The patients were evaluated prospectively, and data were reviewed retrospectively at a minimum of ten years postoperatively. Institutional review board approval was obtained for the study. Clinical evaluations were performed with use of the Harris hip score2, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score3, and a questionnaire that included items regarding any noise during daily activity (the presence of noise; the nature, time of onset, and frequency of the noise as well as activities associated with the noise; awareness of the noise by other people; the intensity of the noise over time; associated pain; and the affect of the noise on the quality of life).
Radiographs were evaluated by two independent observers (Y.-K.L. and H.J.K.) with respect to component stability4, radiolucent lines1, calcar resorption5, osteolysis6, and loosening7. The zones described by Gruen et al.8 and those described by DeLee and Charnley9 were used to assess the location and extent of radiolucent lines and osteolysis. Osteolysis was defined as a periprosthetic cystic or scalloped lesion with a diameter exceeding 2 mm that had not been present on the immediate postoperative radiograph6,10. In addition, serial radiographs were reviewed for femoral neck fretting, which was interpreted as evidence of impingement of the metal neck on the ceramic liner11.
Kaplan-Meier survival analysis12 was performed with use of three end points: revision of any implant for any reason, a reoperation for any reason, and any reoperation or radiographic evidence of osteolysis or loosening.
Source of Funding
This study was funded internally by the Seoul National University Hospital Research Fund (grant number 06-03-063).
At a minimum of ten years postoperatively, of the original eighty-four patients, four (with six total hip arthroplasties) had died and six (with six total hip arthroplasties) had been lost to follow-up, which left eighty-eight hips in seventy-four patients. Of these seventy-four patients, sixty-eight (eighty-one hips) had both clinical and radiographic evaluations and six (seven hips) underwent only a clinical evaluation conducted by telephone and with a questionnaire sent by mail (Table I). The latest clinical and radiographic evaluations were performed at a mean of 131 months (range, 120 to 142 months) and 130 months (range, 120 to 142 months), respectively, after the operation. The study group included forty-five men and twenty-nine women. The mean age of the patients at the time of the index arthroplasty was forty-one years (range, eighteen to sixty-five years), and the mean body mass index was 23.6 kg/m2. The most common diagnosis, recorded for forty-four hips (50%), was osteonecrosis of the femoral head (Table II). A short-neck modular femoral head component was used in forty-seven hips; a medium-neck component, in thirty-two hips; and a long-neck component, in nine hips.
The mean preoperative Harris hip score for the entire series was 59 points (range, 30 to 84 points), and the mean postoperative score for the seventy-four patients followed for at least ten years was 96 points (range, 87 to 100 points). At the last follow-up evaluation, the mean WOMAC score was 12.9 points (range, 0 to 44 points). Twelve (14%) of the eighty-eight hips were associated with an intermittent clicking sound, and one hip had squeaking. None of the "clicking" hips were associated with pain or any modification of daily activities, and all clicking occurred when the patient was rising from a squatting position. With the numbers available, no significant associations were found between hip noise and patient-related or prosthesis-related factors, including age, sex, weight, height, body mass index, position of the implant, or neck length of the femoral component. A patient with bilateral total hip arthroplasty, who was twenty-six years old at the time of the index surgery, reported an audible squeak (several times per day) in the right hip, only during swaying to the side, that had begun seven years postoperatively. However, this noise was not reproducible during our outpatient evaluation. The remaining total hip prostheses were completely problem-free.
All of the eighty-one hips had radiographic evidence of bone ingrowth at the time of the most recent follow-up. Ten hips (11%) in the previous study1 had a radiolucent line only in Gruen zone8 3, 4, or 5. No additional hip was found to have developed a radiolucent line around any implant at the time of the longer follow-up. Furthermore, the radiolucent lines in these ten hips had not changed from their previous appearance. None of the radiolucent lines measured >2 mm in width. Rounding off of the sharp medial edge of the resected femoral neck was a frequent finding (Fig. 1), but no calcar resorption was detected in any hip. No hip demonstrated radiographic signs of loosening. Periprosthetic osteolysis was not detected around any cup or stem. Although no fracture of an alumina liner or notch-sensitive titanium stem occurred in this series, fretting of the femoral neck was observed on the translateral radiographs of six hips (7%) (Fig. 2). The mean interval from the index surgery to the appearance of fretting was 7.2 years (range, 4.7 to 10.3 years). All of the fretting was observed on the anterior portion of the neck. Only one neck had fretting before the five-year follow-up evaluation. With the numbers available, we found no significant associations between fretting and patient-related or prosthesis-related factors, including hip noise, age, sex, weight, height, body mass index, ability to sit in the so-called tailor position, duration that the prosthesis was in situ, position of the implant, or neck length of the femoral component.
Two ceramic femoral head fractures occurred in this series. Previously, we reported on a fracture sustained by a patient who was in a motor-vehicle collision1 and, since that report, an additional ceramic femoral head fracture occurred. That fracture occurred without trauma, eleven years and three months after the surgery. The patient reported a crunching sensation when rising from the toilet, and two days later radiographs confirmed the fracture. The fractured alumina femoral component had a 28-mm short neck, which was reported by Koo et al. to be prone to fracture13. In this patient, we replaced the bearing surface with a cobalt-chromium femoral head and polyethylene insert and left the well-fixed stem and cup in place, as was done in our other patient1. At the time of writing, both patients remained under close observation for wear of the polyethylene liner14.
During the period between the previous and the current report, an additional patient sustained a periprosthetic fracture, after falling from a height of 3 m. The fracture occurred around a well-fixed stem at six years and two months postoperatively. It was treated with open reduction and internal fixation with use of cable and wire, and subsequently it completely healed without an additional procedure.
Since the original report, only one hip has dislocated. This dislocation, of the right hip, occurred seven years and three months postoperatively in a patient who had undergone bilateral total hip arthroplasty. An open reduction was performed; the bearing was not exchanged, and the dislocation did not recur.
Overall, there were only four reoperations during the follow-up period. These included two implant revisions, each for a fracture of an alumina femoral head; one procedure for the treatment of a periprosthetic fracture; and one for the treatment of a dislocation.
Kaplan-Meier survivorship analysis based on eighty-eight hips at risk, with any implant revision for any reason as the end point, revealed a cumulative survival rate of 99.0% (95% confidence interval, 97.0% to 100%) at ten years (Fig. 3). With a reoperation for any reason as the end point, the ten-year survival rate was 96.9% (95% confidence interval, 93.4% to 100%). Survivorship analysis revealed that, at ten years, 96.9% of the hips were not associated with any reoperation or radiographic evidence of osteolysis or loosening.
The third-generation alumina-on-alumina bearings used for primary cementless total hip arthroplasty in this study were found to produce excellent clinical results and implant survival rates, with no detectable osteolysis at a minimum of ten years postoperatively. On the basis of these results, we believe that improved alumina-on-alumina bearing implants are a reasonable option for young, active patients, and we continue to use ceramic-on-ceramic bearing couples for the majority of primary total hip arthroplasties in this patient group. However, our findings suggest that surgeons should be aware of the potential risks of ceramic fracture, noise, and impingement between the metal neck and the ceramic liner when using an alumina-on-alumina articulation; accordingly, patients should be informed of these risks. Additional, longer-term follow-up is necessary to assess the potential adverse effects of these concerns on the longevity of this total hip prosthesis.