The treatment of bone defects in the proximal part of the femur in patients who need a revision total hip arthroplasty can be technically difficult. Disuse osteoporosis and osteolysis, which frequently accompany a loose prosthesis, can compromise the bone stock in the region of the medial portion of the femoral neck that is necessary for the support of a conventional femoral component. A major limb-length discrepancy due to femoral shortening or placement of an acetabular component at a high locus in the pelvis, or both, are additional problems. Occasionally, these defects of bone are so severe that the proximal part of the femur must be replaced completely with a cylindrical proximal femoral allograft. Far more often, the treatment of the defect necessitates the use of a femoral component with a longer head-neck section than is vailable in standard components. To address this special need, Harris and Allen, in 1981, described a calcar-replacement femoral component13.
We are reporting our experience, at a mean of 10.8 years postoperatively, with a series of revision hip replacements in which this calcar-replacement femoral component was used.
*One or more of the authors have 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 or foundation, educational institution, or other non-profit organization with which one or more of the authors are 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.
†Kennedy Orthopaedic Center, Suite 100, Theda Clark Medical Plaza, Neenah, Wisconsin 54956.
‡Orthopaedic Biomechanics Laboratory, GrJ 1126, Massachusetts General Hospital, 32 Fruit Street, Boston, Massachusetts 02114.
Follow-up data were available for thirty-eight hips in thirty-five patients, from a consecutive series of forty-eight hips (forty-four patients) in which we had inserted a calcar-replacement femoral component with cement for a revision operation between April 1976 and December 1986. The mean duration of follow-up was 10.8 years (range, 5.8 to 16.6 years). Of the forty-four patients, one (one hip) had refused follow-up and eight (nine hips) had died of causes unrelated to the hip before the minimum five-year follow-up interval; none of these patients had had a repeat revision. One patient (one hip) had died after the minimum five-year follow-up interval and was included in the study. Eight of the thirty-eight hips had a repeat revision. Of the thirty hips that did not, ten were examined and twenty were evaluated with use of a questionnaire12 and a telephone interview, by the one of us (J. R. McL.) who was not the operating surgeon. Current radiographs were made for all thirty hips.
The mean age of the patients at the time of the operation was fifty-five years (range, twenty to seventy-three years). There were thirteen women (fourteen hips) and twenty-two men (twenty-four hips). Twenty-eight of the index revisions were first revisions, nine were second revisions, and one was a third revision. The original diagnosis was osteoarthrosis in twelve hips, congenital dysplasia in six, fracture in five, rheumatoid arthritis in three, Legg-Perthes disease in two, and slipped capital femoral epiphysis, Paget disease, aseptic necrosis, protrusio acetabuli, and post-traumatic osteoarthrosis in one hip each; in the five remaining hips, the diagnosis was unknown.
The indication for the revision was aseptic loosening of both the femoral and the acetabular component in twenty-four hips, aseptic loosening of the femoral component only in seven hips, fracture of the femoral component in two hips, a bent femoral component in two hips, heterotopic ossification in two hips, and severe osteolysis around a well fixed femoral component in one hip.
The so-called calcar-replacement implant13 is a cast cobalt-chromium long-stemmed femoral component with an extended body (Howmedica, Rutherford, New Jersey). The platform of the extended body is designed to rest on the cortex of the medial aspect of the femur. There are two proximal, or body, lengths (sixty-four and seventy-six millimeters), measured from the center of the femoral head to the junction of the stem with the body. The stem lengths are 150, 200, 250, and 300 millimeters. The initial implant was a so-called monoblock femoral component (non-modular) with a twenty-two, twenty-six, or thirty-two-millimeter diameter of the femoral head.
Both the femoral and the acetabular component were revised in thirty hips, and only the femoral component was revised in eight hips. Of the thirty acetabular revisions, twenty-two were performed with cement and eight, without cement.
In thirty-three hips, a primary indication for use of the calcar-replacement stem was loss of bone from the medial-proximal aspect of the femur. In three additional hips, the indication was the formation of heterotopic bone. These three hips had been operated on before the development of proved modalities for the prevention of recurrent heterotopic ossification. The elongated body of the calcar-replacement stem was used to increase the distance from the acetabulum to the femur. This increased distance was intended to decrease the likelihood of the femur coming into contact with heterotopic bone and thus to maintain the range of motion of the hip in the event that the heterotopic ossification recurred. In two other hips, a calcar-replacement stem was used because of both bone loss in the proximal part of the femur and heterotopic ossification.
