Sixty-eight acetabular revisions in sixty-seven patients were performed without cement by the senior one of us between March 1985 and January 1991. Eleven hips (eleven patients) were excluded from the study because they had been followed for less than five years. Thus, fifty-seven hips (fifty-six patients) were included in the present study. Of the original fifty-nine patients (sixty revisions) who were reported on in the previous study10, eight (eight revisions) died less than five years after the procedure and three (three revisions) could not be located. These eleven patients had a well functioning prosthesis at the time of the last follow-up. Eight acetabular revisions that had not been included in the original cohort and that had been followed for at least five years were included in the present investigation. During the study period, no other type of component or technique was used for acetabular revision; therefore, we are reporting on a prospective, consecutive series.
Forty-five hips had revision of both the femoral and the acetabular component, and twelve hips had an isolated acetabular revision. A cemented acetabular component was revised in thirty-five hips; a bipolar prosthesis, in eight; an acetabular component without cement, in five; a cemented surface replacement, in four; a unipolar hemiprosthesis (with protrusio acetabuli), in three; and a cup arthroplasty component, in two. Seven of the fifty-seven hips had had a chronic infection, and the acetabular component was inserted without cement during the second stage of a two-stage reimplantation.
All patients were followed for a minimum of five years; the mean duration of follow-up was seven years (range, five to twelve years). Twenty-seven patients (twenty-eight hips) were men and twenty-nine patients (twenty-nine hips) were women. The mean age of the patients at the time of the revision was fifty-six years (range, twenty-two to eighty-two years), and the mean weight was seventy-five kilograms (range, forty-three to 107 kilograms).
The initial diagnosis was osteoarthrosis for twenty-two hips, osteonecrosis for sixteen, traumatic osteoarthrosis for six, fracture of the femoral neck for five, rheumatoid arthritis for three, congenital dislocation or acetabular dysplasia for two, slipped capital femoral epiphysis for two, and ankylosing spondylitis for one.
The indication for the revision was painful aseptic loosening in forty-six hips, reimplantation after resection arthroplasty in seven, recurrent dislocation in two, and extrusion and fracture of a bipolar polyethylene liner in two. One patient had aseptic loosening in one hip and recurrent dislocation in the other. The acetabular revision was the first such operation in forty-three hips, the second in eleven hips, and the third in three hips.
The acetabular deficiencies were classified according to the system of the American Academy of Orthopaedic Surgeons Committee on the Hip4. Seven hips had a type-I defect (segmental bone loss); twenty-three, a type-II defect (cavitary bone loss); and twenty-one, a type-III defect (combined cavitary and segmental bone loss). Six hips had no notable deficiency after reaming.
Bone-grafting was performed in forty-five hips (79 per cent). Only morseled autogenous graft from the iliac crest was used in twelve hips, only morseled fresh-frozen or freeze-dried allograft was used in twenty hips, and a combination of morseled allograft and autogenous graft was used in two hips. Fresh-frozen bulk allograft was used in the remaining eleven hips, in combination with some amount of morseled allograft. The bulk allograft was used to augment a large (more than 50 per cent) deficiency of the anterior column in one hip and a large defect of the posterosuperior aspect of the acetabulum in ten hips. A bulk allograft was used only if the host bone could not support the component; its purpose was to allow a mechanical interference fit in a nearly anatomical position. At the time of these procedures, a so-called jumbo (more than seventy-millimeter-diameter) acetabular component was not available. The allografts were secured to the pelvis with either titanium or stainless-steel screws. These screws did not come into contact with the acetabular component. A 10-degree-elevated-rim liner was used in seventeen hips, and a standard liner was used in forty.
A Harris-Galante-I porous-coated acetabular component (Zimmer, Warsaw, Indiana) was used in fifty-six hips, and a Harris-Galante-II porous-coated component (Zimmer) was used in the remaining hip. Both of these components are hemispherical and fully porous-coated; the shell is made of titanium alloy, and the coating is made of titanium-fiber mesh. The Harris-Galante-I porous-coated components were fixed with multiple 4.5 and 5.1-millimeter screws made of chemically pure titanium. The Harris-Galante-II porous-coated component was fixed with three 6.5-millimeter screws made of chemically pure titanium.
The revision arthroplasty was performed through a transtrochanteric approach in forty hips (70 per cent) and through a posterior approach with an extended anterior capsulectomy in seventeen hips (30 per cent). After removal of the acetabular component, cement (if present), and the soft-tissue membrane, the outer edge of the acetabulum was reamed gently. Morseled allograft was then packed into the acetabular defect (if present) and was shaped into a hemisphere with the acetabular reamer in the reverse mode. The acetabular component was press-fit between the remaining portions of the anterior and posterior columns of the acetabulum. The last reamer shell served as a trial component, and the component that was actually used for the procedure was the same size as that reamer shell. Additional fixation was achieved with a mean of 4.8 screws (range, three to seven screws), which were usually directed posteriorly and superiorly. Occasionally, screws were placed medially, toward the iliopectineal line. All components were considered to be rigidly fixed at the time of the revision operation.
