|
|
 |
| home | CME | help | feedback | subscriptions | archive |
|
© 1997 The Journal of Bone and Joint Surgery, Inc. Impaction Bone-Grafting before Insertion of a Femoral Stem with Cement in Revision Total Hip Arthroplasty. A Minimum Two-Year Follow-up Study* , MOORESVILLE, INDIANAInvestigation performed at the Center for Hip and Knee Surgery, Mooresville
Impaction bone-grafting was performed before insertion of a collarless, polished, tapered femoral stem with cement in thirty-four revision total hip arthroplasties (thirty-four patients) that were done because of aseptic loosening. The average duration of follow-up was thirty months (range, twenty-four to forty-two months). The operation was the initial revision in twenty-eight patients (82 per cent). Twenty-two patients (65 per cent) also had revision of the acetabular component. Complications included four intraoperative and two postoperative fractures of the femur as well as one dislocation (at one month). Two patients (6 per cent) needed a repeat revision of the femoral stem because of aseptic loosening at twenty-six and thirty-six months postoperatively. Both of these patients had an associated fracture of the femur (one was intraoperative, and the other was postoperative). Subsidence was common (thirteen patients; 38 per cent) and averaged 10.1 millimeters (range, four to thirty-one millimeters). Although the study group was relatively small, with the numbers available subsidence was not found to be associated with the preoperative or postoperative hip score, segmental or cavitary femoral defects, femoral ectasia, intraoperative fracture of the femur, strut-grafting, trochanteric osteotomy, or varus position of the femoral component. Incorporation of the allograft into the trabecular bone and secondary remodeling were noted radiographically in thirty-two (94 per cent) and fourteen (41 per cent) of the patients, respectively, often within one year. Although the duration of follow-up was relatively short, no localized resorption of the allograft occurred and cortical repair was noted in one patient at three years. At the most recent follow-up evaluation, the Harris hip scores had improved from a preoperative average of 51 points (range, 32 to 90 points) to an average of 87 points (range, 65 to 100 points) and twenty-eight patients (82 per cent) had no or only slight pain. Despite the satisfactory early clinical results, we remain concerned about the high rate of fracture of the femur and the rate and extent of subsidence of the femoral component. On the basis of the worrisome findings after this two-year period, we recommend that impaction bone-grafting be used only when proximal femoral osteopenia is so severe that stability cannot be obtained with insertion of a long-stemmed femoral component without cement. In that setting, impaction bone-grafting may be considered instead of implantation of a massive proximal femoral allograft in combination with insertion of a femoral component with cement.
The loss of bone stock secondary to particles of debris and mechanical instability presents a major challenge in revision total hip arthroplasty10, and the best way to achieve a successful result remains controversial1,6,10,20. Because the bone stock of the proximal part of the femur is often deficient at the time of revision, the identification of a reliable, reproducible approach to these difficult procedures is of continued interest. In addition to other techniques, involving insertion of the prosthesis with or without cement1,4,8,12,14,17, an approach in which cancellous allograft is impacted to reconstitute a deficiency of the proximal femoral bone stock before a collarless, polished, tapered stem is inserted with cement recently has been described6,7,10,15,18,20. Early reports have documented partial restoration of the proximal femoral bone stock, both histologically15,18 and radiographically, with use of this technique6,7,10. Satisfactory early clinical results (at one and one-half to five years) have been reported in two series (of fifty-six patients each), with more than 80 per cent of patients being pain-free7,10. Given the potential for reconstruction of a hip that has deficient proximal femoral bone stock, these results certainly appear promising. The purpose of the current study was to clarify the role of impaction bone-grafting in revision hip operations. We report not only our early clinical results, including complications, but also the degree of radiographic changes (remodeling of the bone graft and reconstitution of bone) in the proximal part of the femur.