In thirteen hips, the center of the hip had been placed proximally, a minimum of thirty-five millimeters proximal to the interteardrop line. This high placement of the center of rotation of the hip necessitated the use of a calcar-replacement femoral component to equalize the lengths of the limbs.
A trochanteric osteotomy was performed in thirty-seven hips. The trochanter was repaired with use of three Vitallium wires in eight hips and four Vitallium wires in twenty-nine16,31. Trochanteric wire mesh14 was used in twenty-eight hips to reinforce the weak bone of the greater trochanter.
A so-called second-generation cementing technique, consisting of use of a polymethylmethacrylate plug and insertion of cement in the doughy phase into the femoral canal in a retrograde fashion, was employed in all thirty-eight hips. Additionally, the use of a femoral compactor22 to pressurize the cement in the femur was introduced during this time interval and was employed in seventeen hips (45 per cent). Reduction of the porosity of the cement5 by centrifugation was introduced in 1982 and was performed in only nine hips (24 per cent).
Radiographs were made for all patients at the time of the review. These included anteroposterior radiographs of the pelvis and the proximal part of the femur, a frog-leg lateral radiograph and a true lateral radiograph of the femur and the hip, and (for twenty hips) Judet radiographs of the pelvis. These radiographs were compared with those that had been made preoperatively, immediately postoperatively, and at each subsequent follow-up visit.
The femoral component was considered loose if there was migration of the component, fracture of the cement, bending or fracture of the stem, or the appearance of any radiolucent line at the stem-cement interface (so-called debonding) that was not present on the immediate postoperative radiographs.
The condition of the bone stock of the femur before the index operation was evaluated radiographically and was graded according to the classification of Estok and Harris10. The femoral bone stock was classified as grade I if there was little or no loss of cortical bone, grade II if there was moderate loss, grade III if there was major loss, and grade IV if there was extreme loss (severe enough that a cylindrical proximal femoral allograft was needed).
The adequacy of the cementing of the femoral component that had been achieved during the index operation was graded on the basis of the appearance of the cement mantle3 on the early postoperative radiographs. Grade A indicated complete filling of the medullary cavity in the diaphyseal area—so-called white-out at the cement-bone interface in the diaphysis. This meant that no distinction could be made between the femoral cortex and the cement mantle. Grade B represented excellent cementing technique, with no voids or defects in the cement mantle but with incomplete white-out, meaning that, in some part of the diaphysis, the cortex could be delineated from the cement mantle. Also, to qualify as grade A or B, there could be no voids in the cement mantle or areas in which the mantle was one millimeter thick or less. Grade C was divided into two categories: grade C1 if there was evidence of a void or voids in the cement mantle, and grade C2 if there was a defect in the cement mantle, meaning that the femoral stem was in contact with the cortex or the cement mantle was one millimeter thick or less. If there were both voids and a defect (or a thin area) in the cement mantle, a grade of C2 was given. If there were gross or multiple defects in the cement mantle, if the cement mantle did not extend beyond the tip of the stem, or if there was a complete radiolucent line along the cement-bone interface on any radiograph, a grade of D was assigned.
Union of the greater trochanter to the proximal part of the femur was evaluated on the most recent radiographs. Heterotopic ossification was evaluated according to the classification of Brooker et al.4.
The locus of the center of the hip was measured on the anteroposterior radiographs of the pelvis. The height of the center of the hip was defined as the vertical distance along a line extending from the center of the femoral head and perpendicular to the interteardrop line. The lateral displacement of the center of the hip was defined as the horizontal distance along the interteardrop line extending from the teardrop to the intersection of the perpendicular line dropped from the center of the femoral head29.
Seven (18 per cent) of the index femoral components were revised again because of symptomatic aseptic loosening. Another, well fixed femoral component was revised again because of extensive lysis of the proximal part of the femur at 13.3 years (Figs. 1-A and 1-B). Four additional femoral components were loose according to radiographic criteria (Figs. 2-A and 2-B). Thus, 68 per cent (twenty-six) of the thirty-eight femoral implants were in place and rigidly fixed more than ten years after the operation. Two of these components were unsupported by the proximal part of the femur secondary to osteolysis but were well fixed distally.
