All patients had revision with use of second-generation cementing techniques for the treatment of aseptic loosening of a femoral stem between February 1976 and June 1980. This was a consecutive series of unselected patients and included all patients who had had such an operation during that interval. Of the original group (forty-three hips in forty-one patients), eight patients (eight hips) had died; the remaining thirty-three patients (thirty-five hips) were evaluated at an average of 15.1 years (range, 14.2 to 17.5 years). Twenty patients were men and thirteen were women. The average age at the time of the primary arthroplasty was forty-five years, and at the time of the index revision it was fifty-one years. The primary total hip replacement was done because of osteoarthrosis in sixteen hips, congenital dislocation of the hip in nine, post-traumatic osteoarthrosis in five, avascular necrosis in two, slipped capital femoral epiphysis in two, and Legg-Perthes disease in one.
In the entire series of forty-one patients (forty-three hips), cultures of fluid that had been aspirated from the hip joint preoperatively and at the time of the revision showed no growth. All patients had received antibiotics prophylactically, but the initial dose had been delayed until after deep material had been obtained for culture at the time of the operation. A trochanteric osteotomy was performed in all of the hips. Trochanteric wire mesh was used to supplement the three-wire technique of trochanteric reattachment15,20,32 in twenty-eight hips because of concern that the wires would pull through the trochanter. Thirty-six operations were a first femoral revision, six were a second, and one was a third.
The second-generation cementing techniques that were employed in all of the index revisions included the use of a polymethylmethacrylate plug to occlude the intramedullary canal of the femur and a cement gun to fill the canal with a doughy mix of Simplex-P bone cement (Howmedica, Rutherford, New Jersey) in a retrograde fashion1,13. Pulsatile lavage was used in operations that were done after January 1978 (twenty-six hips), and a cement compactor was used to pressurize the cement in operations that were performed after November 1978 (twenty-one hips)26. Reduction of the porosity of the cement, centralizers for the femoral stem, and precoating of the femoral component with methylmethacrylate had not yet been developed5-8,29.
Several different femoral components were removed at the time of the index revision. They included fourteen Müller, ten Harris-1, four Aufranc-Turner, four Harris CAD (computer-assisted-design), three Tronzo, three Trapezoidal-28, and two Bechtol prostheses, as well as one CDH, one Harris Micromini, and one McKee-Farrar prosthesis. For the thirty-three surviving patients (thirty-five hips), the femoral components (all made by Howmedica) that were inserted at the index revision included three CDH, five CAD, twelve HD-2, and fifteen so-called calcar-replacement stems14. Calcar-replacement stems have an extended proximal segment for use in revisions in patients who have had major loss of the cortical bone of the medial portion of the femoral neck. The offsets of the calcar replacement prostheses ranged from thirty-five to forty-five millimeters.
All thirty-three surviving patients (thirty-five hips) were interviewed by one of us (W. F. M.), an independent orthopaedic surgeon who had neither done the operation nor participated in the care of the patients. No patient was lost to follow-up. Twenty-two patients (twenty-four hips) returned for a physical and radiographic examination and to complete a questionnaire. The remaining eleven patients (eleven hips) completed a questionnaire and were interviewed by telephone; nine of them submitted current radiographs.
Current radiographs were made for the twenty hips in which at least one index component was still in place. Judet radiographs, an anteroposterior radiograph of the hip and pelvis, a frog-leg lateral radiograph of the involved femur, and a true lateral radiograph of the hip were made. These were compared with the preoperative, the immediate postoperative, and all intermediate radiographs. Our radiographic criteria for loosening of the femoral component were published previously16, and we used the criteria of Hodgkinson et al.19 to evaluate loosening of the acetabular component.
The quality of the femoral bone stock was graded on the basis of its appearance on the radiographs that had been made before the revision10. Grade I indicated little or no loss of cortical bone; grade II, moderate loss of cortical bone; and grade III, major loss of bone stock.
