Wear and aseptic loosening of the acetabular component are the most common long-term problems after total hip arthroplasty performed with use of cement24. The results of primary arthroplasty with cement have been shown to be less successful in younger patients than in older patients1,2,5,6,22, although two of us (B. M. W. and P. D. S.) previously suggested that excessive pessimism about this group of patients was unfounded23.
There have been few reports about the results of revision arthroplasty in young patients. Stromberg et al. reported the results in patients who were less than fifty-five years old. We now report the results of revision of the socket with use of cement in eighty-two young patients (eighty-seven hips) without rheumatoid arthritis who were followed prospectively.
*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. Funds were received in total or partial support of the research or clinical study presented in this article. The funding sources were The Peter Kershaw Trust and The Charnley Trust.
†Royal Orthopaedic Hospital, Northfield, Birmingham B31 2AP, United Kingdom.
‡The John Charnley Research Institute, Wrightington Hospital, Hall Lane, Wigan WN6 9EP, United Kingdom. Please address requests for reprints to Professor Wroblewski.
The criteria for inclusion in the study were an age of fifty-five years or less at the time of the revision, aseptic loosening as the indication for the revision, and a minimum duration of follow-up of two years. One patient who had a repeat revision less than two years after the index revision was also included. Patients who had a primary diagnosis of rheumatoid arthritis or an infected hip were excluded. No patient was lost to follow-up, and no patient died during the study period.
Eighty-seven hips in eighty-two patients (fifty-one women and thirty-one men) were evaluated. Sixty-six patients were referred to us from other centers. The mean age at the time of the index revision was forty-four years and seven months (range, twenty-three years to fifty-five years and eleven months). The mean weight was sixty-eight kilograms. The mean duration of follow-up was approximately six years (range, 1.9 to 18.1 years). The most common underlying condition was primary osteoarthrosis (thirty-nine hips), followed by congenital dysplasia (twenty-four hips), post-traumatic osteoarthrosis (eight hips), avascular necrosis (six hips), slipped proximal femoral epiphysis (three hips), ankylosing spondylitis (three hips), Legg-Perthes disease (two hips), and fracture of the femoral neck (two hips).
The clinical assessment was performed with the method of Merle d'Aubigné and Postel as modified by Charnley. In addition, the patients were asked to grade the result of the arthroplasty subjectively as excellent, satisfactory, or unsatisfactory.
The quality of the acetabular bone stock before the revision was graded according to the criteria of Gustilo and Pasternak. When the quality was reasonably good (the acetabulum was only slightly enlarged or there were local defects involving only one wall), the hip was given a grade10 of I, II, or III and was classified as type 1. When the quality was poor (there was massive or global collapse of the acetabulum and there were defects involving at least two walls), the hip was assigned a grade10 of IV and was classified as type 2. The bone stock was graded on the basis of the radiograph made before the revision and the operative findings. Twenty-three hips were classified as grade I; ten, as grade II; seventeen, as grade III; and thirty-seven, as grade IV; thus, the bone stock before the revision was type 1 in fifty hips and type 2 in thirty-seven. Seventy-three hips had a revision of both components and fourteen, of the socket only.
Serial anteroposterior radiographs of the pelvis, made after the revision (at six weeks, three months, six months, one year, and two years and every two years thereafter), were analyzed by one of us (V. V. R.). The appearance of the bone-cement interface of the socket was recorded as suggested by Hodgkinson et al. Our radiographic criteria for loosening of the socket were very strict: any socket that had migrated any distance or that had a continuous line of demarcation in all three zones as described by DeLee and Charnley, with a demarcation gap that was more than one millimeter thick in at least one zone, was classified as radiographically loose and therefore as a failure, regardless of the patient's clinical status. Vertical migration was determined by measuring the change in the vertical distance of the cup from the teardrop line; horizontal migration was measured with reference to a perpendicular to the teardrop line through the teardrop.
