Since the introduction of modular prosthetic components in 197119, an increasingly diverse variety of such components have been developed to address the complexities of total hip arthroplasty. Both the beneficial and the detrimental effects of modular hip-arthroplasty components have been reported. The predominant clinical benefit, intraoperative flexibility with regard to the choice of components, was emphasized in a study in which the length of the modular femoral head component was changed after insertion of the femoral stem with cement in nineteen (19 per cent) of 100 consecutive total hip arthroplasties20. An economic benefit of modular implant systems offering a wide variety of component sizes is that smaller inventories are necessary than would be required with monolithic systems. Modular systems also offer less intuitive benefits; they have been useful, for example, in the treatment of nerve palsy33, and they can be removed with a universal femoral component extractor4. However, there have also been reports of detrimental effects associated with modular systems, including in vivo dissociation of the components2,27,34; mismatch of the components2; poor component-fit tolerances27; and corrosion, fretting, and generation of wear debris from nonarticulating modular surfaces1,3,8,9,11,14,15,21,30,31,34,40. Despite these potential adverse effects, several authors have reported favorable intermediate-term outcomes with use of implant systems consisting of multiple modular components5,7.
The current study addresses an issue specific to modular femoral head components: the effect of a reinforced extension on the femoral head. In order to offer a component with a longer neck, for the purpose of enhancing the stability of the implant or equalizing limb length, manufacturers have added a flange extension to the modular femoral head so as to maintain adequate strength at the stem-head junction. In addition to the concerns regarding modular components that already were discussed, the addition of a flange extension (or so-called skirt) raises a concern regarding the effect of decreasing the ratio between the diameters of the femoral head and neck. Several authors have shown that this ratio is inversely proportional to impingement of the prosthetic neck on the acetabular rim and directly proportional to the prosthetic range of motion1,6,32. By decreasing the head-neck-diameter ratio, a flange extension increases the likelihood of prosthetic impingement22, which can have dramatic detrimental effects. In vitro studies have shown that ceramic femoral heads are very vulnerable to prosthetic impingement10,38, and case reports have demonstrated a relationship between failure of the component and prosthetic impingement29,37. However, we know of no previous report that has addressed the effect of a modular femoral head with a flange-reinforced neck extension on the outcomes in total hip arthroplasties.
We used a database containing prospectively collected information to analyze the effect of a modular femoral head with a flange-reinforced neck extension on the rate of dislocation of the hip prosthesis, the clinical and radiographic outcomes, and, particularly, the rate of polyethylene wear after total hip arthroplasty.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Division of Orthopaedic Surgery, Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, California 92037.
The rate of polyethylene wear was evaluated in ninety-one patients who had had a total of 100 total hip arthroplasties (nine patients had had a staged bilateral procedure) in order to determine the effect of a modular femoral head with or without an extended neck after an intermediate duration of follow-up (mean, 6.1 years; range, four to eight years). The study cohort consisted of the first ninety hips that had had a total hip arthroplasty without cement and the first ten that had had a so-called hybrid procedure (insertion of the femoral component with cement and the acetabular component without cement) performed at our institution by the two senior ones of us (C. W. C., Jr., and R. H. W.) with use of the same design of acetabular component.
All of the acetabular components (Harris-Galante I; Zimmer, Warsaw, Indiana), which were inserted without cement, consisted of a titanium-alloy outer shell with a titanium-alloy porous fiber-mesh pad for bone ingrowth and a modular nonelevated polyethylene liner. All of the femoral stems and the modular femoral heads were of the Harris-Galante or the Harris Precoat design (Zimmer). The Harris-Galante femoral stems were inserted without cement and were made of titanium alloy, with a titanium-alloy fiber-mesh pad for bone ingrowth at the proximal one-third. The Harris Precoat stems were made of cobalt-chromium alloy, with a polymethylmethacrylate surface preparation at the proximal one-third for enhancement of fixation with bone cement. All of the femoral heads were made of cobalt-chromium alloy and had a diameter of twenty-eight millimeters. These modular components were available in lengths of short, medium, and long, which added zero, seven, and fourteen millimeters, respectively, to the length of the trunion of the femoral stem. The long components included a reinforced neck extension (a flange) for placement over the trunion of the stem. The flange had the effect of increasing the diameter of the neck and decreasing the head-neck-diameter ratio.
