Specimens
Twenty-six paired, fresh-frozen cadaveric feet were disarticulated at the ankle joint, and soft tissue was removed from the medial, lateral, and posterior aspects of the calcaneus. The heel pad and the plantar fascia were preserved. Eight pairs were from women and five pairs, from men. The average age of the donors at the time of death was seventy years (range, fifty-two to seventy-seven years). The dome of the talus was potted in urethane over adjuvant screw fixation to allow for a smooth, flat surface for loading of the specimen onto a testing machine.
Baseline Analysis
A custom extensometer was mounted on the medial aspect of the calcaneus between two Schanz pins that had been placed along the superior and inferior aspects of the calcaneus (Fig. 1-A). The extensometer was calibrated and was found to be accurate to 0.05 millimeter. The specimen was then mounted in an MTS biaxial testing machine (Bionix, model 858.02; MTS, Eden Prairie, Minnesota) and was preconditioned by loading it for ten cycles from zero to 100 newtons at one hertz. Baseline data for stiffness were obtained by loading the specimen from zero to 350 newtons for ten cycles at one hertz. The 350-newton load approximates half of the average body weight (seventy kilograms). Deformation of the calcaneus was measured directly from the extensometer that was attached directly to the medial calcaneal wall.
Preparation of the Fracture
A clinically relevant fracture was made in each specimen by placing stress-risers along the medial, lateral, and posterior aspects of the calcaneus. The stress-risers were made with a narrow, 6.5-millimeter osteotome along fracture lines that are typically encountered in a displaced intra-articular calcaneal fracture of the joint-depression type. The stress-risers were placed from the crucial angle of Gissane toward the plantar aspect of the calcaneus. The stress-risers along the lateral wall outlined a small area to approximate the blowout of the lateral wall in a calcaneal fracture (Fig. 1-B). Stress-risers were also placed along the dorsal aspect of the posterior tuberosity (to simulate the fracture line) and along the medial wall of the calcaneus, 1.5 centimeters inferior to the sustentaculum tali (Fig. 1-A). In addition, they were placed through the middle portion of the posterior facet through the subchondral bone to create a type-IIB fracture as classified according to the system of Sanders et al.20. After these stress-risers were made through the cortical bone, the underlying cancellous bone was similarly divided to produce a nondisplaced fracture along the stress-risers. The specimen was then mounted by the potted talus on a materials testing machine (model 8521; Instron, Canton, Massachusetts), with the foot resting on the base-plate. A block was placed along the lateral aspect of the calcaneus to prevent it from migrating into valgus angulation during loading. In displacement control, the specimen was compressed two centimeters in 0.5 second.
During the placement of the stress-risers and the insertion of the Schanz pins into the medial aspect of the calcaneus, the specimens were found to differ with regard to bone quality. Four pairs were found to be considerably osteoporotic, with poor-quality bone. The remaining nine pairs were classified as having good-quality bone. This determination was made in a qualitative manner on the basis of the resistance encountered when the osteotome was impacted. The specimens with poor-quality bone were severely osteoporotic, which allowed the osteotome to be forced through the cortical bone manually, without the use of a mallet. The specimens were evaluated both as a group and, independently, as subgroups on the basis of whether they had good or poor-quality bone.
Fixation of the Fracture
One specimen of each pair, randomly selected after the fracture, was repaired with conventional internal fixation with a bone graft (the control). The other specimen was repaired with conventional internal fixation augmented with SRS bone cement. The halves of the posterior facet of the control specimen were reduced and were secured with two subchondral, 3.5-millimeter cortical-bone lag-screws inserted from the lateral aspect of the calcaneus. The body of the calcaneus was then reconstructed, and the reduction was maintained with an appropriately contoured calcaneal Y-plate (Synthes, Paoli, Pennsylvania) secured with 3.5-millimeter cortical-bone screws of the appropriate length. Any osseous void was packed with morseled cancellous bone graft that had been removed from a fresh-frozen calcaneus.
The contralateral specimen was fixed in an identical fashion except for the use of SRS bone cement instead of bone graft. After the fracture was reduced and the hardware was placed, a hole was made in the lateral wall through one of the fracture lines. The walls of the osseous defect in the body of the calcaneus were compressed through this hole with the back of a curet to maximize the strength of the osseous bed that would be supporting the bone-cement mass. SRS bone cement was then mixed according to the manufacturer's specifications, and it was injected from the deepest to the most superficial part of the osseous defect through a 12-gauge needle while the cement was in its paste-like form. This process mimicked the clinical procedure with a lateral exposure (Fig. 2). The SRS bone cement set to a hardened mass in approximately ten minutes, during which time the specimen was not disturbed to avoid making discontinuities or voids in the cement6. SRS bone cement has 50 percent of its ultimate compressive strength after one hour of curing and has 100 percent after twelve hours of curing under physiological conditions (37 degrees Celsius and 100 percent humidity)6. The cement-augmented specimens were placed in a bath of phosphate-buffered saline solution at 37 degrees Celsius for twenty-four hours to allow full curing of the cement. The control specimens were placed in the bath for twenty-four hours to allow for uniformity of testing conditions.