Operative Technique
All operations were performed by the senior one of us (B. M. W.) in a Charnley-Howarth enclosure (Howarth Engineering, Brinscall Blackburn, United Kingdom) with use of a body-exhaust suit. A transtrochanteric approach was used. The loose socket was removed, and the femoral stem was checked for loosening or malpositioning. The acetabular floor was thoroughly curetted of any membrane lining it. Any defects in any of the four walls (anterior, posterior, superior, or central10) of the acetabulum were covered with wire mesh. No bone grafts or reinforcement rings were used in any of the procedures. If the quality of the bone stock permitted it, multiple small drill-holes were made in the acetabular walls. An acetabular cement pressurizer (Howmedica International, London, United Kingdom) was used for fifty hips treated since 1985; of these fifty, twenty-nine had type-1 bone stock, compared with twenty-one of the thirty-seven in which the pressurizer was not used. A flanged socket was used, and the aim was to seat the socket on the rim of the acetabulum.
The patients started partial weight-bearing with crutches at about two weeks postoperatively and progressed to full weight-bearing at three months. Hydroxychloroquine (Plaquenil; Winthrop, Guildford, United Kingdom), an antiplatelet agent, was given as prophylaxis against venous thromboembolism.
Complications
Intraoperative complications consisted of a minor fracture of the medial portion of the femoral neck in two patients and fragmentation of the trochanter in two patients. Postoperative complications included trochanteric non-union in ten hips, pulmonary embolism (nonfatal) in six patients, and dislocation in four hips.
Clinical Results
At the latest follow-up evaluation, fifty-eight hips (67 per cent) were classified by the patients as having an excellent result and twenty-one (24 per cent), as having a satisfactory result. The remaining eight hips were classified by the patients as having an unsatisfactory result and were regarded as clinical failures; these included the hips that had subsequently had a repeat revision. Two of the unsatisfactory results were due to loosening of the stem that led to additional revision.
The mean hip score18 for pain improved from 3.3 points (range, 2 to 6 points) before the revision to 5.5 points (range, 2 to 6 points) at the latest follow-up evaluation; that for function, from 3.1 points (range, 1 to 6 points) to 4.8 points (range, 2 to 6 points); and that for motion, from 4.0 points (range, 2 to 6 points) to 5.1 points (range, 3 to 6 points). In one patient, the status of the hip at the latest evaluation was worse than it had been before the revision.
Radiographic Results
Radiographically, sixty-one sockets (70 per cent) were classified as stable and twenty-six (30 per cent), as loose12. Of the twenty-six loose sockets, eleven had a continuous line of demarcation in all three zones with a demarcation gap that was more than one millimeter thick in at least one zone4, and fifteen sockets had migrated. Four of the twenty-six sockets had had a line of demarcation in zone 1 at one year; two, in zone 2; and sixteen, in zone 3; the remaining four sockets had migrated.
Of the fifty hips with type-1 acetabular bone stock, which were followed for a mean of 6.5 years (range, two to 18.1 years), seven (14 per cent) had loosening of the socket at a mean of 7.2 years. Of the thirty-seven hips with type-2 bone stock, which were followed for a mean of 5.8 years (range, 1.9 to 13.4 years), nineteen (51 per cent) had loosening of the socket at a mean of 6.1 years. The relationship between loosening and the quality of the acetabular bone stock before the revision was highly significant (p = 0.0002, chi-square test).
Of the fifty hips for which a cement pressurizer had been used during the operation, twelve had loosening of the socket, compared with fourteen of the thirty-seven for which a cement pressurizer had not been used.
In the fourteen hips that had had a revision of the socket only, all of the stems were radiographically stable and four of the sockets were loose at the latest follow-up evaluation. In this group, there was one unsatisfactory clinical result (in a patient who had a repeat revision because of recurrent dislocation).
None of the sockets that had not had a line of demarcation in any zone at one year had radiographic evidence of loosening at the latest follow-up evaluation (mean, five years; range, three to ten years).