We performed the latest clinical follow-up evaluation of fifty-three patients (fifty-three hips). Three patients (four hips) could not return to our institution for follow-up because of the distance involved, and they were interviewed by telephone with use of a questionnaire. Radiographs of these four hips were made at other institutions according to our guidelines. The clinical function of the hips was assessed with use of the Harris hip score6, both preoperatively and at the most recent follow-up evaluation. A score of 90 to 100 points was considered excellent; a score of 80 to 89 points, good; a score of 70 to 79 points, fair; and a score of 69 points or less, poor.
Anteroposterior radiographs of the hips and pelvis, centered over the pubis, and frog-leg lateral radiographs of the hip were made in the same radiology department (except for the four hips that were evaluated at other institutions) in a standardized fashion with the same tube angle and distance. One of us (E. D. P.), who was not involved in the care of these patients, measured the position of the component on sequential radiographs of the pelvis. Migration was defined as a change in the angle of the cup of 5 degrees or more or a change in the horizontal or vertical position of the cup of more than three millimeters. Radiolucent lines around the component and the screws were measured with use of the zones described by DeLee and Charnley. Only lines that involved 50 per cent or more of a zone were considered important.
Heterotopic ossification was classified according to the method of Brooker et al. A bone graft was considered to be united to the host bone if there was trabecular bridging. Polyethylene wear was measured with an electronic digital caliper according to the method described by Livermore et al., as modified for acetabular components inserted without cement.
Clinical Results
The mean preoperative Harris hip score was 42 points (range, 0 to 82 points), and the mean postoperative score was 84 points (range, 53 to 100 points). Twenty hips (35 per cent) had an excellent result, nineteen (33 per cent) had a good result, fourteen (25 per cent) had a fair result, and four (7 per cent) had a poor result.
Three of the hips that had a poor result had complications unrelated to the acetabular component. Two of these three hips had a reoperation for loosening of the femoral component, and one had operative drainage and débridement because of a late, deep metastatic infection. The acetabular component in all three of these hips was well fixed and was not removed.
Thirty-one hips (54 per cent) were not painful; twenty-two (39 per cent) were only slightly or occasionally painful; three (5 per cent) were mildly painful after unusual activities, but the pain was relieved with aspirin (or another non-steroidal anti-inflammatory medication); and one (2 per cent) was moderately painful, necessitating modification of activities and concessions and occasional use of narcotics. No hip was severely painful.
Twenty-one hips (37 per cent) caused no limp, twenty (35 per cent) caused only a slight limp, fifteen (26 per cent) caused a moderate limp, and one (2 per cent) caused a severe limp.
Twenty-four patients (43 per cent) did not need to use a support to walk, seventeen patients (30 per cent) (eighteen hips) used a cane outdoors only for long walks, nine (16 per cent) used a cane full-time, and six (11 per cent) used two supports.
The infection did not recur in any of the seven hips in which the acetabular component had been inserted during the second stage of a two-stage reimplantation.
Radiographic Results
No acetabular component had been revised or was scheduled to be revised at the time of the latest follow-up examination. No acetabular component had migrated or was loose. Radiolucent lines were frequently seen in the zones in which morseled allograft had been used. Twenty-four hips had a thin, non-progressive radiolucent line in zone III (Figs. 1-A and 1-B), but only eleven lines were at least one millimeter wide. Seventeen hips had a radiolucent line in zone I, but only four lines were at least one millimeter wide. Eight hips had a radiolucent line in zone II, but only three lines were at least one millimeter wide. A total of eighteen hips had a radiolucent line that was at least one millimeter wide, and no hip had a radiolucent line that was that wide in more than one zone.
There were no radiolucent lines around any screw, and there were no broken screws. No component had separation of the mesh from the shell.
The eleven bulk allografts were considered to have healed, as evidenced by trabecular bridging. Four bulk allografts had partial lateral resorption, but it was not progressive. Twenty-seven of the twenty-nine hips in which the morseled bone graft had been placed medially had resorption and remodeling of the graft, which led to the formation of a thin (one-to-two-millimeter) sclerotic rim. In the other two hips, the graft resorbed without the formation of a rim. Two hips had areas of minor peripheral pelvic osteolysis.
The mean rate of polyethylene wear was 0.06 millimeter per year (range, 0.01 to 0.43 millimeter per year).
Complications and Reoperations
There were no vascular or neural complications associated with the multiple small screws used for fixation.