The mantle of cement in the femur was graded on the basis of the immediate postoperative radiograph, according to a slight modification of the criteria of Barrack et al.2. A grade of A was given if cementing had resulted in complete filling of the intramedullary cavity of the diaphysis (a so-called white-out at the cement-bone interface); a grade of B, if there was slight radiolucency (indicating slight non-filling of the trabecular bone adjacent to the cortex) at the cement-bone interface; a grade of C1, if there was more extensive radiolucency (along 50 to 99 per cent of the cement-bone interface) or voids in the cement; and a grade of C2, if there was either a thin (less than one-millimeter) mantle of cement at any site or a defect in the mantle, with the metal in direct contact with cortical bone. A grade of D was given if the gross appearance was poor; this included radiolucency along 100 per cent of the cement-bone interface on any radiograph, no cement distal to the tip of the stem, and multiple defects or large voids in the mantle of cement2. The extent and thickness of any radiolucent lines at the acetabular cement-bone interface were recorded for each of the three zones of DeLee and Charnley9.
The rate of revision for aseptic loosening of the femoral component for the entire group of forty-three hips over the study period (average, 15.1 years) was 16 per cent (seven hips). In contrast, the rate of reoperation was twice as great (35 per cent; fifteen hips), primarily because of acetabular problems.
None of the eight patients (eight hips) who had died before the time of this review had had a revision. Only one patient had had a loose femoral component according to the last available radiograph, and this patient had been asymptomatic. The average age at the time of the index revision was distinctly different for this group of eight patients (sixty-eight years; range, sixty to eighty-two years) compared with the average age for the surviving patients (fifty-one years). The average duration of follow-up for the patients who had died was 11.0 years (range, 3.8 to 14.8 years). Only one of the eight patients had had pain in the hip that was rated as being more than mild at the time of the last follow-up.
At the time of follow-up of the thirty-three surviving patients, fifteen hips (43 per cent) had had a reoperation during which one or both components had been removed; seven femoral components (20 per cent) had been removed because of aseptic loosening. The over-all rate of aseptic loosening of the femoral component was 26 per cent (nine components: seven that had been revised and two that were loose radiographically).
Four of the hips that had repeat revision of the femoral component because of aseptic loosening had been reported on in the previous reviews10,30; repeat revision had been done fifty-nine, 104, 110, and 114 months after the index revision. Since the time of the review at 11.7 years10, three other aseptically loose femoral components have been revised again, 179, 183, and 188 months after the index revision. A fourth femoral component had been removed because of infection.
At the six-year review30, three femoral components were loose radiographically. By the time of the 11.7-year review10, two of the three had been revised again because of aseptic loosening. The third component and three others were loose radiographically. Of these four femoral components, three had been revised again by the time of the present review. (The fourth component was in a patient who died without a reoperation.)
Four hips had a reoperation during which the femoral component was also removed despite its being rigidly fixed. These reoperations included a resection arthroplasty (one hip) at seventy-three months because of irreparable loss of bone from the acetabulum, repeat revision because of recurrent dislocation of the hip (two hips), and revision of a loose acetabular component (one hip) at 151 months. In the acetabular revision, the well fixed femoral component was removed for exposure and then was cemented back into the existing mantle of cement. Two other femoral components were removed during resection arthroplasty because of infection, 111 and 133 months after the index revision.
Radiographs were made of twenty of the twenty-two hips in which the index femoral component was in place at the time of the review, and two femoral components were seen to be loose. In the entire series of forty-three hips, seven (16 per cent) had been revised for aseptic loosening of the femoral component during the study period and three (7 per cent) had a loose femoral component according to radiographic criteria; thus, a total of ten components (23 per cent) were loose.
In the entire series, four (11 per cent) of the thirty-six first femoral revisions were followed by another revision and two (6 per cent) were complicated by radiographic loosening of the femoral component. Of the seven hips for which the index operation was a repeat revision (a second revision for six hips and a third revision for one), three had another repeat revision.
