonate for hydroxide yields an apatite of high crystalline order (type A), whereas substitution of a carbonate for phosphate produces an apatite of low crystallinity (type B). Bone mineral contains primarily type-B apatite18. Type-A apatite is usually formed by sintering at high temperatures and is relatively stable, whereas type-B apatite is produced near room temperature and is relatively soluble at an acidic pH. In order to facilitate resorption of a skeletal substitute material, it would be advantageous to use a type-B apatite.
A synthetic calcium-phosphate material for use in filling defects of bone effectively should possess chemical properties that promote bone apposition, should be readily resorbed and replaced by bone, and should have the mechanical strength required to maintain a stable reduction of a severely comminuted, displaced fracture. The biomechanical properties of the material should resemble those of the native bone, and the material should be resistant to fragmentation during the process of resorption and replacement with new bone. Previous studies of a calcium-deficient carbonated apatite (FractureGrout; Norian, Cupertino, California) have suggested that its biological and mechanical properties might make it a satisfactory structural substitute for cancellous bone in selected clinical applications6,31.
FractureGrout results in the in situ formation of a mineral phase of bone. Monocalcium phosphate, monohydrate, a-tricalcium phosphate, and calcium carbonate are mixed dry and then suspended in a sodium phosphate solution. The total time for preparation of the cement is approximately five minutes. This process results in the formation of an injectable paste that begins to harden in four minutes to form a carbonated apatite (dahllite) through a non-exothermic reaction. In four hours, the material hardens to 90 per cent of its ultimate strength. It reaches full strength (twenty-five megapascals) in twenty-four hours. Both the chemistry and the crystallographic structure of this carbonated apatite material have been described previously6,18,31. The structure of FractureGrout, as determined with x-ray diffraction, closely resembles that of normal human bone, and this appears to explain its reported capacity to elicit bone apposition and to be replaced by bone. FractureGrout is a precursor formula to the Skeletal Repair System (Norian), which is the trade name of a material that is currently under clinical investigation. Skeletal Repair System cures to a compositional and crystallographic structure that is identical to that of FractureGrout, and the biological properties of the two materials appear to be identical. However, Skeletal Repair System has a greater compressive strength (fifty megapascals) than FractureGrout (twenty-five megapascals).
The purpose of the present study was to evaluate the repair of bone in corticocancellous metaphyseal defects with use of a calcium phosphate cement that sets in situ (FractureGrout).
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was a grant from the Norian Corporation, Cupertino, California.
†Orthopaedic Research Laboratories, University of Michigan, G-0161, 400 North Ingalls Street, Ann Arbor, Michigan 48109-0486.
‡The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195.
§Norian Corporation, 10260 Bubb Road, Cupertino, California 95014-4166.
Experimental Design and Operative Models
Seventy-two adult male mongrel dogs that each weighted approximately twenty-five kilograms were used in the study. All animals were screened to ensure that they were in good physical condition and that their growth plates were closed. Each dog was randomly assigned, on the basis of the time of entry into the research facility, to be treated with a partial osteotomy of the proximal part of the tibia in either the right or the left hind limb and to receive two cylindrical defects in the distal part of the contralateral femur. The dogs also were randomly assigned to treatment with either cement or allograft. Each animal received cephalexin (Keflex; thirty milligrams per kilogram of body weight) preoperatively, and the antibiotic therapy was continued for five days after the operation.
Seventy dogs were divided into five equal groups and were killed at twenty-four hours or at two, four, eight, or sixteen weeks postoperatively. Seven animals in each group were treated with calcium phosphate cement, and seven were treated with allograft bone. The remaining two dogs both received the calcium phosphate cement in the defects and were killed at thirty-two or seventy-eight weeks postoperatively.