In March 1992, we began to use the technique of impaction bone-grafting before insertion of a femoral stem with cement in revision arthroplasty at the Center for Hip and Knee Surgery. By November 30, 1993, thirty-four revision arthroplasties (thirty-four patients) had been performed with this technique by the four of us (M. A. R., twenty procedures; E. M. K., ten procedures; P. M. F., three procedures; and J. B. M., one procedure). During this period, a total of 163 revision hip arthroplasties were performed at our institution. Ninety-nine of these revisions involved both the acetabular and the femoral component; twenty-nine, only the femoral component; and thirty-five, only the acetabular component. Therefore, impaction bone-grafting was used in 27 per cent of the 128 revisions of the femoral component performed during this period. The technique was chosen when marked deficiency of the proximal femoral bone stock was noted preoperatively and when reconstitution of the proximal femoral bone stock was deemed necessary for long-term fixation and function of the prosthesis. In all instances, the proximal femoral bone stock was believed to be insufficient for revision with use of cement as well. All patients were followed for an average of thirty months (range, twenty-four to forty-two months). The average age (with standard deviation) at the time of the revision was 69.6 ± 10.4 years (range, forty-one to eighty-six years). Six patients were less than sixty years old, and six were more than eighty years old. There were eighteen women (53 per cent) and sixteen men (47 per cent). The indication for the revision hip arthroplasty was aseptic loosening of the femoral component in all patients. Twenty-six patients (76 per cent) had a loose cemented stem, and eight (24 per cent) had loosening of a stem that had been inserted without cement. Three patients (9 per cent) had loosening after a hemiarthroplasty. The original diagnoses included primary osteoarthrosis in twenty-two patients (65 per cent); post-traumatic osteoarthrosis in three (9 per cent); a previous fracture of the femoral neck in five (15 per cent); and congenital dysplasia of the hip, avascular necrosis, Legg-Calvé-Perthes disease, and pyarthrosis in childhood in one patient each. No patient had an active infection, and only one had had a previous infection at the site of the prosthesis. Twenty-eight patients (82 per cent) had had no previous revision of the hip; three (9 per cent) had had one previous revision involving both components; one (3 per cent) had had two previous revisions, with the first involving both components and the second involving the femoral component only; and two (6 per cent) had had three previous revisions, with the first and second involving both components and the third involving the femoral component only. Thus, all patients who had had a previous revision had had at least one revision of both the acetabular and the femoral component. The average interval between the most recent previous revision and the index procedure was 11.8 years (range, three to twenty-six years). During the index procedure, three patients (9 per cent) had a primary acetabular arthroplasty (revision of a hemiarthroplasty) and nine (26 per cent) did not have revision of the acetabular component. The remaining twenty-two patients (65 per cent) had revision of the acetabular component secondary to aseptic loosening. In fourteen patients (41 per cent) the acetabular component (Nelson Cup; Biomet, Warsaw, Indiana) was inserted without cement, and in eight (24 per cent) an all-polyethylene acetabular component (Bioclad; Biomet) was inserted with cement. A protrusion ring was used in six revisions (18 per cent). The operative technique of impaction bone-grafting has been well described in the literature1,6,10,18. Initially, a thorough débridement of the proximal femoral canal, including all fibrous tissue and cement, was performed. Next, segmental cortical defects, if present, were reinforced with strut allografts and cerclage wires. An attempt was made to secure the graft at a distance of at least twice the femoral diameter both proximal and distal to the defect. Strut-grafting was used in nine patients (26 per cent); two patients had one strut, six patients had two struts, and one patient had four struts. Cerclage wires were used prophylactically without strut grafts in eight patients (24 per cent) because the integrity of the cortical shell raised concern regarding the potential for fracture during the impaction of the graft. The femoral canal then was plugged at least two to three centimeters distal to the most distal cortical defect. In three patients, retained cement was used as a distal plug; in one patient, a bone pedestal was used; and in one patient, a retained prosthetic tip was used. In the remaining patients, a cement restrictor was used in the femoral canal. The canal was then filled, starting distally, with morselized cancellous allograft prepared from fresh-frozen femoral heads that had been retrieved during primary total hip replacements at our institution. Thus, a so-called neomedullary canal was formed10. Finally, a CPT (collarless, polished, tapered) femoral stem (Zimmer, Warsaw, Indiana) was inserted with cement into the cancellous allograft after a trial reduction, according to the technique described by Gie et al. The CPT stem is a highly polished, forged, cobalt-chromium-molybdenum-alloy component with a medial and a lateral taper, measuring 130 millimeters in length. Although it is available in five different sizes, the largest size was not used in this series. In all procedures, a distal polyethylene centralizer was used on the tip of the implant and a twenty-eight-millimeter-diameter cobalt-chromium