Selective cell proliferation can be controlled with CPC particle coatings.

To develop implantable, engineered, cartilage constructs supported by a scaffold, techniques to encourage rapid tissue growth into, and on the scaffold are essential. Preliminary studies indicated that human endothelial cells proliferated at different rates on different calcium phosphate ceramic (CPC) particles. Judicious selection of particles may encourage specific cell proliferation, leading to an ordered growth of tissues for angiogenesis, osteogenesis, and chondrogenesis. The goal of this study was to identify CPC surfaces that encourage bone and vascular cell growth, and other surfaces that support chondrocyte growth while inhibiting proliferation of vascular cells. Differences in bone and vascular cell proliferation were observed when using epoxy without embedded CPCs to encourage bone cells, and when three CPCs were tested, which encouraged vascular cell proliferation. One of these (CPC 7) also substantially depressed cartilage cell proliferation. Only one small-diameter crystalline CPC (CPC 2) supported rapid chondrocyte proliferation, and maintained the cartilage cell phenotype.

Sensate scaffolds can reliably detect joint loading.

Treatment of cartilage defects is essential to the prevention of osteoarthritis. Scaffold-based cartilage tissue engineering shows promise as a viable technique to treat focal defects. Added functionality can be achieved by incorporating strain gauges into scaffolds, thereby providing a real-time diagnostic measurement of joint loading. Strain-gauged scaffolds were placed into the medial femoral condyles of 14 adult canine knees and benchtop tested. Loads between 75 and 130 N were applied to the stifle joints at 30 degrees, 50 degrees, and 70 degrees of flexion. Strain-gauged scaffolds were able to reliably assess joint loading at all applied flexion angles and loads. Pressure sensitive films were used to determine joint surface pressures during loading and to assess the effect of scaffold placement on joint pressures. A comparison of peak pressures in control knees and joints with implanted scaffolds, as well as a comparison of pressures before and after scaffold placement, showed that strain-gauged scaffold implantation did not significantly alter joint pressures. Future studies could possibly use strain-gauged scaffolds to clinically establish normal joint loads and to determine loads that are damaging to both healthy and tissue-engineered cartilage. Strain-gauged scaffolds may significantly aid the development of a functional engineered cartilage tissue substitute as well as provide insight into the native environment of cartilage.

An instrumented scaffold can monitor loading in the knee joint.

No technique has been consistently successful in the repair of large focal defects in cartilage, particularly in older patients. Tissue-engineered cartilage grown on synthetic scaffolds with appropriate mechanical properties will provide an implant, which could be used to treat this problem. A means of monitoring loads and pressures acting on cartilage, at the defect site, will provide information needed to understand integration and survival of engineered tissues. It will also provide a means of evaluating rehabilitation protocols. A "sensate" scaffold with calibrated strain sensors attached to its surface, combined with a subminiature radio transmitter, was developed and utilized to measure loads and pressures during gait. In an animal study utilizing six dogs, peak loads of 120N and peak pressures of 11 MPa were measured during relaxed gait. Ingrowth into the scaffold characterized after 6 months in vivo indicated that it was well anchored and bone formation was continuing. Cartilage tissue formation was noted at the edges of the defect at the joint-scaffold interfaces. This suggested that native cartilage integration in future formulations of this scaffold configured with engineered cartilage will be a possibility.

TGF-beta1-enhanced TCP-coated sensate scaffolds can detect bone bonding.

Porous polybutylene terephthalate (PBT) scaffold systems were tested as orthopedic implants to determine whether these scaffolds could be used to detect strain transfer following bone growth into the scaffold. Three types of scaffold systems were tested: porous PBT scaffolds, porous PBT scaffolds with a thin beta-tricalcium phosphate coating (LC-PBT), and porous PBT scaffolds with the TCP coating vacuum packed into the scaffold pores (VI-PBT). In addition, the effect of applying TGF-beta1 to scaffolds as an enhancement was examined. The scaffolds were placed onto the femora of rats and left in vivo for 4 months. The amount of bone ingrowth and the strain transfer through various scaffolds was evaluated by using scanning electron microscopy, histology, histomorphometry, and cantilever bend testing. The VI-PBT scaffold showed the highest and most consistent degree of mechanical interaction between bone and scaffold, providing strain transfers of 68.5% (+/-20.6) and 79.2% (+/-8.7) of control scaffolds in tension and compression, respectively. The strain transfer through the VI-PBT scaffold decreased to 29.1% (+/-24.3) and 30.4% (+/-25.8) in tension and compression when used with TGF-beta1. TGF-beta1 enhancement increased the strain transfer through LC-PBT scaffolds in compression from 9.4% (+/-8.7) to 49.7% (+/-31.0). The significant changes in mechanical strain transfer through LC-PBT and VI-PBT scaffolds correlated with changes in bone ingrowth fraction, which was increased by 39.6% in LC-PBT scaffolds and was decreased 21.3% in VI-PBT scaffolds after TGF-beta1 enhancement. Overall, the results indicate that strain transfer through TCP-coated PBT scaffolds correlate with bone ingrowth after implantation, making these instrumented scaffolds useful for monitoring bone growth by monitoring strain transfer.