The specific purpose of the present review was to evaluate the long-term results of femoral revision with use of second-generation cementing techniques; however, the outcome of the revised acetabular components was also assessed. The acetabular component was revised during thirty-six of the forty-three index revisions of the femoral component. Seven of the patients (seven hips) who had an acetabular revision died without needing a reoperation. Of the remaining twenty-nine hips, thirteen (45 per cent) had a reoperation. Eleven reoperations (38 per cent) were performed for aseptic loosening of the acetabular component and two (7 per cent), for reasons other than aseptic loosening. Repeat revision of the acetabular component was done in nine of the eleven hips, and resection arthroplasty was done in the other two hips, as already mentioned, because of irreparable loss of bone from the acetabulum (one hip) or infection (one hip). The acetabular component became loose in four of the seven hips in which it had not been removed at the index operation, but no reoperation had been done by the time of this review.
Eighteen of the twenty-nine hips that had an acetabular revision at the time of the index operation did not have repeat revision for aseptic loosening. The acetabular component was removed from two of these hips, despite its being well fixed; one was removed because of infection (as already noted) and the other, because of recurrent dislocation. Therefore, sixteen hips in which the acetabular component had not been revised again or removed were available for radiographic review of the acetabular component. Seven of the sixteen components were radiographically loose. Thus, 62 per cent (eighteen) of the twenty-nine acetabular components that had been revised at the index operation were loose.
The twenty hips in the surviving patients who had not had a reoperation had an average Harris hip score12 of 56 points (range, 42 to 73 points) preoperatively and 82 points (range, 61 to 100 points) at the most recent follow-up examination. Eleven hips (55 per cent) were rated as good or excellent; five (25 per cent), fair; and four (20 per cent), poor. The four patients who had a poor rating (less than 70 points) all were limited by poor function and mobility that were unrelated to the involved hip, and none of them had more than mild pain in that hip.
The radiographs that had been made before the index revision showed that the quality of femoral bone was grade I in eleven hips, grade II in seventeen, and grade III in seven in the group of thirty-three surviving patients (thirty-five hips). Thus, the quality of the femoral bone stock was grade II or III in twenty-four hips (69 per cent). According to radiographs that were made after the index revision, the mantle of cement was graded as A in five hips, B in nine, C1 in five, C2 in thirteen, and D in three. When the population was segregated into subsets according to both the grading of the mantle of cement and the classification of bone stock, the small sample size of each subset precluded statistical conclusions regarding improved durability of the femoral component in hips that had better quality of bone stock or grades for cementing.
There was some evidence of femoral lysis in nine (26 per cent) of the thirty-five hips. For the purpose of this study, the definition of lysis did not include a radiolucent line of two millimeters in width or less in the proximal one centimeter of zones 1 and 7 (according to Gruen et al.11), designated 1A and 7A, and the corresponding regions on the lateral radiograph, commonly designated as zones 8A and 14A. In the entire group of hips (those that did and those that did not have repeat revision), the rate of femoral lysis associated with a loose component was twice that associated with a rigidly fixed component. Specifically, with regard to the twenty-two femoral components that were not revised, five of the twenty fixed components and one of the two loose components were associated with femoral lysis. As for the thirteen femoral components that had been removed, four of the seven loose components and one of the six fixed components were associated with femoral lysis.
The aim of this study was to assess whether revision with use of second-generation cementing techniques, performed because of aseptic loosening of the femoral component, was associated with the same improved duration of fixation that has been reported after primary total hip replacement in older patients and in younger patients managed with good cementing techniques (classified2 as A, B, or C1)21,23,27,31,34,35.
