Design of the Femoral Implants
A collarless femoral stem was chosen on the basis of a radiographic study of femora in dogs. The stem was straight with a cylindrical cross section distally and a generally tapered metaphyseal section that had a slight anteroposterior flare and a straight lateral aspect (Fig. 1). The length of the stem was ninety millimeters, and the diameter was ten millimeters. Commercially pure titanium fiber-metal was diffusion-bonded to the stem from the shoulder to the tip. The extent of this porous coating was chosen to maximize bone ingrowth and thus increase the potential for stress-related bone resorption. The neck was machined to a number-3 Morse taper to accommodate a modular sixteen-millimeter cobalt-chromium-alloy head with a neck length of zero or three millimeters.
Two types of stems with the same configuration and external dimensions were manufactured (Fig. 1). The first type was a conventional all-metal design; it consisted of a machined core of surgical-grade titanium alloy (Ti-6Al-4V) with a porous coating of titanium fiber-metal. The second type was representative of the composite technology; it consisted of a forged cobalt-chromium core, an inner layer of polyaryletherketone polymer, and an outer layer, or wrap, of titanium fiber-metal. Both types of stems had a porous coating that was at least 0.75 millimeter thick and approximately 55 percent dense and that consisted of an interconnecting network of pores averaging approximately 350 micrometers in diameter.
Cross sections of the implants, obtained at one-centimeter intervals, were analyzed with regard to area and moments of inertia in order to compare the structural stiffness characteristics of the two types of stems. Calculations of bending stiffness were based on the product of the modulus of elasticity and the moment of inertia. The individual contributions to stiffness made by the solid-core material (titanium or cobalt-chromium alloy), intermediate polymer, and outer fiber metal (commercially pure titanium) were added to obtain the total stiffness (Table I). Cross sections of femora from six different dogs, in the size-range appropriate for implantation of a porous-coated prosthesis without cement, also were obtained at one-centimeter intervals for calculation of stiffness.
Over most of its length, the titanium-alloy stem had a stiffness that was comparable with the stiffness (range) of the canine femur (Fig. 2). The composite stem was approximately threefold less stiff than the femur proximally and approximately four to fivefold less stiff distally. Comparable differences in relative stiffness were found between the two types of stems with regard to bending in the sagittal plane.
The femoral components had a sixteen-millimeter-diameter modular head. These implants were paired with an acetabular component with an outer diameter of twenty-eight millimeters that was designed to be inserted without cement. The acetabular implants consisted of a titanium backing with a fiber-metal porous coating and a modular polyethylene insert (Fig. 1). Four screw-holes in the metal backing allowed supplementary fixation with 3.2-millimeter screws.
Experimental Protocol
In order to allow for a direct comparison between the osseous response to the titanium-alloy femoral stems and the osseous response to the cobalt-chromium-polyaryletherketone-composite femoral stems, a bilateral total hip arthroplasty was performed, in stages three to four weeks apart, with insertion of one prosthesis of each type in each animal. The dogs were carefully screened radiographically to determine the appropriate size and shape of the femoral component. An effort was made to include only animals in which a press-fit or near press-fit of the distal aspect of the stem within the intramedullary canal was likely. The titanium-alloy stem and the composite stem were used alternately in the first operative procedure to eliminate a time-related bias in the results.
The operation was performed with use of standard sterile techniques. The hip joint was exposed through a modified anterolateral approach. After a lateral skin incision had been made, the interval between the gluteus medius and the tensor fasciae latae was opened to expose the anterior aspect of the hip capsule. The gluteus medius and the vastus lateralis were elevated with a small flake of bone off the anterior portion of the greater trochanter.
After an anterior capsulotomy and sectioning of the ligamentum teres, the hip was partially dislocated and a provisional osteotomy of the femoral neck was performed. The acetabulum was exposed, excess soft tissue was resected, and the transverse acetabular ligament was partially excised. The acetabulum was progressively reamed under irrigation to a diameter of twenty-seven millimeters in order to create a one-millimeter press-fit. The metal backing of the component was impacted into position and was secured with one or two screws before insertion of the polyethylene liner.
Preparation of the femur started with insertion of a small tapered curet into the piriformis fossa in order to broach the femoral canal. The intramedullary canal was progressively enlarged in 0.5-millimeter increments to a diameter of 9.5 millimeters, creating a 0.5-millimeter press-fit for the stem. The femoral canal was rasped to prepare the metaphyseal region, and the stem was impacted forcefully into position. A modular femoral head was attached to the neck taper, and the hip was reduced. The trochanteric bone flake was reattached with two interrupted transosseous nonresorbable sutures. The incision was closed in a standard three-layer fashion with use of resorbable sutures.
