In a recent study of total hip arthroplasty performed with cementing of both the femoral and the acetabular component, 9 per cent (nine) of 102 hips followed for a minimum of fourteen years had aseptic loosening of the femoral component18. However, the prevalence of aseptic loosening of the cemented all-polyethylene acetabular component was 49 per cent (thirty-eight of seventy-seven). In order to reduce the rate of aseptic loosening that was associated with the cemented acetabular components and to retain the longevity of the cemented femoral component, the so-called hybrid total hip replacement, which combines an acetabular component inserted without cement and a femoral component inserted with cement, was designed12.
The hybrid total hip prosthesis that was evaluated in the present study is no longer being manufactured. The acetabular component (Acetabular Reconstruction Component; Howmedica, Rutherford, New Jersey) was fixed with screws through peripheral flanges, so a wide exposure was mandatory. This acetabular component was subsequently redesigned. The newer component has a similar hemispherical ingrowth surface, but screws are inserted through the dome of the shell, obviating the need for the peripheral flanges and the wide exposure. However, the Acetabular Reconstruction Component, which has many of the basic features of the subsequent designs, was used in the first series of hybrid arthroplasties of which we are aware, and that series was followed longer than any other of which we are aware.
Studies of hybrid total hip arthroplasties with other components, in series ranging from 125 to 155 hips followed for a minimum of five years, have revealed rates of 1 to 6 per cent for aseptic loosening of the femoral component, 1 to 3 per cent for revision due to aseptic loosening of the femoral component, 0 to 2 per cent for aseptic loosening of the acetabular component, and 0 to 1 per cent for revision due to aseptic loosening of the acetabular component2,4,9,16,21,23.
In the present study, we evaluated the results of fifty-two primary total hip arthroplasties (forty-seven patients) in which the femoral component was inserted with cement and the acetabular component was inserted without cement. Forty-seven hips (forty-two patients) were followed for a minimum of ten years.
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. In addition, benefits have been or will be directed to a research fund, foundation, educational institution, or other non-profit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was the William H. Harris Foundation, Boston, Massachusetts.
†Tennessee Orthopaedic Associates, 301 21st Avenue North, Nashville, Tennessee 37203.
‡Orthopaedic Biomechanics Laboratory, GrJ 1126, Massachusetts General Hospital, 55 Fruit Street, Boston, Massachusetts 02114. E-mail address: harrisw@helix.mgh.harvard.edu.
Fifty-two consecutive primary total hip arthroplasties were performed in forty-seven unselected patients, between November 1982 and February 1984, by the senior one of us (W. H. H.) at one institution. All primary total hip arthroplasties that were done in this period were performed with the same type of reconstruction. This included a Harris Design-2 femoral component (Howmedica) with a bead-blasted, cobalt-chromium, monoblock stem and a medial collar as well as a cobalt-chromium Acetabular Reconstruction Component with a sintered-bead bone-ingrowth surface, a non-modular shallow-chamfer polyethylene liner, and three peripheral flanges through which three or four screws provided fixation (Figs. 1-A, 1-B, and 1-C). Fourteen of the femoral head components were twenty-two millimeters in diameter (used when the outside diameter of the acetabular component was forty-eight millimeters or less), and the remaining thirty-eight femoral head components were twenty-six millimeters in diameter. The average outside diameter of the acetabular component was fifty millimeters (range, forty to fifty-eight millimeters).
The average age at the time of the index operation was fifty-seven years (range, twenty-nine to seventy-nine years). There were twenty-four men and twenty-three women. Twenty-seven right hips and twenty-five left hips were reconstructed. The average height of the patients was 170 centimeters (range, 143 to 186 centimeters), and the average weight was seventy-three kilograms (range, forty-six to 107 kilograms). Five patients had the hip arthroplasty bilaterally during the study period, with an average of twenty-one days (range, twelve to thirty days) between the two procedures. In addition, nine patients had a contralateral total hip prosthesis already in place at the time of the index operation, and seven patients had a contralateral hip arthroplasty performed after the study period. Eleven hips had been operated on previously, with an average of 1.5 (one, two, or three) operations.