The effect of proximally and fully porous-coated canine hip stem design on bone modeling.

Porous coated canine femoral hip replacement implants were evaluated for biological fixation by bone ingrowth and the effect of the extent of porous coating on bone modeling. The Co-Cr alloy implants were either fully porous coated or coated only on the proximal 40% of the stem. Two implants of each type were studied 9, 16, and 36 months after surgery. Implant fixation and bone modeling were assessed radiographically throughout the implant periods and histologically after the test animals were killed. All 12 implants appeared stably fixed within the femur and were bone-ingrown in the porous region. Radiographic features such as proximal medial and anterior cortical thinning, proximal cancellous bone hypertrophy, and new endosteal bone formation near the stem tip were noted within the first postoperative year, with no appreciable change thereafter. The extent of proximal cortical thinning varied from virtually none to as much as 40%, being more prominent with the proximally coated implants at 16 months and with the fully coated implants at 36 months. Of consistent note was cancellous hypertrophy at the junction of porous and smooth implant surfaces with proximally coated implants and new endosteal bone formation and ingrowth at the stem tip of fully coated implants. These results indicate that the proximally porous-coated implant design causes increased proximal stress transfer, but this does not necessarily preclude proximal cortical resorption.

A long-term in vivo bone strain measurement device.

Bone remodeling resulting from implant insertion has been attributed to changes in the bone's strain state. Since remodeling takes several months, it was this study's purpose to develop a long-term in vivo strain sensor. Porous surfaced metal tabs were attached to a standard strain gauge. Two standard gauges and the porous tabbed gauge were attached to one femur and three standard gauges to the contralateral femur of a greyhound. Tissue ingrowth provided an attachment mechanism for the porous tabbed gauge in vivo. Gauge measurements were compared to those from the standard gauges. Post sacrifice testing allowed further comparisons. After histological preparation the femoral section shapes as well as the gauge locations were examined and photographed. The porous tabbed gauge remained bonded and sensed strain throughout the 8-week implantation period, while the standard gauges debonded and were unable to detect strain after 3 weeks. During testing, the measurements from the porous tabbed gauge were lower than those from the standard gauges. This was consistent with the histology which indicated that fibrous tissue had invaded the gauge's porous surface. Although the use of tissue ingrowth as an attachment mechanism seems to be worthwhile, a means of insuring bone ingrowth is necessary.

In vivo strain analysis of the greyhound femoral diaphysis.

Subminiature single element and rosette strain gauges used for deformation measurement were prepared for surgical implantation using a technique published previously (Szivek JA, Magee FP. J Invest Surg. 1989;2:195-206). During surgery, gauges were placed on the anterior, lateral, and medial aspects of the mid-diaphysis of one femur in six greyhounds. Motion and gait analyses were performed to ensure uniform weight bearing prior to strain monitoring. In vivo strain measurements were obtained during normal gait at several speeds on a treadmill. After a 3-month holding period, strain gauges that were implanted on the contralateral femur were monitored. All animals were euthanized and both their femora explanted. Following embedding and histological preparation of the explanted femora, strain measurements were plotted on diagrams of the section shapes of the mid-diaphysis of each femur. Strain distribution diagrams indicated that peak strain levels and strain distributions changed during different phases of gait. Increases in gait speed increased the peak strain levels. In addition, the anterior rather than anterior-lateral aspect of the femur exhibited the highest strain during midstance. Measurements taken from rosette gauges indicated that the principal compressive strain direction was oriented slightly off axis to the long axis of the femur. Measurements from gauges placed along the length of the femur indicated an average strain change of 22.3 microstrain +/- 12.2% over a 2-cm length in the mid-diaphysis. These measurements provide a baseline describing the strain state of the greyhound femur and can be used in computer modeling.

Strain redistribution in the canine femur resulting from hip implants of different stiffnesses.

Bone remodeling adjacent to orthopedic implants has been attributed to bone strain changes. Although many animal studies have assessed bone remodeling near implants, the altered bone strains and even the strains in the intact bone prior to implantation have not been mapped extensively. Instead, bone changes are often correlated with implant stiffnesses. In this study, a benchtop loading system was developed using measurements from in vivo strain analysis to simulate physiologic loading of a canine femur. The effect on bone strains of three different stiffness canine hip implants with the same anatomic shape were compared by taking measurements from the proximal greyhound femur during loading. Peak compressive and tensile strains of the order of 200 to 400 microstrain were measured in the intact and implanted femora. The measurements indicate that during simulated in vivo loading, none of the implants substantially alter the normal strain state of the bone. If initial axial strains significantly affect the remodeling response of bone, the similarity of measurements with the different implants in place suggests that the same remodeling response would be expected to both the stiffest and least stiff implant, as has been noted in animal studies adjacent to the intermediate stiffness implant. It also suggests that this implant shape and initial bone implant interface condition can compensate for strain reductions expected near stiff straight-stemmed implants.

In vivo strain measurements collected using calcium phosphate ceramic-bonded strain gauges.

Identification of the strains and the strain changes caused by implants is critical to the understanding of bone remodeling and can identify design changes needed to prevent bone loss near orthopedic implants. Calcium phosphate ceramic (CPC) coated strain gauges have been developed to allow long-term in vivo strain measurements. Previously used cyanoacrylate-bonded gauges have uncharacterizable sensing accuracy because the adhesive is resorbed from the instant it is placed in vivo. In this study CPC-coated strain gauges were used to measure physiologically "normal" bone strains collected from the proximal femora of dogs at a series of gait speeds and the postmortem sensing accuracy of the gauges was evaluated. Three male dogs were surgically implanted with up to six wired CPC-coated strain gauges placed around the circumference of their proximal femora. The dogs were trained to run on a treadmill, and in vivo strain measurements were collected following a 12-week period. The animals were tetracyline labeled and then euthanized and their femora explanted. Gauges were attached with cyanoacrylate to the intact contralateral control femora in the same position as the CPC-coated gauges on the test femora. Both femora were tested in cantilever bending to assess the functionality of the gauges and quality of the CPC-bone bond. After testing, all bones were embedded, sectioned, and ground. Sections from each femur were stained with mineralized bone stain and examined with transmitted and ultraviolet light to assess bone formation. Additional sections were examined with backscattered electron microscopy to confirm bone bonding to coatings. Wired gauges attached with the CPC coatings measured strain patterns during gait at several treadmill speeds. Patterns were similar and peak strains the same over a 2-week period. Mechanical testing showed bonding of CPC-coated gauges, and histologic examination showed intimate contact between gauge coatings and bone surfaces. Further development of CPC-bonded strain gauges is expected to result in a measurement system that provides ease of placement, and consistent longer term bone strain measurements with characterizable accuracy.

Bone remodeling and in vivo strain analysis of intact and implanted greyhound proximal femora.

Pre- and poststudy motion and gait analyses of eight size-matched male greyhounds confirmed uniform loading of their femora. Subminiature strain gages implanted on the intact inferior and anterior aspects of the femoral neck in six greyhounds indicated in vivo strain variations among test animals. Motion and gait analyses confirmed uniform loading of femora following unilateral hemiarthroplasty with cobalt-chromium hip implants. In vivo strain measurements adjacent to the implants indicated large variations among test animals. A consistent direction of strain change relative to the intact femur was noted, even though strain changes varied in magnitude. Image analysis of microradiographs indicated insignificant differences in the cortical areas of implanted and intact femora. Extensive new trabecular bone formation was noted along the implant in the endosteal cavity and correlated with a combination of implant placement and exercise level. Most of the bone was formed with centrally placed implants in exercised dogs, and the least with stems on the medial neck surface in rested dogs. Iliac crest biopsies indicated that bone formation rates slowed in rested animals and remained constant throughout the study in exercised animals. All implanted femora had a thin (< 1 mm thick) aligned fibrous tissue layer separating the implant from bone. It varied in thickness as a function of the aspect of the implant. Exercised dogs had a larger proportion of fibrous tissue on the anterior and posterior aspects, while rested dogs had a larger proportion of fibrous tissue on the medial and lateral aspects.

A mechanical and histomorphometric analysis of bone bonding by hydroxyapatite-coated strain gages.

Identification of the strains controlling bone remodeling is important for determining ways to prevent bone loss due to load deprivation, or implant placement. Long-term monitoring of strains can potentially provide the best information. Glues are resorbed within 2-3 weeks. Two formulations of microcrystalline hydroxyapatite (HA) were used to attach strain gages to rat femora to assess their long-term in vivo strain measurement capability. Seven male rats received HA-coated gages, and 2 animals underwent a sham procedure. The gages were prepared using a published technique and placed on the antero-lateral aspect of the left femora. After 6-7 weeks, the animals were euthanized and both femora explanted. Gages were attached to the right femora with cyanoacrylate. All femora were tested in cantilever bending, then embedded, sectioned, and stained with mineralized bone stain. The undecalcified sections were examined using transmitted and ultraviolet light microscopy. Mechanical testing showed one HA formulation provided 70-100% bonding. Histology showed intimate contact between the gage and bone surface. Histomorphometry indicated increased bone activity under the gage compared to the remaining bone, the controls, and the shams. The results indicate that microcrystalline HAs bond to bone quickly and can allow long term in vivo measurements.

An experimental method for the application of lateral muscle loading and its effect on femoral strain distributions.

Experimental models that have been used to evaluate hip loading and the effect of hip implants on bone often use only a head load and abductor load. Anatomic considerations and in vivo measurements have lead several investigators to suggest that these models are inaccurate because they do not incorporate the loads imposed by additional muscles. The aim of this study was to evaluate the strains in the proximal and mid diaphysis of the femur for five hip loading models, one with a head load and abductor load only and four which incorporated lateral muscle loads as well. Head load to body weight load ratios were used to evaluate the physiologic accuracy of these models and strains were compared to determine the extent of strain changes as a function of model complexity. All models which incorporated additional lateral muscle loads more accurately simulated head load to body-weight load ratios than the simple abductor-only model. The model which incorporated a coupled vastus lateralis and iliotibial band load in addition to the abductor load provided the simplest configuration with a reasonable body-weight to head-load ratio.

A biomechanical evaluation of three forms of internal fixation used in ankle arthrodesis.

A biomechanical study was undertaken to evaluate the relative stability of three types of internal fixation used for ankle arthrodesis. Crossed screw fixation, RAF fibular strut fixation, and T-plate fixation were tested in 30 cadaver ankles using an MTS machine. T-plate fixation consistently provided the stiffest construct when compared with the other types of fixation. Failure occurred by distraction of bony surfaces, posterior to the plane of fixation, in the crossed screw and RAF groups. In contrast, failure in the T-plate group occurred through compression of bone anterior to the midcoronal plane of the tibia. Although the stability of fixation is only one factor in determining the success or failure of ankle arthrodesis, the results of this study would support T-plate fixation over the other forms tested.

Trabecular scaffolds created using micro CT guided fused deposition modeling.

Free form fabrication and high resolution imaging techniques enable the creation of biomimetic tissue engineering scaffolds. A 3D CAD model of canine trabecular bone was produced via micro CT and exported to a fused deposition modeler, to produce polybutylene terephthalate (PBT) trabeculated scaffolds and four other scaffold groups of varying pore structures. The five scaffold groups were divided into subgroups (n=6) and compression tested at two load rates (49 N/s and 294 N/s). Two groups were soaked in a 25 °C saline solution for 7 days before compression testing. Micro CT was used to compare porosity, connectivity density, and trabecular separation of each scaffold type to a canine trabecular bone sample. At 49 N/s the dry trabecular scaffolds had a compressive stiffness of 4.94±1.19 MPa, similar to the simple linear small pore scaffolds and significantly more stiff (p<0.05) than either of the complex interconnected pore scaffolds. At 294 N/s, the compressive stiffness values for all five groups roughly doubled. Soaking in saline had an insignificant effect on stiffness. The trabecular scaffolds matched bone samples in porosity; however, achieving physiologic connectivity density and trabecular separation will require further refining of scaffold processing.

A testing technique allowing cyclic application of axial, bending, and torque loads to fracture plates to examine screw loosening.

Orthopaedic internal fracture fixation plates are subjected to combined axial, bending, and torsional loads in vivo which can cause screw loosening and implant failure. This paper outlines a relatively simple technique which allows controlled application of combined axial, bending, and torsional loading to examine the loosening rate of cortical screws used to attach these plates. Fiber reinforced polycarbonate rods with a tensile strength similar to that of cortical bone were cut at half their length to simulate fractured tibii. These were compression plated using a standardized technique and placed in a loading fixture. Joint loads at the knee determined from force plate analysis and statics were applied to a plated fixture during testing. The design of the fixture allowed adjustment of the proportion of bending and torsional loads applied to the test samples. It also allowed a reproducible means of applying a predetermined axial, bending, and torsional load. Screw loosening following cyclical loading was evaluated by measuring the amount of angular displacement required to retighten screws to a prescribed torque value. A torque wrench was modified to allow the measurement of these displacements.

Strain transfer between a CPC coated strain gauge and cortical bone during bending.

The finite element method was used to simulate strain transfer from bone to a calcium phosphate ceramic (CPC) coated strain gauge. The model was constructed using gross morphometric and histological measurements obtained from previous experimental studies. Material properties were assigned based on experiments and information from the literature. Boundary conditions simulated experimental cantilever loading of rat femora. The model was validated using analytical solutions based on the theory of elasticity as well as direct comparison to experimental data obtained in a separate study. The interface between the bone and strain gauge sensing surface consisted of layers of polysulfone, polysulfone/CPC, and CPC/bone. Parameter studies examined the effect of interface thickness and modulus, gauge geometry, partial gauge debonding, and waterproofing on the strain transfer from the bone to the gauge sensing element. Results demonstrated that interface thickness and modulus have a significant effect on strain transfer. Optimal strain transfer was achieved for an interface modulus of approximately 2 GPa. Strain transfer decreased consistently with increasing interface thickness. Debonding along the lateral edges of the gauge had little effect, while debonding proximal and distal to the sensing element decreased strain transfer. A waterproofing layer decreased strain transfer, and this effect was more pronounced as the modulus or thickness of the layer increased. Based on these simulations, specific recommendations were made to optimize strain transfer between bone and CPC coated gauges for experimental studies.

Pullout strengths of cannulated and noncannulated cancellous bone screws.

The pullout strengths of large diameter cannulated and noncannulated cancellous screws were tested in a synthetic polyurethane foam. The foam was fabricated to have mechanical properties equivalent to human cancellous bone and was characterized by compression testing before screw pullout. Long and short thread commercially available screws from four manufacturers were tested. In screws with short threads (16-22 mm), there was no difference in holding power among the four cannulated screw designs. However, the short thread noncannulated screw performed significantly better than the short thread cannulated screw with the lowest pullout strength. There were statistically significant differences in holding power among the different long thread (32-40 mm) cannulated screw designs. Additionally, the long thread noncannulated screw had better holding power than several of the long thread cannulated screws. No differences in pullout strengths between comparably sized cannulated and noncannulated screws produced by the same manufacturer were found, and all long thread screws had significantly greater holding power than all short thread screws. There was no demonstrable effect on holding power when screws were inserted with or without tapping. Thread surface area was found to be a reasonable predictor of holding power.

Average and peak contact stress distribution evaluation of total knee arthroplasties.

Seven total knee arthroplasty systems were tested to determine contact stress patterns and contact areas using a calibrated Fuji film stress analysis technique. Knees were loaded to 2,000 N (204 kg) at 15 degrees, 60 degrees, 90 degrees, and 135 degrees flexion at 24 and 37 degrees C. Evaluation of stresses at 37 degrees C at 15 degrees and 60 degrees using an average contact stress assessment technique indicated that the LCS meniscal bearing knee system, (DePuy, Warsaw, IN), the AMK knee with a constrained insert (DePuy), and the PFC knee with a posterior-lipped insert (Johnson and Johnson, Raynham, MA) had the lowest average contact stresses (near or below 10 MPa). The PFC with a regular insert (Johnson and Johnson) the Ortholoc II (Dow Corning Wright, Arlington, TN), and the AMK with a regular insert (DePuy) had intermediate contact stresses. The AMK with a Hylamer-M insert (DePuy) and the MG II (Zimmer, Warsaw, IN) had the highest average contact stresses (near or above 20 MPa). A stress-calibrated Fuji film measurement technique has shown that an assessment of ranges of contact stress provides much more information about regions of expected wear than an assessment of average contact stresses. Testing of the tibiofemoral articulation of artificial knees revealed that all knees had some areas of contact with maximum stresses in excess of 15 MPa. As the yield strength of ultrahigh-molecular-weight polyethylene is approximately 15 MPa, all tibial inserts could wear to some extent. Peak contact stresses at four test angles of the AMK, Series 7000 (Osteonics, Allendale, NJ:) Genesis (Smith & Nephew Orthopaedics, Memphis, TN), and MG II patellofemoral articulations were high (above 30 MPa). Contact areas varied from line-shaped to bilateral circular or elliptical shapes. The LCS knee system experienced substantially lower patellofemoral contact stresses and larger contact areas. Changes in conformity of knee designs are warranted to overcome wear problems. Peak contact stresses measured from the LCS meniscal bearing tibiofemoral and patellofemoral joint were in excess of 30 MPa in some areas at low flexion angles. This design does create large areas of contact at very low contact pressures, however, and for this reason is expected to wear less than other designs.