Of the eight calcar-replacement femoral components that were revised again, six had initially been implanted in a first revision operation and two, in a second revision. With the numbers available, no significant difference was found between aseptic loosening or repeat revision and the number of previous revision operations (p < 0.18). The seven repeat revisions that were done because of aseptic loosening were performed at a mean of 117 months (range, seventy-four to 175 months) after the index procedure.
The mean Harris hip score for the thirty hips that had not had a repeat revision was 84 points (range, 66 to 100 points). Twelve (40 per cent) of these thirty hips were rated excellent; nine (30 per cent), good; seven (23 per cent), fair; and two (7 per cent), poor. Twenty-nine hips had mild or no pain, one hip had moderate pain, and no hip had severe pain. Twenty-four (80 per cent) of the thirty hips were associated with a slight limp or no limp and six (20 per cent), with a moderate limp; no patient had a severe limp. Seventeen patients (seventeen hips; 57 per cent) needed no assistive devices, and five (five hips; 17 per cent) used a cane for long walks. Three patients (three hips; 10 per cent) used a cane full-time, and five (five hips; 17 per cent) used two crutches or a walker. Seventeen patients (seventeen hips; 57 per cent) could walk an unlimited distance; five (five hips; 17 per cent) could walk five to six blocks; seven (seven hips; 23 per cent) could walk one to two blocks; and one patient (one hip; 3 per cent), who had had a cerebrovascular accident, could walk inside the house only. Of the seven patients who had a fair hip-rating, four (four hips) had a decrease in function that was due to unrelated causes and was primarily responsible for this rating. Of the two patients who had a poor rating, one was an eighty-six-year-old woman who was markedly disabled because of a cerebrovascular accident. The second patient had mild pain in the left (index) hip and a moderate limp; he used a cane for long walks and could walk only five to six blocks because of a non-union of the greater trochanter.
Evaluation of the cementing technique according to the immediate postoperative radiographs revealed that no hip had a grade of A and only one had a grade of B3. Eleven hips were graded C1, twenty-five were graded C2, and one was graded D. A defect in the cement mantle (grade C2 or D) was present in twenty-six hips (68 per cent). The cementing technique had been grade C1 in one of the eight hips that had a repeat revision of the femoral component and it had been grade C2 in the other seven. The technique was grade C1 in one of the four hips in which the femoral component was loose according to radiographic criteria, grade C2 in two, and grade D in one.
An assessment was made of the relationship between the quality of the cement mantle and the fate of the femoral component. Of the twelve hips without a defect of the cement mantle (grade A, B, or C1), one had aseptic loosening of the femoral component and another had a repeat revision because of aseptic loosening. Of the twenty-six hips that had a defect of the cement mantle (grade C2 or D), three (12 per cent) had a loose femoral component and seven (27 per cent) had been revised again because of aseptic loosening, for a total rate of aseptic loosening of 38 per cent. With the numbers available, this difference was not significant (p = 0.35, Student t test).
Of the thirteen hips that had a high center (thirty-five millimeters or more proximal to the interteardrop line), one had a femoral component that was loose according to radiographic criteria and three had been revised again because of loosening. Of the twenty-five hips in which the center was less than thirty-five millimeters proximal to the interteardrop line, four (16 per cent) had a femoral component that had been revised again and three (12 per cent) were loose, for a total rate of aseptic loosening of 28 per cent. With the numbers available, no significant difference in aseptic loosening was found between the hips with a high center and those in which the center was more distal (p < 0.20).
The bone stock of the femur was assessed10 before the index revision on the preoperative radiographs of all thirty-eight hips. Twenty-two hips were grade I; twelve, grade II; three, grade III; and one, grade IV. Of the eight femoral components that were revised again (seven, because of aseptic loosening, and one, because of femoral lysis with a well fixed implant), four were associated with grade-I lysis; three, with grade-II; and one, with grade-III; no component was associated with grade-IV lysis. Of the four femoral components that were radiographically loose but were not revised again, two were associated with grade-I lysis; one, with grade-II; and one, with grade-III; no component was associated with grade-IV lysis.
Of the thirty-seven hips that had had a trochanteric osteotomy, thirty-two (86 per cent) had radiographic evidence of osseous union and five, of non-union. Of these five hips, one had union after operative repair. Two hips in which union had occurred after the index procedure had separation of the trochanter from the femur at a later date, secondary to lysis of the proximal-lateral aspect of the femur, and the femoral component migrated.