The diagnoses leading to the total hip arthroplasty included developmental so-called pistol-grip deformity (fifteen hips), congenital dysplasia (thirteen hips)14, avascular necrosis (eight hips), rheumatoid arthritis (seven hips), osteoarthrosis that was so advanced the primary etiology could not be determined (three hips), slipped capital femoral epiphysis (two hips), Perthes disease (one hip), fracture of the femoral neck with avascular necrosis (one hip), poliomyelitis (one hip), and Paget disease (one hip). Of the thirteen congenitally dysplastic hips, twelve were classified as type I (less than 50 per cent subluxation) and one was classified as type III (75 to 100 per cent subluxation), according to the system of Crowe et al. With use of Charnley's system, fifteen patients were classified as having category-A function (involvement of one joint); twenty-six, category-B (involvement of multiple joints); and six, category-C (systemic disease).
The operative approach was posterolateral transtrochanteric in forty-nine hips and posterolateral without a trochanteric osteotomy in three. Non-structural autogenous cancellous graft was used to fill contained defects in ten acetabula. Cancellous graft from the iliac crest was used to fill screw-holes in one femur.
The femoral components were inserted with a so-called second-generation cementing technique1,18, which included the use of a medullary plug, a cement gun to introduce the cement in a retrograde fashion, and pressurization of the cement. The acetabula were prepared with hemispherical reamers sized to the outer diameter of the acetabular component to be used. The acetabular components were secured to the pelves with three or four screws through the three peripheral flanges.
The patients were followed prospectively with use of serial radiographs, questionnaires, and interviews. Anteroposterior and oblique radiographs of the pelvis and direct lateral and frog-leg lateral radiographs of the hip were made at a minimum of ten years for all forty-two surviving patients who had not had a revision. Questionnaires, which included general questions about function and satisfaction, were filled out by all but one of these patients. Twenty-one of the forty-two patients visited our office for an interview and an examination at least ten years after the operation. Nineteen patients were unable to return for an examination, and they were interviewed by telephone, after their questionnaire had been received, by the junior one of us (S. E. S.) at least ten years postoperatively. The technique of telephone interview after receipt of a completed questionnaire has been validated17. Two patients were examined in another surgeon's office, after which they were interviewed by telephone; one of these two patients did not return the questionnaire. The follow-up examination was done by an orthopaedic surgeon who was skilled in total hip arthroplasty but not the surgeon who had performed the operation. The Harris hip scores for the patients who did not return for an examination were calculated with use of the answers regarding the ability to put on shoes and socks, to replace the data regarding range of motion, as previously described19.
Information about the patients who had died before ten years had elapsed was derived from the radiographs and notes from the last office visit.
All of the radiographs were assessed by an orthopaedic surgeon (not the operating surgeon) who was participating in an arthroplasty fellowship. The definitions of coverage of the acetabular component, abduction angle, and version have been previously described18. The maximum width of gaps between the acetabular component and the host bone were recorded in millimeters and assigned a zone according to the system described by DeLee and Charnley. Any beads that had separated from the acetabular component were counted, and the zone8 in which they were found was recorded. Because the center of the femoral head was obscured by the dense shadow of the chromium-cobalt acetabular shell, measurements of even gross polyethylene wear were not believed to be reliable.
Any subsidence or other change in the position of the femoral component was recorded. The contact of the collar of the femoral component with the calcar, coverage by the collar of the cortical bone of the medial portion of the femoral neck at the site of the osteotomy, any change in the quality of the bone under the collar, and radiolucent lines between the femoral component and the cement mantle were noted according to the zones described by Gruen et al.
The technique for cementing of the femoral component was classified according to the lowest grade recorded from any of the immediate postoperative or follow-up radiographs, as previously described1,18. Grade A represented complete filling of the medullary canal of the diaphysis, whereas grade B meant that a distinction could be made between the cement mantle and the cortical bone in some areas of the diaphyseal region. Grade C-1 was given when a radiolucent line involved more than 50 per cent of the cement-bone interface or when there were voids in the cement, whereas grade C-2 was assigned when the cement mantle was less than one millimeter thick in any area. Grade D represented gross deficiencies in the cementing technique.
Osteolysis was defined as any perceptible area of endosteal, intracortical, or cancellous destruction of bone. These areas usually were not linear and commonly were scalloped. If they were linear, they had to be more than two millimeters in width and to be or to have been progressive to be considered osteolysis18. Femoral cement-bone radiolucency that was isolated to the proximal two centimeters of Gruen zones 1, 7, 8, or 14, was parallel to the cement mantle, and was less than two millimeters thick was not considered osteolysis18.