The arthroplasty was performed because of osteoarthrosis in fifty-four hips, because of osteonecrosis in twenty-seven, because of inflammatory arthritis in eight, because of posttraumatic osteoarthrosis in five, because of congenital dysplasia in five, and because of hemochromatosis in one hip.
All arthroplasties were performed with use of a standardized posterolateral approach, with the trochanter left intact. The acetabular component was inserted with a line-to-line reaming technique and use of two to six supplemental titanium-alloy dome-fixation screws. The decision to use a long femoral head component—that is, one with a flange extension—rather than a medium or short component (without a flange extension) was made intraoperatively by the surgeon. This decision was based on measurement of the limb length and the use of radiographic templates for arthroplasty components preoperatively as well as on measurement of the limb length, determination of the stability of the prosthesis, and anteroposterior radiographs of the pelvis intraoperatively.
A rehabilitation protocol was used postoperatively for all patients.
The clinical evaluation included determination of the Harris hip score18 both preoperatively and postoperatively.
The radiographic evaluation included measurement of the abduction angle of the acetabular component with reference to the horizontal pelvic teardrop line. The presence of acetabular or femoral osteolysis was recorded. Osteolysis was defined as periprosthetic radiolucency demonstrating cavitation (scalloping) with a minimum longitudinal measurement of five millimeters17. So-called droplet lytic lesions in the calcar femorale were classified as femoral osteolysis.
Linear polyethylene wear was measured radiographically according to the technique of Livermore et al.25, with use of digital calipers. The direction of wear also was determined with use of the technique of Livermore et al., by passing a vertical line through the center of the femoral head and perpendicular to the horizontal pelvic teardrop line. This allowed the direction of wear to be classified as superolateral or superomedial (zone I or II, respectively, according to the system of DeLee and Charnley16). If the vector of wear was directly superior (that is, if it was on the vertical line centered on the acetabular shell), the direction of wear was classified as superolateral.
The extent of linear polyethylene wear was evaluated relative to whether or not the femoral head had a flange extension; the age, gender, and weight of the patient; the abduction angle of the acetabular component; the initial thickness of the polyethylene; and the direction of wear.
Data analysis was performed with use of a Sun computer system (Sun Microsystems, Mountain View, California) as well as a Prophet statistical program (release 3.2; Bolt, Beraneck and Newman, Cambridge, Massachusetts).
Eight of the ninety-one patients (100 hips) were followed for less than four years after the index arthroplasty. Four patients (four hips) died, three patients (three hips) were lost to follow-up, and one patient (one hip) had a revision arthroplasty three years postoperatively because of aseptic loosening of the acetabular component.
Eighty-three patients (ninety-two hips) were followed for more than four years after the arthroplasty (Table I). Sixty-two patients (sixty-six hips) had had radiographic evaluation that was adequate for the analysis of polyethylene wear, and twenty-one patients (twenty-six hips) had had inadequate radiographic evaluation; the demographic data and the clinical outcomes were comparable for these two groups.
The demographic data, the results of the radiographic analysis of the acetabular component, the clinical outcome, and the rate of polyethylene wear were analyzed according to the presence or absence of a flange extension (Table II). No significant differences were detected, with the numbers available, between the hips that did and did not have a femoral component with a flange extension with regard to the age, weight, or gender of the patients; the abduction angle of the acetabular component; the initial thickness of the polyethylene; the rate of dislocation or revision of the hip prosthesis; or the Harris hip score18. Two (4 percent) of the fifty-five femoral components that did not have a flange extension were revised because of aseptic loosening five years after the index arthroplasty. Both acetabular components were intact and were included in the five-year radiographic follow-up evaluation.