femoral head was implanted. Walking was initiated on the first day after the operation, with toe-touch weight-bearing continued for eight weeks. This was followed by progressive weight-bearing during the next six to eight weeks, with gradual return to full weight-bearing. All but six patients, who had had an osteotomy of the greater trochanter, were allowed to perform active abduction immediately after the operation. The quality of the femoral bone stock, determined radiographically and grossly at the time of the operation, was graded according to the criteria of The American Academy of Orthopaedic Surgeons Committee on the Hip5. Segmental defects were noted in eight patients (24 per cent); cavitary defects, in twenty-eight (82 per cent); and femoral ectasia, in eleven (32 per cent). Combined cavitary and segmental defects were noted in three patients (9 per cent), and femoral malalignment and femoral stenosis were noted in one patient each. No patient had femoral discontinuity. Clinical scores were determined with the functional rating system described by Harris and were recorded at each follow-up evaluation. The location and pattern of pain (when noted) also were recorded. Complications were classified as intraoperative, early (occurring within six months postoperatively), or late (occurring more than six months postoperatively). The patients were evaluated at eight weeks, six months, one year, two years, and three years. Standard radiographs, including a low anteroposterior radiograph of the pelvis and an anteroposterior and a Lauenstein lateral radiograph of the entire femoral shaft, were made initially (in the recovery room) and at each follow-up interval. At each follow-up evaluation, the patient completed two self-administered questionnaires: the WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index)2, which is a disease-specific evaluation of pain, stiffness, and function, and the HSQ-36 (Health Outcomes Institute, Minneapolis, Minnesota)3,10, which measures not only functional status but also the patient's general perception of health and well-being. The questionnaires were not filled out preoperatively. The distance from the tip of the stem to the plug was measured, in millimeters, on the initial postoperative radiograph. The varus and valgus position of the stem was measured, according to the method described by Salvati et al., at each follow-up interval. Although any change in the varus or valgus position of a cemented stem may be very important, a change of 5 degrees or less was considered to be within the range of measurable error. Subsidence of both the stem within the cement and the stem-and-cement construct as a whole was measured radiographically, according to the method described by Fowler et al., at each follow-up interval. Radiolucent lines, both between the prosthesis and the cement and between the cement and the allograft, were classified on each postoperative radiograph with use of the zonal distribution system described by Gruen et al. Incorporation of the allograft was evaluated subjectively with a review of each radiograph. With use of a method similar to that described by Gie et al., the radiograph was judged as showing no evidence of healing if it demonstrated no change in the allograft. Any change in radiodensity within the allograft was considered indicative of primary healing, a confluent area of radiolucency within the allograft was equated with localized resorption of the graft, stress-oriented trabeculae within the allograft were judged as evidence of trabecular remodeling or secondary healing, and reconstitution of the cortical shell was noted as cortical repair. The zonal distribution of these radiographic changes was recorded with use of the system of Gruen et al. Because the assessment of graft incorporation is quite subjective, all radiographs were evaluated by one of us (J. B. M.) who was blinded with regard to which surgeon had performed the revision arthroplasty, in order to minimize observer bias. The Harris hip scores and the radiographic parameters were evaluated statistically for causal relationships with use of the Fisher exact test (two-tailed) and the Wilcoxon two-sample test, with the aid of the Statistical Analysis System (SAS Institute, Cary, North Carolina).
Clinical Scores
Complications Early complications included a postoperative fracture of the femur, heterotopic ossification (islands of bone within the soft tissues), a dislocation at one month that was treated successfully with closed reduction, and a symptomatic deep-vein thrombosis necessitating therapeutic anticoagulation in one patient each. The fracture occurred at three weeks and was treated with open reduction and internal fixation; despite evidence of healing, the patient had progressive pain in the groin and buttocks and shortening of the involved lower extremity. Twenty-six months after the index revision hip replacement, a repeat revision, with insertion of a calcar-replacement prosthesis without cement, was performed because of aseptic loosening. Intraoperatively, the fracture was noted to be healed. Late complications included one fracture, loosening of two stems, and non-union at the sites of three (of six) osteotomies of the greater trochanter. The average hip score for the latter three patients was 81 points (72, 85, and 86 points). As of this writing, none of the three patients had had a revision. The late fracture of the femur occurred secondary to a fall approximately twelve months after the revision. It healed uneventfully after open reduction and internal fixation.
Repeat Revisions Of the fourteen index acetabular revisions without cement and the eight with cement, none had been followed by a repeat revision as of this writing.