Mechanical Testing
After curing for twenty-four hours, each specimen was placed in the MTS Bionix biaxial testing machine and was subjected to the same cyclical testing as had been used for the intact specimen (that is, ten cycles from zero to 100 newtons at one hertz and ten cycles from zero to 350 newtons at one hertz). Fatigue cyclical testing was then performed with a load of zero to 350 newtons at one hertz for 1000 cycles or until complete failure of the specimen (seven millimeters of displacement). Displacement of as much as seven millimeters (the maximum displacement that could be recorded with the extensometer) was measured. We believed that measurement of more displacement would not be clinically relevant because this degree of displacement represents a complete failure. We also recorded when two millimeters or more of displacement occurred because it represents a more frequently used measure of an adequate reduction of the joint surface clinically21. Deformation per cycle (millimeters per cycle), first-cycle deformation (at a load of zero to 350 newtons), and number of cycles to complete failure were calculated for each specimen.
Radiographic Evaluation
Lateral and axial radiographs were made of each specimen before testing, as a baseline, and after each specimen was fractured, to assess the anatomy of the fracture (Figs. 3-A and 3-B). After reconstruction, the adequacy of the reduction was assessed with additional lateral and axial plain radiographs as well as a coronal computerized tomographic scan performed from posterior to anterior in three-millimeter cuts (Figs. 3-C, 3-D, and 3-E). The filling of the osseous void with the SRS bone cement and the reduction of the articular surface were evaluated with the computerized tomographic scan. More plain radiographs were made after cyclical testing and after testing to failure. Böhler's angle2 was measured with a protractor on each lateral radiograph, and the values for the intact specimen, after fracture, and after reconstruction were compared between the control and cement-augmented groups. The average values and standard errors were calculated for each group.
Analysis of the Data
The average deformation per cycle, first-cycle deformation, and number of cycles to complete failure were calculated and compared between the treatment groups. The average deformation per cycle was the average displacement measured at 350 newtons of load for 1000 cycles or until failure (seven millimeters of displacement). Load-to-failure data were further analyzed to determine when two millimeters or more of displacement had occurred. The results of treatment were compared for all of the specimens, together as well as after they had been divided into subgroups according to whether they had good or poor-quality bone. The continuous variables were tested for normality and equal variance before parametric statistical analyses were performed. Normality was not found in any of the data sets, so a nonparametric test (Wilcoxon signed-rank test) was applied.
Large differences in stability between the specimens augmented with SRS bone cement and those fixed with standard techniques were evident when all of the specimens were analyzed together and when they were separated into subgroups according to whether they had good or poor-quality bone. The cement-augmented specimens displaced an average of 0.00195 millimeter per cycle compared with 1.013 millimeters per cycle for the control group (p < 0.005) (Fig. 4). In the group with good-quality bone, the cement-augmented specimens displaced 0.000727 millimeter per cycle compared with 0.331 millimeter per cycle for the control specimens (p < 0.03). In the group with poor-quality bone, the cement-augmented specimens displaced 0.00476 millimeter per cycle compared with 3.68 millimeters per cycle for the control specimens (p < 0.07) (Fig. 4). The data for each specimen pair are presented (Fig. 5). For all three groups of data, a relatively consistent difference of nearly three orders of magnitude was noted in the average deformation per cycle between the cement-augmented and the control group.
Analysis of the first-cycle deformation revealed an average value of 0.478 millimeter (range, 0.003 to 1.33 millimeters) for the cement-augmented group compared with 3.09 millimeters (range, 0.003 to 6.96 millimeters) for the control group (p < 0.01). The number of cycles to complete failure (seven millimeters of displacement) was 976 ± 128 for the cement-augmented group compared with 584 ± 511 for the control group (p < 0.03). Of the thirteen cement-augmented specimens, only one (a specimen with poor-quality bone) failed before 1010 cycles to 350 newtons, whereas six control specimens (three with good-quality bone and three with poor-quality bone) failed before 1010 cycles to 350 newtons (p < 0.07).