Four dogs (eight hips) were studied for six months postoperatively, and four dogs (eight hips) were studied for eighteen months postoperatively. Previous studies of total hip arthroplasty in dogs have shown little change in the tissue response to a functioning implant, in terms of the extent of both tissue ingrowth and periprosthetic remodeling, after six months2,28,29. Twelve to eighteen months is sufficient for evaluation of the end point of bone-remodeling3,4,28,29. Anteroposterior and lateral radiographs of each hip were made before and within the first week after the operation. Subsequent radiographs were made at six-month intervals and at the end of the study period. The radiographs were inspected for canal fill (defined as the extent of filling measured at the isthmus), resorptive bone-remodeling, bone ingrowth, and regions of formation of radiopaque lines that would indicate fibrous encapsulation. Radiographic characteristics of the femur were documented by location according to the seven contiguous zones described by Gruen et al.11. Radiographs of the acetabular cup were examined for the presence or absence of periprosthetic radiolucency.
The animals were killed with an intravenous dose of Nembutal (pentobarbital). The femora and acetabula were removed and stripped of soft tissue, and contact radiographs were made with high-resolution film and use of a Faxitron apparatus (Hewlett-Packard, San Diego, California). The specimens were fixed in 70 percent ethanol and were processed for histological analysis of undecalcified thin sections. This involved dehydration in ascending solutions of ethanol, defatting in ether and acetone, and embedding in methylmethacrylate. Small holes were drilled in the femora and acetabula in locations that were not critical in order to facilitate penetration of the various solutions.
The femora were sectioned transversely at one-centimeter intervals, starting at the shoulder of the implant. Sections approximately two millimeters thick were obtained at nine intervals. The intervals were subdivided to correspond with the seven zones of Gruen et al.11 (Table II). The ninth interval was not included in the data analysis because it represented the nonporous tip of the stem. Additional thinning and polishing were performed before sputter-coating with gold-palladium to enable analysis with backscattered scanning electron microscopy. Sections from the third through seventh intervals were analyzed with regard to the volume fraction of bone ingrowth, defined as the percentage of the available porous space that was filled with new bone. Care was taken to account for the reduced thickness of the porous coating available for ingrowth in the composite implants, which was due to the partial infiltration with polyaryletherketone polymer. The paired bone-ingrowth data were compared between the two stem designs at each time-period. Statistical significance was assessed with use of analyses of variance and two-tailed paired Student t tests.
In addition, each radiograph of each section was examined for the presence of a thin radiopaque line encircling the stem but separated from it by a small gap of 0.5 to 1.0 millimeter. This characteristic has been documented in numerous radiographic and histological studies of porous metal-bone interfaces and is indicative of fibrous-tissue encapsulation7,16,23. Although each region was not evaluated histologically to confirm the presence of fibrous tissue, these sections were recorded as not having bone ingrowth. The percentage of sections that had a radiopaque line was tabulated, for each type of implant, and these data were compared with use of the chi-square test.
The serial thin sections from both femora were paired at each level, and radiographs were made. Computer-aided image analysis was used to digitize and quantify the cortical area defined by the periosteal and endosteal envelopes of each section. Paired comparisons of the cortical area were made between the titanium-alloy stems and the composite stems at each one-centimeter interval and were averaged for each time-period. This provided a macroscopic measurement of periprosthetic bone-remodeling. The depth of bone ingrowth was not calculated. Differences were analyzed for significance with use of analyses of variance and two-tailed paired Student t tests. The sections from the third, fifth, and seventh intervals were analyzed to determine intracortical porosity, a histological parameter of resorptive bone-remodeling. The mean intracortical porosity at each interval was compared between the two stem designs with use of two-tailed Student t tests. A Bonferroni adjustment was made to compare the overall values. For all statistical analyses, the level of significance was p < 0.05.
The acetabular cup was cut coronally to produce three one-millimeter-thick sections for contact radiography and scanning electron microscopy. The implants were categorized as those with a stable interface, denoted by multiple sites of bone ingrowth, or those with uncertain functional stability, denoted by continuous fibrous-tissue encapsulation.