Heterotopic ossification occurred in twenty-nine (76 per cent) of the thirty-eight hips. Of these, nineteen (66 per cent) had class-I ossification4; six (21 per cent), class-II; three (10 per cent), class-III; and one (3 per cent), class-IV. In the five hips in which the calcar-replacement stem had been used for the treatment of severe heterotopic ossification, two had class-II recurrent heterotopic ossification; two, class-III; and one, class-IV.
The acetabular components, although not the focus of the current study, were also evaluated. Thirty acetabular components had been placed at the time of the index revision operation. Of the twenty-two acetabular components that had been inserted with cement, five were revised again because of aseptic loosening and six were loose according to radiographic criteria, for a total rate of loosening of 50 per cent at the most recent follow-up examination. None of the eight acetabular components that had been inserted without cement in the index revision had been revised again, and none were loose according to radiographic criteria. However, the duration of follow-up of the acetabular components inserted with cement was 11.3 years compared with 7.8 years for those inserted without cement. Loosening of the acetabular component was evaluated with respect to a high center of the hip (a minimum of thirty-five millimeters proximal to the interteardrop line), placement of the femoral component in a varus or valgus position, and the estimated length of the abductor muscles. As none of the acetabular components that had been inserted without cement were loose and none were revised again, there was no relationship between loosening and the level of placement of the acetabular component or the length of the abductors for those components. There was also no relationship between those factors for the twenty-two acetabular components that had been inserted with cement in the index revision.
Osteolysis of the cortex of the femur was detected at the time of the most recent follow-up in six (20 per cent) of the thirty hips that had not had a repeat revision of the femoral component. All six components were well fixed; that is, the osteolysis was not associated with loosening. The osteolysis involved one zone, as defined by Gruen et al.11, in three patients, two zones in two patients, and four zones in one patient. Lysis of the femur was noted in one hip with a loose femoral component.
Complications
Intraoperative complications occurred in eleven (29 per cent) of the thirty-eight hips. The complications included perforation of the femoral cortex in ten hips, four of which later had a repeat revision of the femoral component and one of which later had loosening of the femoral component according to radiographic criteria.
Postoperative complications were associated with seventeen (45 per cent) of the thirty-eight hips. There were ten instances of deep-vein thrombosis (detected on bilateral ascending venography, which was done routinely in all patients); four wound hematomas (all treated non-operatively); two instances of bleeding of the upper gastrointestinal tract; one transient arrhythmia; and one partial peroneal-nerve palsy, which resolved. One patient needed a reoperation to remove a broken trochanteric wire that was causing symptoms. Another patient had a reoperation consisting of bone-grafting because of a delayed union of a fracture that had been reduced and stabilized with use of the calcar-replacement femoral component.
Dislocation occurred in eight hips (21 per cent), and one patient had a fracture of the greater trochanter, which was osteopenic.
Two patients (two hips) were managed with prolonged intravenous and oral administration of antibiotics because of a positive culture of joint fluid obtained from the hip at the time of the operation. Cultures of joint fluid aspirated from the hip preoperatively were negative. Neither patient had an infection of the hip at the time of the most recent follow-up (at nine years for one patient and at fourteen years for the other).
The so-called calcar-replacement femoral component may be useful in several types of complex revision total hip replacements, including those requiring a high placement of the acetabular component, those in hips in which a major loss of bone stock has occurred in the proximal portion of the medial part of the femur, and those involving a major limb-length discrepancy (Figs. 3-A and 3-B).
Seven (18 per cent) of the calcar-replacement femoral components in this series of thirty-eight hips needed a repeat revision for aseptic loosening, and eight components (21 per cent) needed a repeat revision for any reason. At the time of the most recent follow-up, thirty (79 per cent) of the implants were still in place, with a mean Harris hip score of 84 points. Four of the thirty hips (11 per cent of the thirty-eight) had definite radiographic evidence of loosening of the femoral component.