Two types of metaphyseal defects were created. The first type was produced in the proximal tibial metaphysis (Fig. 1). An incision was made over the anterior spine of the tibia, just proximal to the tibial tuberosity, and was extended to the distal border of the tibial crest. The insertion of the patellar ligament was preserved, and the tissues were reflected subperiosteally on the medial and lateral sides of the tibial crest until the borders of the medial and lateral collateral ligaments were visualized near the posterior cortex. After a guide-hole had been drilled into the proximal part of the tibia, a rectangular defect was produced with use of a reciprocating saw under constant irrigation. The saw was equipped with a specially manufactured blade that created a defect that was 3.5 millimeters in vertical height. The resected region extended from the anterior surface (just proximal to the insertion of the patellar ligament) toward the posterior cortex. All cortical and cancellous bone was removed from the defect, which was circumferentially complete except for approximately six millimeters of intact cancellous and cortical bone posteriorly. As the defect was located just proximal to the insertion of the patellar ligament, no fixation other than the calcium phosphate cement or the allograft was necessary.
After the defect had been filled with either the calcium phosphate cement or the allograft, the tissues were closed in layers and the skin was secured with staples. A Robert Jones dressing was applied to the limb for the first twenty-four hours postoperatively, and the animal was allowed to walk freely.
The site of the second metaphyseal defect was chosen to simulate a region that required a graft but was more protected from large compressive and torsional loads. In the contralateral hind limb of each animal, the lateral and medial femoral condyles were exposed and an eight-millimeter-diameter hole that penetrated eight millimeters into the area of trabecular bone was drilled in each condyle (Fig. 2). The femoral defect model was used to evaluate the mechanical integrity of the cement as a function of time after implantation. In both models, the defects were completely filled with calcium phosphate cement or morseled canine allograft.
Allograft trabecular bone was obtained from the proximal part of the tibiae and the distal part of the femora of large mongrel dogs. It was stored frozen at -20 degrees Celsius before being processed and was sterilized with gamma irradiation. The allograft specimens were defatted by soaking for more than one hour in 100 per cent isopropyl alcohol. Marrow was removed by spraying the solution through the bone. The specimens then were rehydrated in sterile water at 37 degrees Celsius for fifteen to thirty minutes, washed, and air-dried in a sterile hood. The specimens were sealed in double pouches, sterilized with 2.5 megarad (25,000 gray) of gamma irradiation at another facility (Sterigenics, Tustin, California) within one day, and shipped to the University of Michigan. The sterile packs were stored at room temperature. At the time of the operation, the bone was hydrated for five minutes with saline solution, morseled with rongeurs, and densely packed into the defects. The graft material was not mixed with blood, except for that on the recipient host surfaces.
The calcium phosphate cement was mixed, according to the manufacturer's recommendations, at the time of the operation. The mixing was done with a mortar and pestle, and the prepared paste was placed in a three-milliliter syringe. Preparation of the cement took approximately five minutes. It was then injected through a large-bore needle into the defect. The injection was continued until the material completely filled the defect and was contiguous with the periosteal surface. Excess material was removed before soft-tissue closure. As with the allograft material, the volume of the cement that was used depended on the size of the tibia in each animal. No specific for pressurizing the cement were used.
Of the seven animals in each group that were killed in the first sixteen weeks after the operation, five were used for biomechanical studies and two were used for histological analysis. Radiographs were made on a weekly basis to ensure that the graft material remained within the experimental site. Both tibiae and the distal part of treated femur from each animal were dissected free and stored for mechanical testing and histological analysis. The two animals (both treated with the calcium phosphate cement) that were killed at thirty-two and seventy-eight weeks were only evaluated histologically, in order to assess the long-term biological response to any remaining cement.
Mechanical Testing
The femora and tibiae of the animals scheduled for mechanical testing were dissected free of all soft tissues. Each bone was wrapped in paper towels that had been soaked in saline solution and then was frozen at -20 degrees Celsius. Before testing, the specimens were thawed in a water bath at room temperature. The tibiae were embedded in aluminum channels with polymethylmethacrylate and use of a standardized fixture to ensure consistent gauge length. Two small bone screws were inserted into the articular surface of the tibia, and the region proximal to the osteotomy site was potted in the channel. With maintenance of a gauge length of 5.5 inches (14.0 centimeters), the distal end of the tibia was fixed in a similar manner.