The mean rate (and standard deviation) of linear polyethylene wear in the sixty-six hips that had had adequate radiographic evaluation was 0.12 ± 0.08 millimeter per year (range, zero to 0.44 millimeter per year). The mean rate of wear per year was 0.17 millimeter in the eleven hips that had a long femoral head component (a flange extension), 0.11 millimeter in the fifty-two that had a medium-length component, and 0.15 millimeter in the three that had a short component. The rate of wear in the eleven hips that had a long femoral component was significantly greater than that in the fifty-five that had a medium-length or short component (p = 0.009). According to multivariate analysis, an increased rate of linear wear was more strongly associated with the presence of a flange extension (F = 2.86) than it was with a younger age (F = 1.72), a more vertical abduction angle of the cup (F = 0.49), a heavier weight (F = 0.14), male gender (F = 0.11), or a smaller initial thickness of the polyethylene (F = 0.02).
The direction of wear in the sixty-six hips was significantly more likely to be superolateral when the femoral component had a flange extension than it was when the femoral component did not have a flange extension (p = 0.045, chi-square test) (Table III). The presence of a flange extension was associated with superolateral wear of nine of eleven cups and with superomedial wear of the other two, whereas the absence of a flange extension was associated with superolateral wear of twenty-five (45 per cent) of fifty-five cups and with superomedial wear of thirty (55 per cent). The mean rate of wear in the thirty-four hips that had superolateral wear was significantly greater than that in the thirty-two hips that had superomedial wear (0.15 compared with 0.10 millimeter per year; p = 0.01).
The rate of wear also was significantly greater in the thirteen hips that had femoral periprosthetic osteolysis than it was in the fifty-three that did not (mean, 0.18 compared with 0.11 millimeter per year; p = 0.01). None of the sixty-six cups had acetabular periprosthetic osteolysis.
The radiographic evaluation in this study showed that long modular femoral head components with a flange extension resulted in a significantly greater rate of polyethylene wear than the components without the extension. In addition, the femoral heads with a flange extension were associated with a superolateral direction of wear (p = 0.045), a superior direction of wear was associated with greater wear (p = 0.01), and greater wear was associated with femoral osteolysis (p = 0.01).
The mean rate of wear that was measured in this study (0.12 millimeter per year at a mean of 6.1 years postoperatively) was similar to the rates that have been reported for this prosthetic system previously. Woolson and Murphy reported a mean rate of 0.14 millimeter per year at a mean of 5.7 years39; Latimer and Lachiewicz, 0.1 millimeter at seven years24; and Tompkins et al., 0.11 millimeter at 8.7 years35.
Of the 100 hips in the entire series, ninety had a cobalt-chromium-alloy femoral head with a titanium-alloy femoral stem. There are no conclusive data with regard to the effect on polyethylene wear of a so-called mixed-alloy prosthetic system compared with that of an all-cobalt-chromium modular system. Collier et al. reported increased corrosion but no increase in fretting in association with mixed-alloy modular systems11,13. However, other authors have predicted or reported no increase in corrosion23,26,28,42 or polyethylene wear36 with use of such systems.
In the current study, the use of a flange extension for the femoral head component had no adverse effect on the rate of dislocation or on the clinical outcome as determined on the basis of the rate of revision and the Harris hip score18. Power analysis for this small cohort of patients did not allow assignment of significance to these negative findings; however, this study is the first of which we are aware to demonstrate an association between a flange extension and an adverse effect (a significantly greater rate of polyethylene wear) in a clinical series.
Laboratory studies have raised concern regarding an increase in prosthetic impingement in association with a flange extension1,6,22. Krushell et al., in a study of cadavera, evaluated the same modular system for total hip arthroplasty that was used clinically in the current series22. They found that a change from a medium-length femoral head component (seven millimeters, without a flange) to a long component (fourteen millimeters, with a flange) decreased flexion from 137 to 117 degrees (with the hip in neutral abduction and rotation) and decreased internal rotation from 18 to 2 degrees (with the hip in 90 degrees of flexion and neutral abduction) before prosthetic impingement occurred.