Clinical Function and Fixation of the Acetabular Cup
After the operation, the dogs were allowed unrestricted activity in their cages. Limited walking usually began within two or three days, and full activity was typical within two to three weeks. The animals were allowed one hour of daily unrestricted activity, including running and jumping, in a large run.
There were three early postoperative dislocations. All necessitated open reduction, which was successful without additional complications.
Careful attention was paid to the function of the limbs and signs of the animal favoring one limb or limping. Six animals had a normal gait pattern throughout the study period. Two dogs—one that was evaluated at six months (Dog 4) and one that was evaluated at eighteen months (Dog 8)—limped on the side with the composite stem (Table III). The animal in the six-month group had a slight limp, which was noticeable during the middle two months of the study period. The animal in the eighteen-month group had periodic episodes of more obvious limping between nine and eighteen months after the operation. In both animals, the acetabular cup on the side with the composite stem was surrounded by a continuous radiolucent line and was encapsulated by fibrous tissue, a finding consistent with the gait abnormality. The other fourteen acetabular implants appeared stable radiographically (Table III) and were found to be fixed by bone ingrowth on histological analysis.
The two animals that had fibrous fixation of the acetabular cup were included in the analysis of bone ingrowth and radiopaque lines. This was because a limp and the resultant unloading of the composite femoral stem would be expected to promote stability of the implant and therefore would have no adverse effect on the potential for bone ingrowth compared with that on the contralateral side. Calculations also were made with exclusion of these two dogs to determine if there was a change in the relative magnitudes of the differences and the level of significance. Inclusion of these data did not alter the results for bone ingrowth. However, the same two animals were excluded from the analyses of cortical resorption and intracortical porosity because favoring one limb could induce bone resorption due to disuse atrophy, thus rendering comparisons of bone-remodeling on the basis of differences in the stiffness of the stem inequitable.
Radiographic and Histological Findings
Bone Ingrowth
Radiographically, all implants had characteristic signs of bone ingrowth in various regions along their length. There were no apparent differences in the amount or distribution of bone ingrowth between the six and eighteen-month evaluations. Densification of bone immediately adjacent to the porous fiber metal, suggestive of bone ingrowth and load transfer, was observed locally at both time-periods (Figs. 3 and 4). Marked new endosteal bone formed about the tip of both the titanium-alloy and the cobalt-chromium-composite stems in three animals (Figs. 4 and 5-A).
Histological analysis of the cross sections from all femora confirmed the presence of bone ingrowth at some of the serial one-centimeter intervals (Figs. 5-B, 6, 7, 8-A through 8-B). The most proximal section, where there was incomplete filling of the femur or where there had been inaccurate sizing during operative preparation of the canal, often had regions of fibrous-tissue encapsulation. At six months, there was 9.5, 8.9, and 17.2 percent more bone ingrowth at the third, fourth, and fifth intervals (zones 2 and 6 of Gruen et al.11) of the titanium-alloy stems (Table IV). When these three intervals were considered as a group, paired analysis revealed a significant difference between the two stem designs (p = 0.03). No significant differences were detected, with the numbers available, between the two types of stems with regard to the paired six-month bone-ingrowth data for the sixth and seventh intervals. Also, overall analysis of all five intervals revealed no significant difference in bone ingrowth between the two types of stems at six months (mean [and standard deviation], 19.0 ± 5.72 percent for the titanium-alloy stems and 14.1 ± 10.13 percent for the composite stems; p = 0.21).
At eighteen months, there were substantially greater differences in bone ingrowth between the two designs (Table IV). Paired analysis revealed significantly more ingrowth at the third, fourth, and fifth intervals (zones 2 and 6) of the titanium-alloy stems (p = 0.001); the values were 17.3, 22.4, and 25.3 percent, respectively. In addition, overall analysis of the third through seventh intervals (zones 2, 3, 5, and 6) revealed significantly more bone growth in the titanium-alloy stems (p = 0.004). The mean value for bone growth in the titanium-alloy stems was 28.1 ± 5.31 percent compared with 9.7 ± 5.38 percent for the composite stems (p = 0.003). Calculation of the relative magnitudes and the significance with exclusion of the data for the dogs that had a limp did not change the significance of the difference with respect to bone ingrowth between the composite stems and the titanium-alloy stems.