Biomechanical testing of the tibiae was performed at room temperature with use of servohydraulic torsion-testing machine (model 810; MTS, Minneapolis, Minnesota). The specimens were kept moist and were mounted so that the proximal end could be rotated while the distal end was held fixed. Both the experimental tibiae (those treated with cement or allograft) and the intact, contralateral tibiae were subjected to torsional loading in external rotation at 100 degrees per second until failure. Continuous load and rotation data were collected with an analog-to-digital converter and were stored for analysis on a Macintosh microcomputer. From the recorded torque versus angular displacement curves, we calculated the values of maximum torsional load, the displacement at maximum torsional load, torsional stiffness, and energy to failure. In an effort to reduce errors associated with geometric differences among the animals, all mechanical test data were expressed as a percentage of the results for the intact, contralateral tibiae. As a test for consistency across the animals, the mechanical data for the intact, contralateral tibiae were compared among all groups. Statistical analysis of the results was performed with use of a two-way analysis of variance across time and group. The result was considered to be significant if the p value was less than 0.05. Pairwise analysis was performed with use of a correction for multiple comparisons.
For mechanical evaluation of the femoral metaphyseal plugs, the distal part of each femur was cleaned of all soft tissues and the condyles were separated. The distal part of the femur was then carefully positioned in a vise so that a 7/16-inch (approximately eleven-millimeter) hole saw, powered by a drill press, would encompass the defect. Under constant irrigation, the bones were cored such that the specimen included a thin sheath of trabecular bone surrounding the cement or allograft plug. The specimens then were cut to produce a right cylinder of cancellous bone and cement or cancellous bone and allograft with a constant height of six millimeters. For testing, the specimens were loaded in compression at a constant displacement rate of 0.015 inch (0.381 millimeter) per second for six seconds with use of a servohydraulic testing machine (model 810; MTS). The variables recorded from the load and displacement data included yield load, stiffness, and energy to yield. In addition to the specimens from the animals treated with allograft or cement, a group of ten control specimens of trabecular bone from matched locations in normal dogs were evaluated with use of the same procedure.
Histological and Morphological Analysis
To improve infiltration, the specimens for histological analysis were cut into smaller specimens to include the region of the osteotomy gap or the cylindrical plugs, the surrounding trabecular bone, and the cortical bone. The specimens were fixed in 70 per cent ethanol and embedded in methylmethacrylate. Each block was sectioned into 800-micrometer slices with use of a diamond-impregnated band saw (model 300; Exakt Apparatebau, Hamburg, Germany). This process yielded seventeen to twenty-five vertical coronal sections from each tibial specimen and nine sagittal sections from each femoral specimen; the sections included the defects and the surrounding bone. The odd-numbered sections were used for light microscopic evaluation, and the even-numbered sections were used for electron microscopic analysis.
Sections that were designated for electron microscopic analysis were ground and polished, coated with carbon, and then examined with a scanning electron microscope (model 1810; Amray, Bedford, Massachusetts) that was equipped with a back-scattered imaging detector. Each section was scanned sequentially at low magnification to observe the entire region of interest and then at multiple higher magnifications to observe the interface between the implanted material and the native bone.
Sections that were to be evaluated with light microscopy were mounted on Plexiglas plates with use of an adhesive (Permabond 910; Permabond, Eastleigh, United Kingdom) and were allowed to dry overnight. The sections were then ground and polished by hand, on a grinding and polishing table (Ecomat IV; Buehler, Lake Bluff, Illinois), to a thickness of less than fifty micrometers. Sections were stained with either hematoxylin and eosin or toluidine blue.
Histomorphometric Studies
Histomorphometric analysis was used to estimate the rate of resorption of the dahllite and the replacement by new bone. The amounts of cement resorption and new-bone formation in the cortical area and in the medullary defect were quantified separately. An image-analysis system, consisting of a microscope, digitizing pad, microcomputer, and image-analysis software (BQ System IV; Bioquant, Nashville, Tennessee), was used to measure the total area of the defect, the area within the defect that was occupied by cement, and the area within the defect that was occupied by bone. The boundaries of the cortical portion of the defect were estimated on the basis of the location of the endosteal and periosteal margins of cortical bone proximal and distal to the defect. The area of the medullary defect was estimated by constructing a trapezoid, in which the superior and inferior borders were parallel lines corresponding to the superior and inferior edges of the defect. These boundaries were intended to define the interface between bone and cement, and the preexisting host bone was not included in the area of the defect. The anterior and posterior borders of the medullary defect corresponded to the endosteal margin of cortex as just described. The relative proportion of the defect that was occupied by cement and by bone was expressed as a percentage of the total area of the defect. Measurements were obtained from three to six serial histological slides of each specimen; however, the statistical analysis was based on the number of animals, not the number of slides, in each group. Descriptive statistics, including analysis of variance and the Student t test, were used to detect differences among the groups, and a p value of less than 0.05 was considered to be significant. Data points from the specimens from the dogs that were killed at thirty-two and seventy-eight weeks were not used for statistical comparisons, but the micrographs of those specimens were used to demonstrate long-term resorption of the cement.