A late, deep metastatic infection developed in one hip 2.7 years after a revision that had been performed because of aseptic loosening of both components of a cemented total hip prosthesis. The patient had had Hodgkin disease, which had been treated with a splenectomy and irradiation of the neck. The infection was treated with emergent drainage and débridement of the hip; the components were rigidly fixed and were not removed. At the time of the latest follow-up, the patient had occasional pain in the hip but the clinical result was considered poor because of a severe limp. The patient was able to walk about the house only because a contralateral resection arthroplasty had been done for a chronic infection. Leukemia later developed, and the patient died six years after the revision with the components in place.
Seven patients had a reoperation for removal of trochanteric hardware because of symptomatic bursitis. Eight patients had a non-union at the site of the osteotomy of the greater trochanter.
Four hips, all of which had had a revision of both components, had a dislocation. Three of the dislocations occurred in the immediate postoperative period, and the hips were immobilized in a hip-spica cast for six weeks. The fourth dislocation was associated with an acute avulsion of the greater trochanter at eight weeks. The greater trochanter was repaired after a closed reduction, and it subsequently united.
One patient, who had an arthrogryposis-like syndrome and had had a complex revision of both components, had rhabdomyolysis and a gluteal compartment syndrome on the contralateral side. This patient had a fasciotomy of the involved compartment with subsequent complete recovery. Another patient, who had had a complex revision that included fixation of a femoral fracture, had contralateral rhabdomyolysis and a contralateral sciatic-nerve palsy, presumed to be due to pressure from a pelvic bolster, that resolved completely.
In our prospective study of the mid-term results of fifty-seven acetabular revisions performed by one surgeon, insertion of a titanium fiber-metal-coated hemispherical component with multiple screws and no cement provided stable fixation, even in the hips that had severely deficient acetabular bone. None of the revision acetabular components were revised again because of loosening, and no revisions were pending at the time of the latest follow-up, a mean of seven years postoperatively. No component was surrounded by a progressive, circumferential radiolucent line or had migrated. Many hips had a thin, non-progressive radiolucent line in the zones in which morseled bone graft had been placed. Medially placed bone graft usually remodeled into a thin, sclerotic rim.
The results of the present study appear to be superior to those reported after acetabular revisions performed only with cement3,8,14,16. Another approach to acetabular revision with cement has been to utilize a reinforcement ring or cage supplemented with cancellous allograft. Peters et al. reviewed the results at a mean of thirty-three months after twenty-eight revisions done with the Burch-Schneider antiprotrusio cage. Migration was noted after 14 per cent (four) of the reconstructions. Zehntner and Ganz reported the mid-term results of twenty-seven revisions performed with an acetabular reinforcement ring. Migration was seen in twelve hips (44 per cent) and in six of the twelve hips that had segmental defects.
An important goal of acetabular revision arthroplasty is the reconstitution of the acetabular bone structure. Brien et al. used morseled bone graft and a bipolar prosthesis to restore bone stock in eighteen hips, but the graft was maintained in only four hips. In the present study, forty-five of the fifty-seven hips needed bone-grafting for major acetabular deficiencies. In thirty-four hips, only morseled allograft or autogenous graft was used to fill cavitary defects or to reconstitute a deficient medial wall. Although none of these grafts were studied histologically, the radiographs showed apparent healing with filling of the defects. Heekin et al. studied three postmortem specimens from hips that had been treated with morseled allograft, and the grafts appeared to have been incorporated. A fresh-frozen bulk allograft was used in eleven hips in the present study to obtain mechanical stability of the acetabular component. Although there was partial lateral resorption in four hips, no graft collapsed and no component failed. Because of concern regarding late resorption and collapse of bulk allografts10, the senior one of us currently uses bulk allografts only rarely; instead, he uses a so-called jumbo acetabular component for a rim fit in revisions in hips with a large segmental deficiency.
The purpose of the present study was to reevaluate the results of this technique of acetabular revision, which were previously reported in a short-term study10, after a longer period of follow-up. It is difficult to report the clinical results for only one component of a revision total hip arthroplasty, as pain and functional limitations may be due to factors that are not related to the acetabular component alone. The results of the present study thus represent a worst-case scenario. However, despite the complex nature of these revisions, twenty hips (35 per cent) had an excellent clinical result and nineteen (33 per cent) had a good clinical result.
The results of acetabular revision with use of this specific porous-coated component have been extremely favorable17. Silverton et al. reported the results of 115 acetabular revisions performed by three surgeons. At a mean of eight years postoperatively, thirteen components had been revised but none had been revised because of aseptic loosening. The results of the present study support the use of a titanium fiber-metal-coated component fixed with multiple screws and no cement. No component migrated or had a continuous radiolucent line at the bone-implant interface, and there were no vascular complications associated with the screws. The senior one of us continues to use this technique and implant for all acetabular revisions, except those in patients who have had high-dose irradiation of the pelvis for the treatment of a malignant tumor. Until the results with other techniques and implants are reported after similar durations of follow-up, the technique of so-called line-to-line implantation of titanium fiber-metal-coated acetabular components with multiple-screw fixation should be considered the so-called gold standard for acetabular revision.