Despite signs of gross stability of the implant and bone ingrowth, the plain radiographs of some femora revealed isolated regions in which a distinct, thin, generally noncontinuous radiopaque line was present. Analysis of the titanium-alloy stems revealed a radiopaque line in one zone of Gruen et al.11 in two femora and in two zones in two femora. Analysis of the composite stems demonstrated a radiopaque line in six of the eight femora, four of which had a line in at least two zones.
Examination of the cross sections for the presence of radiopaque lines more clearly revealed different reactions to the two stem designs (Table III). Five of the eight femora containing a titanium-alloy stem had a partial radiopaque line in at least one cross section. Six (21 percent) of the twenty-eight sections had a radiopaque line at six months compared with four (14 percent) at eighteen months. In none of these sections was the radiopaque line continuous around the entire circumference of the implant. In contrast, all eight femora containing a composite stem had a radiopaque line in at least one cross section (Figs. 9, 10-A, 10-B through 10-C). Thirteen (46 percent) of the twenty-eight sections had a radiopaque line at six months; with the numbers available, this rate was not found to be significantly higher than that seen in association with the titanium-alloy stems. Fifteen sections (54 percent) had a radiopaque line at eighteen months; this rate was significantly higher than that associated with the titanium-alloy stems (p = 0.02). In six sections, the radiopaque line was continuous around the implant. There tended to be more radiopaque lines on the side containing the composite stem in the dogs that had less canal fill. All dogs that did not have filling of the canal had three or four more sections with a radiopaque line on the side with the composite stem compared with the side with the titanium-alloy stem (Fig. 9).
Bone-Remodeling
The immediate postoperative anteroposterior radiographs indicated that, in general, the diaphyseal portion of the femoral canal was well filled by the implant (Table III). In ten of the sixteen femora, the stem either completely filled the canal or was within one millimeter of the endosteal cortex (Figs. 3 and 5-A). In the remaining six femora, the stem was undersized by 1.5 to two millimeters (Fig. 4).
On the basis of plain radiographs, the patterns of bone-remodeling were generally similar at six and eighteen months. Within three to six months after the operation, there was clear radiographic evidence of a localized loss of cortical density or thinning in eight femora. This was evident in the proximal medial cortex (zone 7) and, to a lesser degree, in the proximal anterior cortex. Bone loss in zone 7 was minor or was similar in the two femora in two animals at six months and in two animals at eighteen months. In one animal that was evaluated at six months bone loss was slightly more marked in association with the composite stem, and in one animal that was evaluated at eighteen months it was much more evident in association with the titanium-alloy stem (Figs. 5-A and 5-B).
The radiographs of the cross sections revealed some variability in the bone-remodeling response between animals but demonstrated no general differences between time-periods. Six of the twelve paired femora had only minor resorptive changes along the entire length of the stem (Fig. 6). The other six femora had more noticeable resorption (Figs. 5-B and 7). Proximal anterior and medial resorption (in zone 7) was typically more marked than the resorption in the other zones. The thin sections clearly revealed marked localized cortical thinning, leading to exposure of the anteromedial edge of the titanium-alloy implant in two animals and the anteromedial edge of the composite implant in another animal (Figs. 5-B and 7). Other than these three instances of local resorption, there was no full-thickness cortical resorption in any zone of any femur. The two femora in which the composite stem was paired with a fibrous-encapsulated acetabular cup showed a generalized loss of cortical thickness around the stem in zones 2, 3, 5, and 6. In the dog that was evaluated at eighteen months, cortical thinning was also present distal to the level of the tip of the stem, a finding that is consistent with disuse atrophy from unequal limb-loading.
The intracortical porosity of the cortical bone surrounding the implants was highest proximally (range, 6.6 to 7.1 percent) and lowest distally (range, 1.6 to 1.9 percent); however, with the numbers available, the difference between the femora containing a titanium-alloy stem and those containing a composite stem was not found to be significant at any of the three intervals that were analyzed (Fig. 11).
At six months, there was significantly less cortical area (range, 8 to 22 percent) at the fourth through eighth intervals in the femora containing a composite stem (p = 0.03) but more cortical area at the second and third intervals. The overall difference for all seven intervals was not found to be significant, with the numbers available (p = 0.19). At eighteen months, there was slightly more cortical area (range, 11 to 14 percent) at the third through seventh intervals in the femora containing a composite stem; however, the difference for these five intervals as a group was not found to be significant, with the numbers available (p = 0.07). The overall difference for all seven intervals was also not found to be significant (p = 0.37).