The use of a so-called napkin-ring calcar allograft (a circumferential segment of femoral allograft three centimeters in length or less) in association with a standard or long-stem femoral component has been less successful in such patients. Allan et al. reported on the use of thirty such grafts in association with femoral components inserted with and without cement1. Resorption occurred in 50 per cent (fifteen) of the grafts; in twelve of the fifteen, the resorption exceeded one-half of the graft. In fourteen hips, the graft had been used in connection with a femoral component inserted without cement, and eleven had resorption of more than one-third of the graft. One of ten hips in which the graft had been cemented to the prosthesis had resorption of more than one-third of the graft. The grafts subsided in 43 per cent (thirteen) of the thirty hips.
At the time that the index procedures were performed, there was only one size of medullary stem (large) for the calcar-replacement prosthesis. The quality of many of the cement mantles was rated3 as C2 rather than as A or B, solely because the wide medullary stem made contact with the femoral cortex, creating a defect in the mantle. For this reason, the results related to the quality of the cementing technique—the rates of well fixed implants, repeat revision, and femoral lysis—were not as favorable in the current study as in our other series of femoral revisions with use of conventional stems inserted with cement21. Subsequently, an additional version of the calcar replacement was developed with a smaller-diameter medullary stem (the medium-sized stem). In two other series, in which the results of revision with a thinner femoral stem inserted with cement were reviewed at a mean of 8.8 years25 and 11.7 years10, the prevalences of aseptic loosening were 4 per cent25 (two of forty-six hips) and 11 per cent10 (four of thirty-eight hips), and the rates of repeat revision for aseptic loosening were 7 per cent25 (three of forty-six hips) and 11 per cent10 (four of thirty-eight hips). These values are all lower, at corresponding durations, than those obtained in the present series, in which a large-stem component was used. In the study in which the mean duration of follow-up was 8.8 years, the prevalence of defects in the cement mantle (grade C2) was only 13 per cent (six of forty-six)25. In the study in which the mean duration of follow-up was 11.7 years, femoral osteolysis (excluding that in zones 1A and 7A) occurred in only 8 per cent (three of thirty-eight) of the hips10.
Fixation of the femoral component with so-called first-generation cementing techniques (finger-packing of the cement in the femur without the use of a cement gun, a medullary plug, pressurization, or reduction of the porosity of the cement) and, often, the use of a cast femoral stem with a narrow medial border and sharp corners in revision hip operations have led to inadequate fixation and high rates of failure2,7-9,17-19,23,24,26-28,32,33. Kavanagh et al. reported a prevalence of probable loosening of the femoral component of 43 per cent (seventy-two of 166 hips) at a mean of 4.5 years postoperatively18. Amstutz et al. reported a rate of repeat revision of 9 per cent (six of sixty-six) in a series that was followed for only 2.1 years2. Strömberg et al. reported their experience with 4664 revision total hip arthroplasties performed with cement in Sweden. In a subgroup of 204 hips in patients between fifty and seventy years old, 27 per cent of the femoral components were loose radiographically, 10 per cent of which had been revised again for aseptic loosening, at a mean of six years after the operation33. Similar results have been reported by Pellicci et al.23,24, Retpen et al.26,27, Eftekar et al.8, and others6,15,32,33.
The poor results with use of first-generation cementing techniques in femoral revision total hip arthroplasty led to the use of femoral components inserted without cement as an alternative. We know of no other investigators who studied, as we did, the ten-year results of the use of femoral components inserted without cement in revision hip operations and, specifically, who provided long-term data on patients who had deficient bone stock in the medial-proximal aspect of the femur.
Contemporary cementing techniques involving the use of a femoral plug, a cement gun, and so-called super-alloy metal stems with rounded corners have markedly improved the duration of fixation of femoral components in primary and revision total hip arthroplasties that do not require use of a calcar-replacement device3,10,20,30. Mulroy and Harris reported a 3 per cent prevalence of loosening of the femoral component (three of 105) at a mean of eleven years after primary total hip replacement performed with such techniques20. Estok and Harris reported the results in thirty-eight patients at a mean of 11.7 years after revision of the femoral component with so-called second-generation cementing techniques10. Repeat revision of the femoral component for aseptic loosening was performed in four of the thirty-eight patients and an additional four had a loose component, for an over-all prevalence of loosening of the femoral component of 21 per cent (eight of thirty-eight). This represents a marked improvement in the fixation of the femoral component compared with the results reported with use of first-generation cementing techniques.
Our results with use of a calcar-replacement femoral component for the special problems in our patients were favorable. It is likely that use of the medium rather than the large-stem components employed in the current series will improve the quality of the mantle of cement.