Repeat Revisions
Nine hips had a repeat revision: five, because of aseptic loosening of the socket (which was asymptomatic in one), and one each, because of aseptic loosening of the socket and the stem, aseptic loosening of the stem, recurrent dislocation, and unexplained pain. With use of repeat revision for any reason as the end point, the cumulative rate of survival14 of the prosthesis was 90 per cent (95 per cent confidence interval, 8.8 per cent) at six years. With use of radiographic loosening as the end point, the cumulative rate of survival was 75 per cent (95 per cent confidence interval, 13.8 per cent) at six years. After six years, the number of patients in the life table was less than forty; thus, interpretation of the data is of doubtful importance16.
The high rates of loosening, especially of the socket, that have been reported after primary arthroplasty with use of early cementing techniques in young patients2,5 has encouraged the insertion of prostheses without cement in these patients. However, problems that have been reported in association with bone-ingrowth prostheses at an early stage appear to make it a far from satisfactory option. Hernandez et al. found a higher rate of linear wear with use of porous-coated acetabular cups inserted without cement than with use of all-polyethylene cups inserted with cement (0.22 compared with 0.08 millimeter per year). Owen et al. reported disappointing results with use of porous-coated anatomic primary hip replacements inserted without cement: of ninety-nine implants that had survived in situ for more than five years, 36 per cent were associated with acetabular osteolysis and 12 per cent, with femoral osteolysis. In the total series of 241 porous-coated anatomic hip replacements inserted without cement, twenty-six acetabular components (11 per cent) failed at a mean of five years and the cumulative rate of survival was only 73 per cent at seven years. Owen et al. attributed these findings to increased polyethylene wear. Engh et al. reported a 41 per cent prevalence of osteolysis in association with seventy-four porous-coated total hip replacements. Kim and Kim, in a study of the results of fifty consecutive primary total hip arthroplasties with use of the anatomic medullary locking (AML) prosthesis (DePuy, Warsaw, Indiana) without cement, found a 32 per cent prevalence of acetabular osteolysis (sixteen hips) and a 58 per cent prevalence of femoral osteolysis (twenty-nine hips) at a mean of seven years after the operation. The osteolysis occurred only in young patients.
We know of only one previous report on the results of revision operations in young patients. Stromberg et al., in their series of sixty-seven revision arthroplasties performed with use of cement in patients younger than fifty-five years, found a 15 per cent prevalence of loosening of the acetabular component at a mean of four years. In the current study, the rate of loosening of the socket in type-1 acetabula (those with reasonably good bone stock) was 14 per cent (seven of fifty hips) at a mean of 6.5 years, while in type-2 acetabula (those with poor bone stock) the rate of loosening was 51 per cent (nineteen of thirty-seven hips) at a mean of 5.8 years (p = 0.0002). As mentioned earlier, our criteria for loosening were very strict: even a slight change in position led to the socket being classified as loose. In many other studies, migration of five millimeters or more has been the definition of loosening9,20,21. The significant relationship between the rate of loosening and the quality of the acetabular bone stock in our series points to the importance of regular follow-up evaluation of all young patients so that loose components can be revised early, before excessive deterioration of the bone stock. The fact that almost all (twenty-one; 81 per cent) of our twenty-six patients who had a socket that was radiographically loose either were completely asymptomatic or had only slight symptoms also underscores the importance of regular follow-up. This finding is in keeping with those of Charnley, Engelbrecht et al., and two of us (B. M. W. and P. D. S.)24. We advise more frequent follow-up of asymptomatic patients who have a radiographically loose socket but no evidence of progressive bone loss. Ideally, the revision should be performed before the acetabular bone stock has deteriorated to type 2.
The prevalence of pulmonary embolism in the present series was much higher than that encountered by Stromberg et al.
The findings on the anteroposterior radiograph of the pelvis that was made one year postoperatively had a high predictive value with regard to the future success of fixation of the socket. This was demonstrated by the fact that the seventeen sockets that had no line of demarcation in any zone at one year were radiographically stable at the time of the latest follow-up.
Thus, the use of cement did not yield good results in young patients who had poor acetabular bone stock at the time of revision of the socket. The use of bone grafts to try to reconstitute some of the lost bone may improve the results17,20. However, the results of revision of the socket with use of massive allografts in patients who have very poor acetabular bone stock have not been very encouraging9,13.