Two types of fully porous-coated femoral stems that had identical dimensions but different structural stiffnesses were compared in the current study. The titanium-alloy stem had a stiffness comparable with that of the canine femur, whereas the cobalt-chromium-polyaryletherketone-composite stem was threefold less stiff than the canine femur proximally and four to fivefold less stiff distally. Use of the composite stem resulted in no clear decrease in stress-shielding compared with that produced by the titanium-alloy stem. An increase in the flexibility of the stem beyond that provided by the titanium-alloy stem resulted in a negligible alteration in femoral bone-remodeling and a decrease in the extent of bone ingrowth. Finite-element analysis of stress-shielding relationships between femoral prostheses and surrounding bone has indicated that the extent of periprosthetic stress-shielding, and hence the expected level of bone resorption, can be increasingly attenuated with increasing flexibility of the stem12,13,31. However, in the same studies it was predicted that the stresses between the implant and the bone that are due to interfacial shear motion also will increase with increasing flexibility of the stem. Excessive motion at the interface may jeopardize the stability of the implant. Therefore, a balance must be struck between the competing objectives of stress-shielding and interface stability.
The stiffness relationship between the implant and the femur can be adjusted by changing either the material or the geometry. Titanium-alloy implants, for example, have approximately one-half the stiffness of cobalt-chromium-alloy implants because of a twofold difference in the elastic modulus. Both radiographic and quantitative evaluations of bone-mineral density with use of dual-energy x-ray absorptiometry have revealed that titanium-alloy stems fixed without cement are associated with less periprosthetic bone resorption than are cobalt-chromium stems fixed without cement17,18,24. Inevitably, even with use of titanium alloy, larger stems are quite stiff relative to the femur and cause more resorption. Geometric features such as flutes, grooves, or slots help to reduce stiffness by decreasing the cross-sectional area or moment of inertia. However, there are practical limitations to the extent that the geometry of all-metal femoral stems can be modified. Because of strength limitations of surgical alloys, particularly after the heat treatments required for bonding of porous coatings, stiffness-reduction characteristics have limited applicability. The metaphyseal portion of the stem cannot tolerate meaningful geometric modification (such as the creation of grooves) without fatigue strength being jeopardized. Similarly, flutes or slots, which can reduce the geometric contribution to stiffness by as much as 25 percent in larger stems, can be incorporated only into the distal few centimeters of the stem because of the need to preserve mechanical properties in the middle third.
For many years, attempts to reconcile the divergent goals of flexibility of the stem and strength have focused on composite-materials technology22,25. It must be recognized, however, that increased flexibility of the implant may increase the shear stresses at the bone-implant interface during loading, and this in turn could create differential motion between the implant and the bone and cause loosening or mechanical abrasion of the implant. Polymer-coated prostheses originated with the isoelastic design of Mathys, which consists of a stem with a stainless-steel core and a thick casing of polyacetyl resin20. Although the results of use of this design have been favorable in some short-term follow-up studies1,20, rates of aseptic mechanical loosening as high as 32 percent (eleven of thirty-four) within a mean of forty-two months after the operation also have been reported15. Instability of the stem and the formation of radiopaque lines were attributed to excessive flexibility of the implant, even when there had been an initial press-fit within the intramedullary canal15. Clinical trials with use of a similar design, a titanium stem coated with a relatively thin layer of porous polysulfone (a high-strength thermoplastic material with well documented bone-ingrowth characteristics), generally were not successful21, and high rates of revision, evidence of mechanical failure of the polymer coating, and associated osteolysis led to the abandonment of this design21.
Carbon-carbon and carbon-polymer-composite materials have been studied extensively as candidates for a new generation of implants14,19,22,25. Their potential advantage is their combination of high fatigue strength and elastic moduli that are closer to those of bone than to those of metals. These materials can be fabricated in nonhomogeneous and nonisotropic combinations so that structural stiffness can be varied within different regions of the implant and more closely matched to that of bone. However, recent clinical trials with use of a carbon-fiber-polysulfone-composite femoral component revealed problems with stability21. Roentgen stereophotogrammetric analysis of the first eleven implants indicated progressive subsidence and rotation of all of the implants for as long as three years after the operation21.