Although confirmation may require a retrieval study of cadavera, several mechanisms have been postulated to explain the increased polyethylene wear seen in association with a flange extension. First, frequent prosthetic impingement, with increased contact between the femoral neck and the polyethylene rim of the acetabular component, has been implicated as a cause of increased wear of the rim, generation of polyethylene debris, and debris-related central acetabular wear29,37,38,40. Chandler et al., in an arthroplasty model in cadavera, found that an increase in the ratio between the diameters of the head and neck resulted in an increased range of motion and in decreased prosthetic impingement6. They noted that "prosthetic impingement leads to socket rim wear." Second, prosthetic impingement also has been implicated as a cause of repetitive levering of the femoral head from the acetabular articular surface, resulting in multiple microsubluxation events, a so-called point-contact wear phenomenon, loss of concentricity of acetabular wear, and accelerated wear1,8,40. Krushell et al. conjectured that "impingement and slight subluxation of heads with flanges during daily activities might not be unusual."22 Yamaguchi et al., using a three-dimensional wear analysis of 104 cups retrieved at revision from patients who had had a total hip arthroplasty, found significant relationships between polyethylene wear in multiple vectors and a smaller femoral head (p = 0.03), prosthetic impingement (p = 0.0006), the duration that the implant had been in situ (p = 0.007), and maximum linear wear (p = 0.002)41. Those authors concluded that "the relationship between multiple wear vectors and the size of the femoral head may be related to impingement." Third, levering of the femoral head from the acetabulum, resulting in multiple microsubluxation events, reportedly allows an avenue for the introduction of a third body into the central acetabular polyethylene surface40. Conceivably, all three proposed mechanisms could contribute to increased polyethylene wear.
The current study showed that a hip with a femoral head that has a flange extension is at risk for more rapid linear polyethylene wear. On the basis of our findings and those of previous studies, we hypothesized the following mechanism for more rapid wear: (1) a flange extension decreases the ratio between the diameters of the femoral head and neck and increases the potential for impingement of the femoral neck on the acetabular rim; (2) increased prosthetic impingement results in an increase in wear of the rim and in polyethylene debris; (3) repetitive impingement also results in repetitive levering-induced microsubluxation of the femoral head; (4) repetitive levering of the femoral head results in acentric so-called point-contact wear; (5) microsubluxation allows an avenue for introduction of a third body into the central acetabular surface; and (6) the combined phenomena of a greater amount of polyethylene debris, point-contact wear, and exposure to a third body result in a greater likelihood of accelerated linear polyethylene wear.
The presence of a flange extension was significantly more likely to result in superolateral wear as opposed to superomedial wear, which is more aligned with the vector of hip-reactive forces (p = 0.045). Livermore et al., in a radiographic evaluation after 385 total hip arthroplasties, found the mean vector of wear to be "cephalad and medial," or superomedial25. We believe that the association between a flange extension and superolateral wear is due to repetitive prosthetic impingement at the inferolateral part of the rim occurring with weight-bearing terminal extension and external rotation during the stance phase of gait, repetitive lift-off of the femoral head with external rotation, repetitive superolateral point-contact wear, and resultant acentric wear in a superolateral direction.
The present study also shows, as has been reported previously35, that a greater rate of polyethylene wear places the patient at greater risk for osteolysis around the femoral component. Although the flange extension was not found to have a detrimental clinical effect in this intermediate-term follow-up study, our findings of accelerated polyethylene wear and an associated increased prevalence of femoral osteolysis warrant concern that such adverse effects will be demonstrated after more long-term follow-up.
Despite our findings, the ability to select variable femoral neck lengths at the time of a hip arthroplasty, made possible by the modularity of the femoral head, affords a technical advantage that is worth maintaining. Modularity of the femoral head also provides an opportunity for intraoperative adjustment regarding the diameter of the head, equalization of limb length, tensioning of soft tissue, stability of the implant, and soft-tissue impingement. Our findings suggest that other available options for optimizing these factors should be explored before a femoral head with a flange extension is selected. We have begun to use a femoral stem with a variable neck (Morse-taper trunion) length. If a long femoral neck is desired after critical intraoperative evaluation of limb length and stability of the implant, a femoral stem with a trunion that is five millimeters longer than the standard trunion, rather than a femoral head with a flange extension, is selected.