Further analysis of the load-to-failure data for the good-quality-bone specimens revealed that six of the nine control specimens displaced at least two millimeters (range, 2.5 to seven millimeters) before 1010 cycles compared with only one of the nine cement-augmented specimens (3.2 millimeters) (p = 0.05).
Radiographic evaluation revealed that a reproducible displaced intra-articular fracture of the calcaneus with impaction of bone (type IIB according to the classification system of Sanders et al.20) had been created in all of the specimens. The coronal computerized tomographic scans showed that a nearly anatomical reconstruction had been achieved, with less than one millimeter of articular incongruity, in all of the specimens. Böhler's angle was 35.9 ± 1.4 degrees (average and standard error) in the cement-augmented group and 34.4 ± 1.4 degrees in the control group before the fracture, 13.9 ± 1.4 degrees in the cement-augmented group and 13.2 ± 1.8 degrees in the control group after the fracture, and 34.5 ± 1.2 degrees in the cement-augmented group and 35.9 ± 2.7 degrees in the control group after the reconstruction. No differences between the treatment groups could be detected with the numbers available. The computerized tomographic scans confirmed complete filling of the bone defect (without voids) with the SRS bone cement in all of the specimens. The radiographs revealed no visible damage to the bone-cement mass in any specimen after the fatigue or failure testing. In the cement-augmented specimens, failure occurred with impaction of the cancellous bone around the bone-cement mass. In the control specimens, failure occurred with impaction of the bone primarily in and around the defect that had been filled with bone graft. None of the specimens in either treatment group had visible or radiographic evidence of an articular incongruity following failure. None of the hardware failed. These findings were consistent with those on visual inspection of the specimens.
With current methods for operative treatment of displaced intra-articular fractures of the calcaneus, weight-bearing must be delayed for ten to twelve weeks to allow for initial consolidation of the fracture because of the lack of stability of the operative construct. This delayed weight-bearing contributes to prolonged disability, increased disuse osteoporosis, and increased muscular atrophy of the involved extremity. Also, some surgeons use autogenous bone graft, which can increase the morbidity and the duration of hospitalization9. The present study demonstrated a significant increase in the compressive fatigue strength of an in vitro calcaneal fracture-repair construct that included filling of the entire defect with SRS bone cement (p < 0.005). We hypothesize that the use of this material could increase the initial strength of the fracture construct in vivo, which could lead to more-rapid rehabilitation and a shorter period of disability.
Although calcium phosphate materials such as coralline hydroxyapatite have been widely investigated as osteoconductive compounds, none have served as a structural substance. SRS bone cement is injectable and can fill a void of any shape, providing almost immediate structural stability. The compressive strength of SRS bone cement has been documented to be fifty-five megapascals, which is equivalent to the strength of intact cancellous bone6. In contrast to currently available calcium phosphate materials, SRS cancellous bone cement consists primarily of type-B dahllite, which is very similar to the bone mineral found in human bone6.
An in vivo canine study demonstrated that SRS bone cement that had been used in a defect in the proximal tibial metaphysis had immediate load-bearing capacity and did not inhibit bone-healing8. Cortical healing between four and eight weeks resulted in whole-bone torsional strength equivalent to that of the contralateral, control tibia. Additionally, cancellous defects in the distal femoral metaphysis that had been filled with SRS bone cement had greater compressive strength and stiffness throughout the first two-week period after the procedure than defects treated with morseled bone graft. Compressive strength was equivalent to that of the intact cancellous bone of the distal aspect of the femur. The SRS bone cement was partially remodeled over a four-month period through an osteoclast cell-mediated remodeling process, with the composite of bone and remaining SRS bone cement retaining the strength and stiffness of the initial construct throughout the time to cortical healing. At thirty-two and seventy-eight weeks after the procedure, normal amounts of cortical and cancellous bone had been reestablished.
Previous studies have documented that SRS bone cement, when used with hardware, can increase the strength of a fracture construct. Stankewich et al. demonstrated an increase of 50 percent in the load to failure of a fracture construct with cannulated screws and SRS bone cement in a femoral-neck fracture model22. Moore et al. found similar screw-fixation stiffness, yield load, and energy to failure when comparing screw fixation augmented with SRS bone cement and that augmented with methylmethacrylate17. Lotz et al. demonstrated a 50 percent increase in energy to failure during cyclical loading of pedicle screws that had been augmented with SRS bone cement in human cadaveric bone15. Two clinical studies of the use of SRS bone cement in the treatment of fractures of the distal aspect of the radius demonstrated impressive results, with maintenance of fracture reduction without the use of hardware or bone graft11,12. These studies established the biocompatibility of the bone cement in a human fracture model. Kopylov et al. found that patients who had been managed with SRS bone cement had an earlier return of function13.