Overall, the tissue response to the composite stems in the present study resulted in less bone ingrowth and more radiopaque lines than did the tissue response to the titanium-alloy stems. Only at the more distal intervals, where the stem was closely apposed to (and thus supported by) endosteal bone, was the bone-ingrowth response similar between the two stem designs. It has been well documented that excessive motion at the bone-implant interface of porous-coated femoral stems results in fibrous-tissue ingrowth rather than osseointegration16,23. Radiographically, this is manifested by a thin radiopaque line that is adjacent to but separate from the ingrowth surface. It is interesting to note that none of the composite stems were completely encased in fibrous tissue or surrounded by a radiopaque line. Instead, the radiopaque lines had a tendency to form in diaphyseal regions, where the differences in stiffness between the stem and the femur were maximum. Transverse sections showing partial bone ingrowth and incomplete radiopaque lines suggested that deformation of the composite stem can vary enough locally to result in a differential tissue response.
The development of radiopaque lines was more marked in femora that had incomplete filling of the canal. This suggests that stems with low stiffness characteristics are more susceptible to excessive motion and failure of bone ingrowth in association with suboptimum fit and fill of the canal; that is, they rely more on structural support from the femur to stabilize the bone-implant interface. Although the goal of all procedures in which the implant is inserted without cement is to fill the canal and maximize initial stability, this is not always possible, despite careful preoperative planning and operative technique. Thus, stems with a distal stiffness that is four to fivefold lower than that of the femur appear to be at higher risk for fibrous as opposed to osseous integration. It is possible that suboptimum fixation of a low-stiffness stem, resulting from extensive fibrous ingrowth in the diaphysis, could cause favoring of one limb and bone resorption due to disuse atrophy.
Another important finding was that the adverse effect of the composite stem on the bone-implant interface was not associated with a meaningful gain in periprosthetic bone retention. Only at six months was there a significant difference (p = 0.03) between the stem designs with regard to the cortical area over a portion of the bone-implant construct (at the fourth through eighth intervals), and there was more bone in the femora with the titanium-alloy stems. Other differences at six and eighteen months were relatively small and were not found to be significant, with the numbers available. Intracortical porosity was similar between the two types of stems. This is an important finding because it indicates that there is a limit to the beneficial effect of increased flexibility of the stem on resorptive bone-remodeling. It would be counterproductive to design implants in order to achieve minor gains in bone retention at the risk of interface stability.
In a previous study, implantation of fully porous-coated cobalt-chromium stems that were three to fourfold stiffer than the canine femur resulted in marked periprosthetic cortical thinning and extensive regions of full-thickness diaphyseal cortical resorption2,4. In the present study, with the exception of localized proximal medial and anterior cortical thinning in some femora, the amount of bone resorption surrounding the fiber-metal-coated titanium-alloy stems was insubstantial. None of the femora had full-thickness cortical resorption. These findings are comparable with earlier findings in a canine study in which porous-beaded hollow titanium stems were used4. The titanium-alloy implants in the present study and the hollow implants in the earlier study had stiffness characteristics that were generally the same as or slightly lower than those of the canine femur. The results of the current study indicate that the additional flexibility of the composite stems was not particularly helpful in decreasing resorptive bone-remodeling.
The findings with regard to bone-remodeling in the present study are not in complete agreement with those of a previous comparison of the same two stem designs in a unilateral hemiarthroplasty model in dogs; in that study, there was 30 to 50 percent less cortical resorption in the proximal regions around the composite stem30. It is difficult to know if this difference was due to differences between dogs that could result from use of a unilateral as opposed to a bilateral model. No details of the extent of filling of the canal by the stems among the different animals were provided for the unilateral model. In addition, there were no comparisons of the extent of bone ingrowth between the two stem designs and no details about the formation of radiopaque lines; both of these factors could substantially alter patterns of periprosthetic bone-remodeling.
Because of the strength limitations of metal alloys, only composite technology provides the latitude necessary to overcome the mismatches in stiffness between the stem and the femur that occur in association with larger implants. Unlike metal stems, which have substantially greater stiffness proximally compared with distally, composite technology enables both distal and proximal stiffness to be maintained at low levels relative to the femur. It is this unique property that has led to the use of composite technology in the clinical setting, in the hope that the benefits of hip arthroplasty without cement can be maximized while the risk of stress-shielding is minimized. However, this study indicates that there is a limit to which flexibility of the stem can be increased if the optimum balance between bone-remodeling and bone ingrowth is to be achieved.
NOTE: The authors thank K. Smith and U. Kanaan, B.Eng., for their invaluable technical assistance.