All animals recovered from the operative procedure without incident. By the second postoperative day, the dogs began to walk freely and were active throughout the duration of the study. There were no infections or other complications associated with the model or with the implanted materials.
Biomechanical Findings
The effect of the osteotomy in the proximal part of the tibia in the dogs that were killed at twenty-four hours was a reduction in torsional strength of approximately 50 per cent, when compared with that of the contralateral, control tibiae (Table I). Thus, the defect significantly compromised the whole-bone torsional integrity of the tibia (p < 0.001). The tibiae that had been treated with allograft reached approximately 100 per cent of the strength of the control tibiae by four weeks, whereas the tibiae that had been treated with cement reached their maximum load to failure by eight weeks (Table I). Although a significant difference could not be detected with the numbers available for study, the trend suggests that the allograft bone incorporated slightly faster than the cement. Similar trends were found for stiffness, displacement at maximum load, and energy to failure (Table I).
The locations of the fractures that occurred during mechanical testing supported the test results. As expected, all of the control tibiae broke in the mid-part of the diaphysis. At twenty-four hours after the operation, both the tibiae treated with cement and those treated with allograft fractured through the defect. At two weeks after the operation, all tibiae (except for one treated with allograft) demonstrated the same response. At four weeks after the operation and later, all tibiae treated with allograft fractured through the mid-part of the diaphysis. At eight weeks after the operation, four tibiae treated with calcium phosphate cement fractured through the diaphysis, and, at sixteen weeks, all such tibiae fractured through the diaphysis.
The femoral-plug compression test provided an assessment of whether the cement degraded mechanically over time because the test evaluated the integrity of the implanted material in a bounded defect and not interfacial incorporation (Fig. 3). At all time-periods, the mechanical properties of the cement were equal to or greater than the properties of the trabecular bone in the control specimens. In comparison, the mechanical properties of the allograft demonstrated lower values at twenty-four hours after the operation, but normal values were attained by two weeks. Although the specimens containing calcium phosphate cement showed a trend toward a higher yield load at twenty-four hours compared with those containing allograft, no significant difference was found with the numbers available. Test results for stiffness and energy to failure revealed similar findings, although there was a difference between the cement and the allograft at twenty-four hours and at eight weeks postoperatively.
Histological Findings
Group Treated with Allograft
The size of the tibial defect could be easily appreciated at twenty-four hours after the operation. However, by as early as two weeks after the operation, woven bone in the defect made the margins difficult to recognize with certainty. By four weeks, a few spicules of necrotic bone consistent with allograft were still evident, but remodeling of the graft was relatively advanced with abundant new-bone formation. These findings corresponded well with the increase in measured torsional properties of the whole-bone composite (Table I). Remodeling continued throughout the sixteen weeks, with apparent complete incorporation of the allograft and replacement with viable lamellar bone in the defect. The sites of the femoral defects showed a similar profile of incorporation and replacement, although perhaps at a slightly slower rate.
Group Treated with Cement
Histological sections from tibiae demonstrated that the cement was both osteoconductive and bio-compatible. At two weeks, there was extensive apposition of bone, which was composed of mostly woven bone and unmineralized osteoid (Fig. 4). There was no fibrous tissue between the cement and the bone and no inflammation. Occasional osteoclasts also could be identified, and in some sections they appeared to be directly on the surface of the cement in locations that suggested active mineral resorption (Fig. 5). Adjacent to the osteoclasts were areas of enlarged osteoblasts that were associated with new-bone formation and appeared to be directly on the surface of the cement (Fig. 5). Although these histological features suggested intense cellular activity on the cement surface at two weeks, low-magnification photomicrographs showed that only a small proportion of the overall volume of cement was resorbed during the first several months (sixteen weeks) (Fig. 6).