Radiographic Results Cavitary defects were noted about the plug in ten patients (29 per cent) and about the tip of the stem in seventeen (50 per cent). It is noteworthy that both postoperative fractures occurred at a level well proximal to the plug. In both of these patients, cavitary defects were noted about the tip of the stem. The average distance from the tip of the stem to the plug in the entire series was 32.3 millimeters (range, thirteen to seventy-five millimeters). No patient had a radiolucent line at the bone-cement interface about the stem, and only two patients had a radiolucent line at the prosthesis-cement interface. Both of these patients had a radiolucent line in zone 1; one also had a radiolucent line in zone 6, and the other also had a line in zones 2 and 7. All radiolucent lines had a maximum thickness of less than two millimeters and were non-progressive. Although no patient in this series needed a revision of the acetabular component, there was loosening of one cup that had been inserted without cement. Subsidence of the stem within the cement occurred in three patients (9 per cent) and averaged 5.3 millimeters (five millimeters in two patients and six millimeters in one). Measurable migration of the stem and the cement mantle was noted within the allograft in twelve patients (35 per cent) and averaged 11.3 millimeters (range, four to thirty-one millimeters). The stem and the cement subsided in two patients who had a fracture of the femur. If the data for these two patients are omitted, then measurable subsidence of the stem and the cement in the remaining ten patients averaged 9.2 millimeters (range, four to twenty-five millimeters). Overall, subsidence occurred in thirteen patients (38 per cent) and averaged 10.1 millimeters (range, four to thirty-one millimeters) (Table I). Subsidence averaged 3.9 millimeters for the series as a whole. In general, most subsidence (average, 9.5 millimeters) occurred within six months, but in seven patients it occurred between six months and one year and in four patients, between one year and two years.
We found no significant association, with the numbers available, between subsidence and any single parameter that was evaluated in this series (hip score, femoral defects, intraoperative fracture of the femur, use of strut-grafting, osteotomy of the greater trochanter, distance from the tip of the stem to the plug, varus position of the prosthesis, and location of cavitary defects). No measurable subsidence was noted in the three patients who had a non-union of the greater trochanter. Most changes involving primary and secondary trabecular repair were noted before one year, although primary trabecular healing was noted as late as two years (Figs. 1-A, 1-B, 2-A, 2-B and Fig. 2-C). There was a general trend toward increased primary trabecular repair and secondary remodeling in the more proximal zones (Table II). All patients had trabecular healing in at least one zone. No patient had localized trabecular resorption, and cortical repair was noted in only one patient (Figs. 3-A, 3-B, and 3-C), in zone VII, at three years.
Patient-Related Outcomes Twenty-six patients (76 per cent) needed no support for walking at the time of the latest follow-up. Of the remaining eight patients, four (12 per cent) used a cane part-time, three (9 per cent) used a cane full-time, and one (3 per cent) used a walker full-time. Thirty-two patients (94 per cent) stated that they were able to walk about the community, and two (6 per cent) stated that they were able to walk about the house only. Twelve patients (35 per cent) believed that their overall general health was much better at the time of the most recent follow-up than it had been preoperatively, six (18 per cent) thought that it was somewhat better, thirteen (38 per cent) considered it to be the same, and three (9 per cent) maintained that it was somewhat worse. Although twenty-one patients (62 per cent) stated that they were pain-free, fewer patients stated that this was the case when they were questioned specifically about physical function. Twenty-seven patients (79 per cent) stated that they were pain-free while lying in bed, but only eighteen (53 per cent) reported that they were pain-free while performing light domestic duties. Seventeen patients (50 per cent) stated that they were pain-free while getting in and out of a bath, and fourteen (41 per cent) reported being pain-free while shopping or getting in and out of an automobile. When asked about heavy domestic duties, only ten patients (29 per cent) stated that they were pain-free, adding that they tended to avoid such activity.