Loosening of the acetabular component was defined as a change in the abduction angle of more than 4 degrees or as vertical or horizontal migration of more than four millimeters20. Loosening of the femoral component was defined as radiographic evidence of fracture or bending of the femoral component, fracture of the cement mantle, subsidence of the femoral component in the cement mantle or of the femoral component-cement mantle composite in the medullary canal, or a radiolucent line of any width between the femoral component and the cement mantle (so-called debonding).
None of the fifty-two hips (forty-seven patients) was lost to follow-up, and radiographs made at least ten years postoperatively or before death or revision were available for all patients. Only one (2 per cent) of the fifty-two hips was revised during the follow-up period. This patient had two dislocations nine years after the index arthroplasty, and the hip was revised at a different hospital 9.7 years after the index arthroplasty to correct the tendency toward dislocation. Both components were well fixed. No femoral or acetabular component was revised because of aseptic loosening or periprosthetic osteolysis, and no revisions were pending at the time of writing.
No femoral component was radiographically loose at the time of the most recent follow-up or at the time of death. One (2 per cent) of the fifty-two acetabular components was aseptically loose at the time of the most recent follow-up. This component had migrated superiorly six millimeters, but the patient had good function, with a Harris hip score of 94 (of 100) points and a pain score of 40 (of 44) points at 12.4 years.
Pelvic osteolysis that involved an eight by twelve-millimeter area of zone I developed in one (2 per cent) of the fifty-two hips. Femoral osteolysis developed in eight hips (15 per cent). There were six lesions in zone 7; two lesions in each of zones 1, 2, 3, 4, and 8; and one lesion in each of zones 12 and 14. Excluding four hips in which the osteolysis developed in the proximal half of zones 1, 7, 8, or 14 but was three to four millimeters thick, linear, parallel to the cement, and less than two centimeters long, the prevalence of femoral osteolysis was 8 per cent (four of fifty-two hips). Distal osteolysis (in zones other than 1, 7, 8, and 14) was seen in three hips (6 per cent).
Four patients (four hips) died, for reasons unrelated to the arthroplasty, at an average of 8.2 years (range, five to 12.5 years) after the index operation. At the last examination before death, three of the four patients had good function of the hip, with an average Harris hip score of 86 points (71, 92, and 95 points). The only one of these three patients who had a score of less than 80 points had severe rheumatoid arthritis and diabetic neuropathy. The pain score was 30 points. The fourth patient who died had had recurrent dislocations of the hip seven years after the index arthroplasty but was too ill for an operation. She died of a cerebrovascular accident.
The duration of clinical follow-up for the forty-two patients (forty-seven hips) who were alive and had not had a revision after ten years averaged 12.3 years (range, 10.8 to 13.3 years). The average duration of radiographic follow-up was 12.1 years (range, 10.0 to 13.0 years). The preoperative Harris hip score for these forty-seven hips averaged 48 points (range, 26 to 63 points). The score for pain averaged 14 points (range, 10 to 30 points); the score for limp, 6 points (range, 0 to 11 points); the score for support needed for walking, 7 points (range, 0 to 11 points); and the score for distance that the patient could walk, 6 points (range, 2 to 11 points). At the time of the most recent follow-up, the Harris hip score averaged 89 points (range, 67 to 100 points). The score for pain averaged 41 points (range, 20 to 44 points); that for limp, 10 points (range, 0 to 11 points); that for support needed for walking, 9 points (range, 0 to 11 points); and that for the distance that the patient could walk, 9 points (range, 0 to 11 points). Three hips in three patients were given a total score of less than 70 points, but all three had a pain score of between 40 and 44 points and function was generally limited by a comorbid medical or arthrotic condition. The result was rated as good or excellent for forty (85 per cent) of the hips, fair (score, 74 to 79 points) for five (11 per cent), and poor (score, 67 and 69 points) for two (4 per cent).
Of the forty-two patients, five (12 per cent) described themselves as sedentary; thirteen (31 per cent), as semi-sedentary; eighteen (43 per cent), as able to perform light manual labor; and six (14 per cent), as able to perform moderate manual labor. All patients responded that both pain and the need for medication had decreased after the operation. Forty-one patients (98 per cent) were completely satisfied with the hip replacement, and forty (95 per cent) had a higher level of function than they had had preoperatively.
The technique for cementing of the femoral component was graded as A in three hips, B in ten, C-1 in eighteen, and C-2 in sixteen. Of the C-2 defects in the cement mantle, eight were in zone 5; seven, in zone 2; two each, in zones 3, 8, and 9; and one each, in zones 6, 7, and 12. The stem was in a valgus position in twenty-five hips, in a neutral position in fourteen, and in a varus position in eight. Contact between the collar and the calcar had been achieved in forty-three hips (91 per cent), and the collar covered an average of 34 per cent (range, 0 to 100 per cent) of the cortical bone at the level of the osteotomy. Contact between the collar and the calcar was maintained fully or in part in thirty-nine (91 per cent) of the forty-three hips.
Evaluation of the immediate postoperative radiographs revealed separation of beads from nine (19 per cent) of the forty-seven acetabular components. The average number of separated beads was 1.6 (range, one to four). At the time of the most recent follow-up, five of the nine components had no additional separation of beads and twenty-one hips had separation of beads that had not been seen on the immediate postoperative radiograph. The average number of separated beads was five (range, one to twenty-eight) at the time of the most recent follow-up.
The acetabular cancellous graft healed in nine of the ten hips, and the one femoral cancellous graft healed as well. Heterotopic ossification developed in thirty-three of the forty-seven hips. According to the system of Brooker et al., the ossification was grade I in sixteen hips, grade II in ten, grade III in four, and grade IV in three.
In addition to the one revision already described, complications occurred in eighteen (35 per cent) of the fifty-two hips (sixteen of the forty-seven patients) in the entire series. Seven hips (13 per cent) dislocated. Routine venography revealed five instances (10 per cent) of deep venous thrombosis. Three hips (6 per cent) had marked heterotopic ossification that was treated with operative resection. Three (6 per cent) of the forty-nine osteotomies of the greater trochanter were followed by non-union. There were two temporary peroneal palsies (one ipsilateral and one contralateral), and two hips had trochanteric bursitis. There was one wound hematoma, which did not necessitate incision and drainage; one decubitus ulcer; one non-fatal pulmonary embolism; and one infection. The infection, which developed thirteen years after the index arthroplasty in a patient who had rheumatoid arthritis, involved several joints, including the hip with the prosthesis, and was secondary to an infection at the site of an arthrodesis of the occiput and the second cervical vertebra. The patient was managed with irrigation and débridement of the hip joint with retention of the components.
A reoperation other than a revision was performed in six hips (four patients). As already noted, three hips (two patients) were treated with resection of heterotopic bone (one also had a trochanteric non-union, and the greater trochanter was advanced) and one patient had irrigation and débridement for infection. In addition, trochanteric wire was removed from two hips (one patient) because of bursitis and pain.
Certain long-term follow-up studies of primary total hip arthroplasty with cement have demonstrated excellent longevity of the femoral component, exceeding that of the acetabular component18,22. Schulte et al. reported that, during a twenty-year follow-up interval after 322 arthroplasties in which a Charnley prosthesis (Thackray, Leeds, United Kingdom) had been inserted with cement, 2 per cent (eight) of the femoral components and 6 per cent (eighteen) of the acetabular components had been revised because of aseptic loosening. The over-all rates of aseptic loosening of the femoral and acetabular components were 6 per cent (twenty) and 13 per cent (forty-three), respectively, although it should be noted that 22 per cent (seventy) of the hips in that series did not have radiographs at twenty years (the minimum follow-up interval). Mulroy et al. reported the results a minimum of fourteen years (average, fifteen years) after 102 arthroplasties (ninety patients) in which the femoral stem had been inserted with a so-called second-generation cementing technique and the acetabular component had been inserted with cement as well. Two per cent (two) of the femoral components and 10 per cent (eight) of the eighty-one all-polyethylene acetabular components were revised because of aseptic loosening. The total rates of aseptic loosening were 9 per cent (nine) of the 102 femoral components and 49 per cent (thirty-eight) of the seventy-seven all-polyethylene acetabular components for which radiographs were available at the most recent follow-up examination. The evolution of the so-called hybrid total hip arthroplasty was in response to this disparity between the longevity of cemented femoral and acetabular components12.
Several studies of primary hybrid total hip arthroplasty have demonstrated low rates of complications after a minimum of five years of follow-up2,4,9,16,21,23. These investigations involved hemispherical ingrowth acetabular components, fixed either with no screws or with screws placed through the dome of the component, in a total of 782 hips. At an average of 7.3 years (range, five to eleven years) after the arthroplasty, one acetabular component (0.1 per cent) and fourteen femoral components (2 per cent) were revised because of aseptic loosening. Of 627 hips that were included in analyses of osteolysis of the pelvis, eleven (2 per cent) were found to have that complication. Of 497 hips that were included in analyses of femoral osteolysis, twenty (4 per cent) had that complication. Therefore, excellent intermediate-term results, with regard to both revision and osteolysis, have been found after the insertion of different designs of acetabular components without cement in hybrid total hip arthroplasties. The rate of revision of the cemented femoral components and the rate of femoral osteolysis have also been low.
The duration of follow-up of the current series (average, 12.3 years for the surviving patients who had not had a revision) was longer than that of any reported series of hybrid total hip replacement of which we are aware. The results of the hybrid total hip arthroplasties in our series were very good, as determined on the basis of the Harris hip score and the fact that there was only one revision (for late-onset recurrent dislocation, 9.7 years after the index arthroplasty). There were no revisions because of aseptic loosening, and no femoral component loosened. There was aseptic loosening of one acetabular component, but it caused no symptoms.
We have not provided data on so-called possible loosening13, as experience has shown us that the criteria for this finding are more often representative of adaptive remodeling than they are of formation of a membrane and actual loosening. Studies of specimens obtained at autopsy15 have shown that the most common abnormality causing the appearance of a radiolucent zone at the cement-bone interface after a total hip replacement is the adaptation to the implant and cement, rather than lysis and formation of a membrane. Implants associated with less than complete radiolucent lines are usually found to be firmly fixed when tested during an operation.
Pelvic osteolysis developed in only one hip (2 per cent), and the extent was limited. This low rate of pelvic osteolysis might have been related in part to the peripheral fixation of the acetabular component, the lack of screw-holes in the acetabular shell, the use of a twenty-six or twenty-two-millimeter-diameter femoral head, the non-modular design of the acetabular component, the thickness of the polyethylene, and the monoblock construction of the femoral component6. Polyethylene wear could not be assessed quantitatively due to the radiographic density of the metallic shell overlying the center of the femoral head.
Femoral osteolysis developed in 15 per cent (eight) of the fifty-two hips in the entire group. Excluding the hips in which the osteolysis was present in only the proximal half of the four proximal zones10, the prevalence was 8 per cent (four) of the fifty-two hips. No hip was revised or needed a revision because of femoral osteolysis.
Features of the prosthetic design may have led to the high rate of dislocation (13 per cent) in this study. The small (twenty-two or twenty-six-millimeter) femoral head, the shallow chamfer at the margin of the acetabular liner, and the extensive pelvic dissection required for insertion of the component all may have contributed to this high rate.
The strengths of the design of the acetabular component are evidenced by the clinical results. The major drawback was that the flanges required wide acetabular dissection and frequent trochanteric osteotomy with or without release of the psoas insertion. The component was later redesigned, so that fixation was achieved with screws through the dome and the shallow chamfer was deepened.
The average Harris hip score for the forty-seven non-revised hips in the forty-two surviving patients was 89 points (range, 67 to 100 points) at the time of the most recent follow-up. Three hips had a score of less than 70 points, but all three had no or only slight pain. Ninety-eight per cent (forty-one) of the patients were completely satisfied with the hip replacement, and 95 per cent (forty) had a higher level of function than they had had preoperatively, despite the fact that they were older. Thus, function and relief of pain appeared to be excellent after the duration of follow-up in the present study.
While the operative technique and the design of the acetabular component in the present study were different than those in shorter-term studies of hybrid total hip arthroplasty2,4,9,16,21,23, longer-term results in the present series (average duration of clinical follow-up, 12.3 years; range, 10.8 to 13.3 years) add valuable data regarding the longevity of hybrid total hip replacements. It appears that hybrid total hip arthroplasty provides very good function over an intermediate time-period.