Although results after durations of fifteen years and longer have been reported for primary total hip arthroplasties with use of first and second-generation cementing techniques21,23,31,34,35, we report the results of revision total hip arthroplasty with cement after an average of 15.1 years. The rate of revision for aseptic loosening of primary femoral components that had been inserted with cement has been low at fifteen to twenty years (2 per cent [four of 162 hips]25 to 3 per cent [three of ninety-eight]31); in the present study, the rate of repeat revision for first revisions was 11 per cent (four of thirty-six hips) and the over-all rate of repeat revision of the femoral component was 16 per cent (seven of forty-three). This highlights the difference in the longevity of femoral fixation between primary and revision total hip arthroplasty. The less satisfactory results after revision of femoral components with cement is predictable, given the compromised femoral bone stock in many patients4. This challenge has led many surgeons to abandon the use of cement in revision arthroplasty in the hope of gaining more lasting fixation22. However, in reports after short-term follow-up, the rates of repeat revision and loosening have often been higher after arthroplasty with ingrowth implants (no cement)3 than in our reports after longer follow-up of femoral revisions with use of second-generation cementing techniques10,30. Hedley et al.17 reported that 4 per cent (two) of fifty-one porous coated anatomic (Howmedica) femoral stems that had been used in revisions were revised again because of loosening after two years or less. At the time of the six-year follow-up for the current series, 2 per cent (one) of the forty-three femoral components had been revised again because of aseptic loosening30.
In several series, the age of the patient has been a determining factor in the durability of the femoral fixation after revision. Stromberg et al.33 reported a rate of repeat revision of the femoral component of 19 per cent (thirteen of sixty-seven hips) at 2.9 years in patients who were fifty-five years old or less and in whom first-generation cementing techniques had been used. Herberts et al.18 reported on a Swedish population of patients who were fifty-five years old or less at the time of revision with use of second-generation cementing techniques; after four years, fourteen (21 per cent) of sixty-six hips had had repeat revision because of aseptic loosening of the femoral component. For patients who were fifty-five to seventy years old at the time of the index revision, the rate of repeat revision was 16 per cent (thirty-three of 204 hips) at six years. In our thirty-three patients (thirty-five hips) who were alive at the time of the review, the femoral revision had been done at an average age of fifty-one years; this relatively young age added an extra challenge to the long-term success of second-generation cementing. However, of the fifteen hips (43 per cent) in patients who were fifty years old or less at the time of the index revision and were evaluated at least fourteen years postoperatively, two had a repeat revision. Of the twenty hips (57 per cent) in patients who were fifty-one years old or more at the time of the index revision, five had repeat revision by fourteen years. With the numbers available, this difference was not significant (p < 0.43).
Our results can be compared with those in the study by Pellicci et al.28, one of the longest follow-up studies (average, 8.1 years) that has been published after femoral revision with use of first-generation cementing techniques (finger-packing of the cement without use of a medullary plug or cement gun; no centrifugation, pressurization, or reduction of porosity; and, often, use of a cast stem, many of which have sharp corners and a narrow medial border). In their study, the rate of repeat revision for aseptic loosening of the femoral component was 18 per cent (eighteen of ninety-nine hips in ninety-seven patients); in our study, this rate was 16 per cent (seven of forty-three hips) after nearly twice the duration of follow-up and the rate of repeat revision after first femoral revisions was 11 per cent (four of thirty-six hips). Moreover, in contrast to the study of Pellicci et al., none of the reoperations in the current series was done for fracture of the stem, a reoperation that is known to be associated with a more favorable prognosis4. Marti et. al.24 reported a rate of repeat revision of 5 per cent (three of sixty hips) after 8.9 years, but they used first-generation cementing techniques for some patients and second-generation cementing techniques for others. Moreover, the average age at the time of revision in their series was seventy-one years. The importance of age as a factor in the durability of femoral revision with cement has already been noted.
The present study provides insight regarding the fate of femoral components that are loose according to radiographic criteria. By an average of 15.1 years, repeat revision had been necessary for all of the six femoral components that had been considered loose at either six or 11.7 years.
Our data add support to the argument for the use of second-generation cementing techniques for fixation in revisions of femoral components that have become aseptically loose. In many patients who need revision, the bone stock in the femur is compromised, and there is an increased prevalence of complications associated with revision. Therefore, the results of revision will always be less satisfactory than in an otherwise comparable series of primary reconstructions. However, the improved cementing techniques (plugging of the intramedullary canal, use of a cement gun, improved cleaning of the intramedullary canal, and pressurization of the cement) are believed to have played an important role in the results reported here because of their contribution to better distribution and interdigitation of cement.