A reproducible two-part displaced intra-articular fracture of the calcaneus with impaction of cancellous bone (type IIB according to the classification system of Sanders et al.20) was achieved in each of the specimens. Sangeorzan et al. previously developed a model of an intra-articular fracture of the calcaneus to determine the change in joint-contact stresses with varying degrees of displacement21. They performed a simple osteotomy without osseous impaction in order to evaluate the change in load distribution with varying degrees of articular step-off of the posterior facet. Carr et al. evaluated a series of embalmed cadaveric specimens in which an intra-articular fracture of the calcaneus had been created by dropping the specimen with a weight attached to an intramedullary rod in the tibia3. Their method yielded variable fracture patterns because they had not used stress-risers, and their study did not involve repair of the fracture. In another study, Carr et al. produced fractures in thirteen cadaveric specimens by dropping weights onto them while they were mounted in a fracture stand5. Only a small stress-riser was placed on the sinus tarsi. A variable fracture line through the posterior facet was made. They performed cyclical testing to 100 newtons for 500 cycles to simulate walking with crutches. In our study, the main fracture lines were first made with an osteotome and the calcaneus was subsequently impacted in a reproducible fashion on a testing machine by rapid loading in displacement control. In eight pilot specimens, we attempted to produce a fracture with less-extensive stress-risers, such as multiple perforations with a drill through the cortical bone. With impaction on the testing machine, the body of the calcaneus impacted but the posterior facet did not split. We were able to produce a reproducible split in the posterior facet only by first creating a nondisplaced fracture with an osteotome and subsequently impacting the specimen on the testing machine. The radiographic evaluation confirmed our visual observation that a two-part displaced intra-articular fracture of the posterior facet had been successfully made. Also, the fracture fragments displaced significantly, as reflected by the decrease in Böhler's angle after the fracture (p < 1 x 10-14).
The plain radiographs made after the procedure demonstrated a nearly anatomical reconstruction of the calcaneus with reconstitution of the joint surface and Böhler's angle. The computerized tomographic scan demonstrated a nearly anatomical reconstruction of the joint surface and good filling of the osseous void with SRS bone cement. The radiographs made after failure did not demonstrate any visible damage to any SRS bone-cement mass; this indicated that the mode of failure was through further impaction of the bone or pullout of the hardware as opposed to failure of the SRS bone cement. This finding was not surprising because SRS bone cement has been demonstrated to have a compressive strength as great as that of cancellous bone6.
We used paired specimens to minimize any differences between the reconstructed groups due to the quality of the bone. We divided the specimens according to a qualitative assessment of whether they had good or poor-quality bone to determine whether any differences that we observed were more prevalent with one type of bone quality. As expected, the specimens with poor-quality bone demonstrated less stability on testing in both the cement-augmented group and the control group. Although SRS bone cement could be helpful with poor-quality bone, it is possible for the bone to be so poor that operative fixation would be contraindicated. However, the relative difference between the control and cement-augmented groups was approximately the same in the good and poor-quality-bone groups.
We measured osseous displacement directly through the calcaneus by placing an extensometer on pins that were mounted on its superior and inferior aspects. This method allowed for much greater accuracy in measurement than there would have been with displacement of the actuator. When we attempted to measure through the actuator, we also measured compression of the calcaneal fat pad and displacement through the subtalar joint.
We loaded the specimens to 350 newtons because this value represents approximately half of the body weight of a seventy-kilogram person. We did not load the specimens to 700 newtons because we believed that it would lead to rapid failure in more of the control specimens. Our loading level was 3.5 times greater than that used by Carr et al.5. Because of this difference in loading and our differing methods for creation of the fracture, no meaningful comparison of the data from the two studies can be made.
The major limitation of our study is that the data were obtained in vitro and the results would have to be extrapolated to an in vivo setting. Our data reflected the initial compressive strength of these fracture constructs before any healing had occurred. Clearly, the healing process could alter the strength of the fracture repair. However, the healing process did not have an appreciable effect on the repair of a fracture augmented with bone cement in one canine study, in which no change in the rate of healing or the strength of a metaphyseal tibial defect was noted8. We believe that the difference in the stability of these reconstructions will justify a more intensive postoperative weight-bearing protocol.
In summary, in a model consisting of a clinically relevant two-part intra-articular fracture of the posterior facet with compaction of the underlying cancellous bone, augmentation with SRS bone cement dramatically increased the cyclical compression resistance in both good and poor-quality bone.