With more time in vivo, the cement showed evidence of vascular penetration associated with resorption and new-bone formation. This process appeared to occur more rapidly in the cortex, such that by sixteen weeks much of the cement within the cortex had been resorbed and replaced by bone (Fig. 7). By thirty-two weeks, vascular penetration into the cement was evident in the medulla and many vessels were found to have circumferential lamellae of bone that resembled developing haversian systems (Fig. 8). This haversian-like bone formation histologically resembled the process of normal bone-remodeling. New-bone growth in both bone and cement areas was beginning to recapitulate marrow spaces (Fig. 9). There was no evidence of fibrosis or inflammation at any time-interval.
The process of osteoclastic resorption followed by vascular ingrowth and new-bone formation was especially evident at seventy-eight weeks. Developing haversian systems that consisted of central blood vessels and overlapping circumferential lamellae and that were bounded by peripheral reversal lines were clearly evident (Figs. 10-A, 10-B, and 10-C). Photomicrographs at low magnification suggested a process of gradual resorption of cement with restoration of a physiological proportion of trabecular bone and marrow space (Fig. 11). This process appeared to have occurred more rapidly in the cortex, and it was incomplete in the medulla at seventy-eight weeks (Fig. 6).
Histomorphometric Findings
Qualitatively, both the tibial and the femoral defects showed little resorption of cement in the cancellous bone by sixteen weeks postoperatively. However, histomorphometric studies provided a quantitative estimate of the rate of resorption and replacement of the cement by host bone in both types of defects. In the tibial defects, the mean percentage of the cavity occupied by the cement progressively decreased from 90.5 per cent at twenty-four hours to 34.1 per cent at seventy-eight weeks (Fig. 12). This decrease was associated with a progressive increase in the mean amount of bone in the area of the defect, from 0 per cent at twenty-four hours to 26.2 per cent at seventy-eight weeks. Furthermore, the pattern of trabecular bone that was being formed in the area of the defect was qualitatively similar to the adjacent trabecular bone superior and inferior to the defect (Fig. 13), such that the trabecular bone volume fraction at this site appeared to be approaching the normal value (range, 31 to 48 per cent) found in metaphyseal regions in dogs24. In the femur, the mean percentage of the defect area occupied by cement decreased from 97 to 67 per cent, whereas the mean percentage of the defect area occupied by bone increased from 0 to 23 per cent (Fig. 14).
Calcium phosphate materials have been widely investigated as osteoconductive coatings on metal implants, as fillers for voids in bone, and even as carriers for bioactive compounds1,5,21,26,29,37,38. These studies have clearly demonstrated that calcium phosphate materials have a number of attractive properties, including resorbability, osteoconductivity, and formulation-dependent mechanical strength. The properties of these materials depend on the conditions under which they were formed and on their inherent chemistry and final porosity. Some formulations result in a granular mixture with limited handling properties, whereas others resorb very rapidly and lack intrinsic mechanical strength. Several calcium phosphate materials that have compressive strength close to or greater that that of bone remain brittle and resorb very slowly or not at all. Previously, no calcium phosphate cements appeared to have both substantial compressive properties (similar to that of bone) and a chemistry that leads to continuous resorption by osteoclasts followed by replacement with bone8,18,20,22,24,26,28,34.
The results of the present study demonstrate that FractureGrout calcium-phosphate cement (Norian) is both osteoconductive and resorbable, with simultaneous replacement by bone through normal osteoclast-osteoblast coupling. The injectability of the material made it possible to fill gaps easily and efficiently during the operative procedure. Perhaps even more importantly, after hardening in the defect, the cement provided mechanical support and ensured full function of the tibia during normal activities of the animals.
The dog model used in the present study was thought to be appropriate as many attributes of canine trabecular bone are similar to those of human bone19. It should be noted, however, that bone remodels slightly faster in dogs than in humans and that the replacement of the cement may be slower in humans19. The use of canine allograft as a control provided a clinically relevant comparison for the cement as well as a viable alternative for filling such defects. A negative control, in which no material was used to fill the tibial defects, would have led to catastrophic failures.