The technique of impaction allografting before insertion of the femoral stem with cement is a relatively new approach to revision hip arthroplasty. The polymethylmethacrylate, inserted under pressure, provides the initial stability for the stem-cement-bone graft construct, with the potential advantage of reconstitution of the proximal femoral bone stock7,10. Gie et al. noted a satisfactory result in most of fifty-six patients at an average of thirty months postoperatively. Elting et al.6 noted that 74 per cent of twenty-seven patients had no pain in the early period after impaction bone-grafting. In a later report from the same institution7, those authors stated that fifty-six patients had an average hip score of 90 points and 80 per cent were pain-free at an average of thirty-one months. Concern has been raised about perioperative fractures of the femoral shaft associated with this impaction technique1,7,10. We agree with Elting et al.7 that cortical strut-grafting should be considered when the integrity of the cortical shell is questionable. A major problem associated with these fractures is loosening of the stem. One of the two postoperative fractures of the femur in the current series progressed to loosening of the stem and necessitated a subsequent revision. Subsidence of this type of tapered stem has been reported previously7,10. Elting et al.7 noted that twenty-seven (48 per cent) of fifty-six patients had subsidence (average, 2.8 millimeters), and Gie et al. noted that forty-four (79 per cent) of fifty-six patients had subsidence (average, 6.1 millimeters) within the polymethylmethacrylate mantle. Although subsidence of more than one millimeter did not occur in more than 60 per cent of our patients, the remainder had subsidence that averaged 10.1 millimeters (range, four to thirty-one millimeters). Even if the data for the two patients in whom the subsidence followed a fracture of the femoral shaft are excluded, subsidence averaged 9.6 millimeters in the remaining eleven patients. However, subsidence of these cemented, collarless, polished, tapered stems has been well documented within the first two years after both primary and revision operations6,7,9,10,20. Because of the geometry of the stem, which allows self-tightening9,10,20, some degree of subsidence is anticipated. In the current series, subsidence was first noted in seven hips at one year and in four hips at two years. Although Elting et al.7 noted that the subsidence ceased after two years, we remain concerned about the possibility of late subsidence and the degree of incorporation of the allograft in these patients. Unfortunately, on the basis of our findings and the limited number of patients available for study, both the rate and the overall amount of subsidence of the stem remain unpredictable. Furthermore, the overall clinical score may not accurately reflect the rate or amount of subsidence. According to the functional rating system of Harris, points are deducted only if the limb-length discrepancy is more than three centimeters; also, only 3 points are deducted for a slight limp and 6 points are deducted for a moderate limp. Even if subsidence occurs, it may cease because of the design of the stem9,10,20, and the patient can have a Trendelenburg gait and an abductor lurch despite a clinically stable stem. Therefore, neither subsidence of the stem nor limb-length discrepancy is always associated with an inferior clinical result (hip score). The apparent discrepancy between the Harris hip and WOMAC scores for pain reflect the more specific nature of the WOMAC evaluation and the more limited nature of the Harris pain score. The current study may be criticized on the basis of the relatively small number of patients and the so-called learning curve, which may increase the risk of subsidence of the stem. Only two surgeons performed all thirteen operations that were followed by subsidence, and these surgeons performed most of the revisions. Also, only two of the fifteen stems that were implanted early in the series (between March and December 1992) subsided. Therefore, subsidence did not appear to be directly related to the operative technique. Radiographic changes consistent with incorporation of the impacted allograft within the proximal part of the femur have been reported6,7,10,18,20. We realize that these changes were not confirmed by documented vascular ingrowth or histological evidence of remodeling of the allograft10. Nevertheless, a relative increase or decrease in radiodensity within the allograft was considered to be consistent with vascular ingrowth and primary repair. The presence of stress-oriented trabeculae, indicative of secondary trabecular remodeling, is consistent with the compressive forces on the cement mantle as the collarless, polished, tapered stem not only engages the cement7,9,10,18 but also loads the morselized allograft in compression7,15,20. Remodeling of the allograft has been documented histologically. Ling et al. evaluated a biopsy specimen obtained seven years after impaction bone-grafting and found that the allograft had been replaced by viable bone peripherally, with dead trabeculae entombed in cement located more centrally about the cement column. Nelissen et al., in a histological analysis of four biopsy specimens obtained during removal of trochanteric wire eleven to twenty-seven months after revision hip arthroplasty, confirmed the presence of viable trabeculae as well as new cortical bone formation. On the average, one or two femoral head allografts were used for each of our patients, at a cost of $950 per allograft. The price of the CPT femoral stem averaged $2000 during this period. In the current era of cost containment, the expense of not only the prosthesis and the femoral head allografts but also the numerous strut allografts ($500 each at our institution) must be considered. In conclusion, the impaction bone-grafting technique provided satisfactory early clinical results with respect to hip scores and pain. Trabecular incorporation and secondary remodeling were noted radiographically in thirty-two (94 per cent) and fourteen (41 per cent) of our patients, respectively. We remain concerned, however, about the relatively high rate (38 per cent; thirteen patients) and the unpredictable nature of early subsidence of the femoral component as well as about the number of femoral fractures (six patients; 18 per cent) associated with this technique. Although impaction bone-grafting may be useful for revision hip operations, we no longer employ it routinely. However, it may be useful when the deficiency of the proximal femoral bone stock is so severe that stability cannot be achieved with use of a long-stemmed component inserted without cement. The cortical shell must remain intact, and we recommend that, when the bone-impaction technique is used, strut-grafting be strongly considered if the cortical bone is intact but deficient.
*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.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |