To date, the optimal surgical treatment for young patients with end-stage osteoarthritis remains controversial. Hip arthrodesis and resection arthroplasty are considered salvage procedures1, and hip arthroplasty seems to be a more attractive option for the treatment of painful, end-stage osteoarthritis in patients thirty years of age or younger2-13. However, evaluation of total hip arthroplasty in this population reveals a high complication rate, with frequent loosening and limited component durability, mainly because of high conventional polyethylene wear resulting from demanding and active lifestyles14. To improve implant longevity15, the introduction of hard-on-hard bearings with excellent tribological properties seems to be an attractive option for this group and should be considered16-18. Ceramic-on-ceramic bearing surfaces with properties such as chemical inertness may have an advantage over metal-on-metal bearings. However, ceramic-on-ceramic-bearing cup fixation, femoral head or insert fracture, and squeaking remain concerns in young, active patients19-21. Favorable findings have been reported for metal-on-metal bearing components, but the majority of studies have not focused on these younger patients22.
The aim of the current work was to assess the data on second-generation metal-on-metal total hip arthroplasty performed as a primary procedure in patients thirty years of age or younger and to evaluate the survival rate.
Patients
Between 1995 and 2003, forty-eight total hip arthroplasties with metal-on-metal bearings were performed consecutively in thirty-five patients who were thirty years of age or younger; thirteen patients had a bilateral procedure. This retrospective study was undertaken after ethics approval and informed patient consent had been obtained. The use of metal-on-metal bearings was justified by the young age of the patients and the high activity level anticipated after hip arthroplasty.
The study group comprised twenty-three male patients and twelve female patients. The mean age at the time of surgery was twenty-five years (range, fifteen to thirty years), and the mean body mass index was 25.2 kg/m2 (range, 16.3 to 38.6 kg/m2). One patient with steroid-related femoral head osteonecrosis died from leukemia after one year of follow-up. The remaining thirty-four patients (forty-seven hips) were retrospectively reviewed after a minimum duration of follow-up of sixty months (mean, 108 months; range, 62.4 to 153.6 months) (Table I).
The preoperative diagnosis was femoral head osteonecrosis for twenty-six hips, sequelae of pediatric hip disease for seventeen, inflammatory disease for two (ankylosing spondylitis for one and juvenile idiopathic arthritis for one), and posttraumatic osteoarthritis for three. Of the seventeen patients with sequelae of pediatric hip disease, eight had developmental hip dysplasia, five had slipped capital femoral epiphysis, and four had osteochondritis dissecans.
Before total hip arthroplasty, thirteen hips (thirteen patients) had had at least one previous hip operation; the previous procedures included two Chiari pelvic osteotomies, three hip shelf arthroplasties, two internal fixation procedures for the treatment of fractures, and six core decompression procedures.
According to the Devane scoring system, the majority of patients were active. Three patients (five hips) were rated as grade IV (light job, noncontact sports), sixteen patients (twenty-four hips) were rated as grade III (leisure activities), fifteen patients (eighteen hips) were rated as grade II (semi-sedentary), and one patient (one hip) was rated as grade I (sedentary).
According to the Charnley classification system23, fourteen patients (eighteen hips) were rated as A (one hip involved), twenty patients (twenty-eight hips) were rated as B (both hips involved), and one patient (both hips) was rated as C (multi-articular disease).
Surgical Procedure
All prostheses were inserted through a posterolateral approach in an operating room with vertical laminar air-flow. No patient received medications to prevent heterotopic ossification, but all underwent extensive intraoperative lavage with sterile saline solution.
Thirty-two patients (forty-five hips) received an uncemented stem (Alloclassic Zweymüller SL stem; Zimmer, Winterthur, Switzerland) (Figs. 1-A and 1-B), and three patients (three hips) received a cemented stem (Müller self-locking stem; Zimmer). With regard to the acetabular component, the eighteen patients (twenty-two hips) who were managed before 2000 received an Armor impacted cup (coated with titanium mesh and no hydroxyapatite) (Zimmer) and the seventeen patients (twenty-six hips) who were managed after 2000 received the Allofit cup (grit-blasted surface) (Zimmer).
The head diameter was 28 mm in all cases but five, in which a 32-mm head was used. A Metasul (Zimmer) wrought, high-carbon-content cobalt-chromium-alloy (Co-28Cr-6Mo, 0.2% to 0.25% carbon) femoral head and a modular Metasul polyethylene sandwich-type acetabular insert were implanted as bearing surfaces in all cases (Fig. 2). Surface roughness was <0.005 µm, sphericity deviation was <10 µm, and radial clearance was approximately 45 µm. The mean external cup diameter was 54 mm (range, 48 to 62 mm).
Full weight-bearing was allowed on the day after surgery, except in the cases of nine patients in whom bulk autografting had been performed to reconstruct the acetabulum; for these nine hips, weight-bearing was introduced at six weeks postoperatively.
Clinical and Radiographic Evaluation
All patients were followed annually and then were reviewed by an observer who did not participate in the surgery (D.B.). One patient was lost to follow-up, and another died one year after the surgical procedure from preexisting leukemia.
Clinical outcome was assessed with use of the Merle d'Aubigné scale24 and the Devane activity level25.
Radiographic outcome was recorded on the basis of anteroposterior radiographs of the pelvis and lateral radiographs of the affected hip or hips. Radiographic examinations were performed annually for all patients and, at the time of the latest follow-up, radiographs were evaluated by two observers who did not participate in the surgery (D.B. and J.G.).
Femoral stem migration was ascertained by measuring the distance between the top of the proximal aspect of the femoral stem (the lateral spike of the SL stem or the lateral shoulder of the Müller stem) and the top of the greater trochanter.
Likewise, cup migration was detected by quantifying variations in the position of the articulation center with respect to the acetabular teardrops. The following radiographic signs were considered to indicate acetabular loosening: component migration of >4 mm in either the horizontal or vertical direction, breakage of fixation screws, changes in component orientation by >5°, or radiolucent lines measuring >2 mm wide in all zones.
Cementless component fixation was assessed according to the femoral zones of Gruen et al.26 and the three acetabular zones of DeLee and Charnley27 by searching for osteolysis, spot welds, bone condensation, or reactive lines. Heterotopic ossification was graded according to the system of Brooker et al28.
For cemented stems, prosthetic alignment and subsidence were measured directly from calibrated radiographs. Radiolucent lines measuring >1 mm wide and >5 mm long at the stem-cement and cement-bone interfaces were evaluated in the Gruen femoral zones.
Statistical Analysis
Statistical analysis was conducted with SPSS version 13.0 for Windows (SPSS, Chicago, Illinois). Significance of the findings was evaluated with use of the paired t test to compare paired variables. The level of significance was set at p < 0.05. Kaplan-Meier survival analysis was used to estimate the probability of retention of the total hip replacement in relation to revision for any reason and included the entire cohort. Ninety-five percent confidence intervals were calculated.
Source of Funding
No external funding was received for this study.
The mean duration of follow-up was 108 months (range, 62.4 to 153.6 months). One patient was lost to follow-up, and another patient died after one year of follow-up.
The mean Merle d'Aubigné24 score was 10.6 (range, 1 to 14) preoperatively and increased to 17.1 (range, 12 to 18) at the time of follow-up. The Merle d'Aubigné score was excellent (18 points) for twenty-four patients (71%), good (16 or 17 points) for seven (21%), fair (15 points) for two (6%), and poor (<15) for one (3%).
At the time of the latest follow-up, the patient activity level rose according to the Devane25 scoring system relative to preoperative values (p < 0.01): ten patients (sixteen hips) were rated as grade V (strenuous job or contact sports), seventeen patients (twenty-two hips) were rated as grade IV (light job, noncontact sports), four patients (four hips) were rated as grade III (leisure activities), and four patients (six hips) were rated as grade II (semi-sedentary). According to the Charnley23 classification system, eleven patients (eleven hips) were classified as A, nine patients (twelve hips) were classified as B1, thirteen patients (twenty-one hips) were classified as B2, and two patients (four hips) were classified as C.
The clinical results for the nine hips that underwent bulk acetabular autografting were not different, except that six weeks were required before full weight-bearing was introduced. For these nine hips, features of graft union with the pelvis were identified on radiographs at the time of follow-up without signs of cup loosening.
All hips that had a cementless femoral stem had at least one sign indicating osseous fixation (Fig. 1-B): forty hips had spot welds in zone 2, 3, 5, or 6; twenty-one hips had bone condensation in zone 2 or 3; and five hips had calcar atrophy. In contrast, two hips with a cementless stem had partial radiolucencies limited to the proximal aspect of zones 1, 2, and 7. Among the three hips with a cemented stem, one had isolated and partial radiolucencies that involved Gruen26 zones 1 and 7, and one had nonprogressive radiolucency in zones 1, 2, 3, and 7.
Femoral osteolysis, strictly limited to zones 1, 2, and 7, without stem-loosening or migration, was noted in three hips. None of the femoral components showed migration. The mean cup inclination was 45.1° (range, 26° to 54°). On the acetabular side, three hips had radiolucent lines limited to zone 1, one had osteolysis limited to zone 2, and one had osteolysis limited to zones 1 and 2. The cup was considered to be stable in all hips except for one, which had acetabular osteolysis that occurred in two zones and was treated with a reoperation. Only fourteen patients (fourteen hips) had ossifications that were limited to Brooker28 grade 1 (twelve patients, twelve hips) or 2 (two patients, two hips).
There were no intraoperative complications, such as cracks of the proximal part of the femur. Postoperatively, two patients who had received corticosteroids for the treatment of the initial disease had infections that were treated successfully with lavage and antibiotics, one had a deep venous thrombosis, and three had transient sciatic nerve palsy. We were unable to determine the etiology of these neurological complications as there was no leg-length discrepancy, no hematoma, and no lumbar spine problem. No patient reported thigh pain. One dislocation occurred three months postoperatively and was treated successfully with closed reduction, without recurrence. Two revisions were done, one because of painful impingement secondary to cup malpositioning and the other because of acetabular osteolysis. In these two patients, only the cup was removed. In the case of the first patient, the cam effect was detected by means of clinical evaluation, which revealed abrupt and noisy stoppage of movement during internal rotation combined with flexion. In this patient, the cup (Allofit cup with 32-mm Metasul bearing) was reoriented twenty-two months after the initial procedure, and, at the time of the last review, the cam effect had disappeared. For the patient with aseptic acetabular osteolysis and loosening, metallic ion levels in whole blood samples were considered to be normal (chromium, 1.4 µg/L; cobalt, 0.9 µg/L). During the cup revision procedure at six years of follow-up, no metallosis or cam effect was observed, and the stem was left in place. Five years after the reoperation, the clinical result was considered to be excellent (Merle d'Aubigné score, 18).
Kaplan-Meier analysis with revision of either component as the end point revealed a ten-year survival rate of 94.5% (95% confidence interval, 80% to 98.6%) (Fig. 3). The survival rate was 100% for the femoral stem. We noted no significant differences in the survival rate according to component design (cementless or cemented stem, Allofit or Armor cup, etc.), preoperative diagnosis, or previous hip surgery.
Young patients with painful hip disorders seek total hip arthroplasty to relieve the pain and to return to an active lifestyle29. In the Swedish Registry, the fifteen-year prosthetic survival rate following total hip arthroplasty was 77% for patients below the age forty-five years and 92% for those above the age of sixty-five years30. Total hip arthroplasty in young patients remains a procedure that is associated with the risks of high conventional polyethylene wear and osteolysis2,31. Hard-on-hard-bearing components appear to decrease the risk of wear32.
We identified eleven previous studies2-7,9,11-13,22 with the same inclusion criteria: performance of total hip arthroplasty in patients thirty years of age or younger, a duration of follow-up more than five years, and inclusion of more than forty patients (Table II). Of those eleven investigations, only five2,3,11-13 included variable preoperative diagnoses (not only juvenile idiopathic arthritis), as in our work. The implants were cemented in all studies except the one by Odent et al.12. In the current investigation, the revision rate of 4.3% was lower than those in all others (mean, 34%; range, 14% to 67%). In ten of the eleven studies cited, the revision rate was significantly correlated with high activity level6, unilateral disease7, the number of preoperative hip operations, weight (>60 kg)7,8, and etiology (developmental hip dysplasia9, trauma7, osteonecrosis6, juvenile idiopathic arthritis9,11). All of those series highlighted the greater risk of revision compared with that in older and/or less active populations. The revision rate was correlated by all authors with aseptic loosening secondary to conventional polyethylene wear. According to Torchia et al.7, the high revision rate was a relative contraindication to total hip arthroplasty in young, active, and heavy patients (>60 kg) with unilateral osteoarthritis. The revision rate seemed to be directly correlated with etiology9,11. In our series, no primary osteoarthritis was observed. This could have led to a worse survival rate13 because secondary osteoarthritis represents more challenges, including abnormal morphology and comorbidities13.
In our study, the Merle d'Aubigné score increased from 10.6 to 17.1. The clinical result was excellent for twenty-four patients (69%). Odent et al.12 reported a 90% rate of excellent clinical results after thirteen years of follow-up, and Torchia et al.7 reported a 75% rate after twelve years of follow-up.
In previous studies, complications included fracture (prevalence, 1.7% in a study involving cementless stems), neurological problems (prevalence, 4% to 41%, with higher risk being associated with developmental hip dysplasia11,12), dislocation (prevalence, 0% to 18%3,4,7,13), infection (prevalence, 3% to 7% with higher risk being associated with the use of anti-inflammatory steroids or immunosuppressants4,7,12,13).
Currently, hip resurfacing with a metal-on-metal bearing has become a suitable option for young, active patients. The use of a large-diameter head, reduced dislocation risk by the suction-fit effect33, and lower wear according to bearing type are major advantages. Moreover, acetabular and femoral bone stock preservation after hip resurfacing should be considered as an attractive goal in case further revision is necessary34.
Concerns linger about elevated blood chromium and cobalt levels in patients with metal-on-metal bearings. Theoretically, there are still potential risks of mutagenicity and hypersensitivity35. Carcinogenic risk does not seem to be increased in subjects receiving metal-on-metal hip prostheses as compared with those receiving metal-on-polyethylene hip prostheses (relative risk for cancer, 0.95% for metal-on-metal prostheses and 0.76% for metal-on-polyethylene prostheses; p > 0.05)36 or the general population37.
The frequency of pseudotumors after the insertion of metal-on-metal implants appears to be more common in association with large-diameter as compared with small-diameter heads (28 or 32 mm). To our knowledge, there has been only one case report in the literature on pseudotumor after the insertion of a 28-mm metal-on-metal head38. The risk of pseudotumor is estimated to be 1% for patients who have metal-on-metal resurfacing within five years after implantation39. Some risk factors, e.g., female sex, an age of less than forty years, smaller component size, and edge loading with a vertical cup, are now well known40.
Teratogenicity follows exposure to various metals35, but, to date, no reports of mutagenic/teratogenic effects in patients with metal-on-metal total hip arthroplasty have been published, to our knowledge. Cobalt and chromium are able to cross the placenta, which modulates the rate of metal ion transfer41. Hence, the possibility of teratogenic and chromosomal changes in developing fetuses is still unclear. In our institution, we use metal-on-metal bearings in women of child-bearing age after informing them about these potential risks and after obtaining their informed consent.
Revision because of acetabular loosening and osteolysis is related, at the same time, to backside wear at the polyethylene liner-cup interface and cup mobility that could increase wear debris. With a metal-polyethylene sandwich, there is still a possibility of wear and osteolysis resulting from the backside of the liner42. With the cups used in this trial, backside wear debris could pass through screw holes and could lead to acetabular osteolysis.
Patients who have a reoperation because of cam impingement underscore the necessity of perfect cup and stem positioning. In one patient in the present study, inadequate acetabular preparation and retroverted cup positioning associated with wide range of motion created a cam effect. If a cam effect occurs in association with a metal-on-metal bearing, the risk of elevated blood metal levels can lead to recommendations for revision surgery43. A cam effect between the stem and the liner could increase titanium ion release from the neck of the stem, but few research studies have addressed the biological outcomes of elevated titanium ion levels in the human body44. The very low immunogenicity of titanium leads us to consider it a nonproblematic concern45.
The present study had some limitations. It was a retrospective case series with no case controls and with only intermediate-term follow-up. Two different cups and stems were used. No routine metal ion measurements were performed. However, the metal-on-metal bearing that was used was always the same. The two cup designs had the same insert, the same mode of fixation in the metallic shell, and the same opening angle. They were made of the same titanium alloy and differed only in porous coating, which was not the source of a specific adverse effect or failure of both models.
The encouraging intermediate-term results in our patients indicate that total hip arthroplasty with metal-on-metal bearing components may be a suitable solution for young patients with painful hip osteoarthritis. It is necessary to consider, in this population, that the first total hip arthroplasty should facilitate eventual revision for successful long-term outcomes. Thus, hard-on-hard bearings seem to enhance the long-term outcomes due to excellent tribological performance as well as very low osteolysis and wear rates.
In conclusion, with satisfactory intermediate-term results, metal-on-metal bearings appear to be a suitable alternative to metal-on-polyethylene bearings in patients thirty years of age or younger, but long-term outcomes are needed to further confirm these data.