The tibial defect model proved to be valuable for studying the effects and behavior of the calcium phosphate cement. At twenty-four hours postoperatively, the torsional strength of the bone was reduced by approximately 50 per cent and the energy to failure, by approximately 80 per cent. As time progressed, both the allograft and the cement constructs followed a normal healing pattern, and the allografts appeared to be incorporated at a slightly faster rate. The appearance of osteoclast resorption of the cement followed by new-bone formation suggests that all of the implant material may be replaced eventually. At no time during the seventy-eight weeks was there an adverse biological or mechanical response to the remaining cement.
The mechanical testing procedure employed in our study was selected to provide a way to evaluate healing of the cement-bone composite quantitatively. Although the proximal part of the tibia may primarily resist compressive loads, this mode of loading would not have revealed the progressive healing response. As the calcium phosphate cement has compressive properties equal to those of trabecular bone, the compressive properties of the composite (cement and bone) would be nearly normal immediately after the operation. In fact, this was documented in a pilot study in our laboratory. As a result, the combination of the normal activity of the animal and the torsional tests provided a more suitable means for documenting the effect of the cement on fracture-healing.
Testing of the femoral sites along the medial-lateral axis was carried out in order to evaluate material properties over time in bounded homogenous specimens. Although the femoral compression tests resulted in differences between the two materials at twenty-four hours and at eight weeks, large standard deviations were observed. These variations were due, in part, to the amount and distribution of the cement or allograft contained within the specimens. These variations could be substantially explained by difficulties in localizing the grafts and in machining the specimens along the same axes as defined during the operation.
The histological fields that were chosen for morphological analysis consisted of the defect only and did not include the surrounding host bone. This allowed for a more objective choice of boundary. In addition, the host bone was studied qualitatively and was found to be normal with no signs of stress-shielding.
The preferential replacement of the cement in the cortical regions, in conjunction with the increased rate of resorption in the tibial sites compared with that in the femoral sites, strongly suggests that the healing process is influenced by mechanical, geometric, and physiological factors. This hypothesis is based on three interpretations. First, the most dominant feature that contributes to whole-bone strength is the geometry and dimension of the outer surface of the bone. To attain the greatest gain in whole-bone strength, an intact outer cortex would be the highest priority. The observation that the earliest and most rapid replacement of the cement occurs on the outer cortex is evidence of a mechanical driving force. Second, the femoral sites, because of their cylindrical shape and eight-millimeter length, had a smaller ratio of surface area to volume than the tibial defects, which were 3.5 millimeters high. Therefore, a geometrically based theory would tend to be partially supported by the reduced rate of replacement at the femoral sites. Third, there may have been physiological reasons for the observed differences in regional resorption and replacement of the cement. Differences in vascular supply, and perhaps in access to progenitor cells, may be important. For example, it might be argued that the outer cortical regions are replaced by bone faster because of the proximity to the reforming periosteum. However, whatever the reason, it is important to note that the cement is replaced by bone through a normal process of osteoclast-osteoblast coupling.
The results of this study demonstrate that FractureGrout is both osteoconductive and physiologically well tolerated. The material is stable once it hardens in situ and, although it is replaced by bone more slowly than allograft is, replacement is through normal mechanisms. The cement has important osteoconductive properties that enable it to become fused with host bone. These composite structures do not catastrophically lose mechanical integrity while being slowly replaced. Although we observed some cracks that may have been initiated by mechanical factors during use of the limb by the animals, they appeared to be associated with vascular ingrowth and had bone deposition on their surfaces. Finally, we found that the handling properties of the cement, which include its injectability, capacity for non-exothermic setting, ability to set in blood, and relatively rapid hardening time, make it a feasible material for orthopaedic applications.
In conclusion, this injectable calcium-phosphate cement material that sets in situ and attains properties similar to normal cancellous bone may be useful in metaphyseal regions for the repair of cancellous bone defects associated with severly comminuted, displaced, unstable fractures for which satisfactory reduction is difficult to achieve and maintain with use of conventional methods.
NOTE: The authors thank Morton Brown, Ph.D., Ming Jiang, M.D., and Doug Moore for their important contributions.