Abstract
Large-head metal-on-metal (MoM) total hip arthroplasties were introduced because of their purported advantages. Inflammatory pseudotumours occasionally occur after MoM hip arthroplasty and often lead to revision. The purpose of this study was to assess the outcome of revision of large-head MoM total hip arthroplasties after an extensive screening protocol for all MoM articulations with a minimum 2 year follow-up. We identified 50 hips that had undergone large-head MoM total hip arthoplasty and required revision at a mean of 44 months after index operation. Of these, 38 were revised for pseudotumours, 7 for loosening, 2 for infection and 3 for instability. There was bone loss in 44 hips. The majority were revised to a 28 mm metal or ceramic head on a polyethylene-cemented cup. In 12 patients there was a complication. There was a decrease of whole blood cobalt from 20.8 µg/L preoperatively to 1.8 µg/L 1 year after revision. We found 6 residual masses on routine postoperative CT scans. The indication for revision of this cohort is inevitable but the clinical outcome 2 years after revision for pseudotumour is disappointing. Revising a MoM hip arthroplasty to a conventional cemented polyethylene or dual-mobility cup with bone impaction grafting and a 28 mm head can adequately treat the high ion levels and probably the disease of ARMD.
Introduction
Large head metal-on-metal (MoM) total hip arthroplasties were introduced because of their purported advantages over conventional metal-on-polyethylene bearings. Despite good clinical results and survival at medium-term follow-up (1), serious complications are now increasingly reported (2-3-4). It is assumed that generated metal debris can induce a local tissue response generally referred to as ARMD (adverse reaction to metal debris). Such a local tissue reaction can give rise to a periarticular mass known as a pseudotumour. Pseudotumours are of specific concern because they can give rise to swelling, pain, local pressure effects, subluxation, bone and soft-tissue destruction (4-5-6). Systemic effects of metal ions released from local debris have also been reported, although these appear to be infrequent (7). Compared to hip resurfacing, where metal debris is believed to arise mainly from the articulation, stemmed modular designs have an additional mechanism of failure. The different taper contact areas are susceptible to corrosion and wear, generating an additional source of metal debris and ion release. Revision of the MoM hip arthroplasty is associated with poor clinical results and a high complication rate (8-9-10-11-12-13), and a high rate of recurrence of the pseudotumour has also been reported in some studies (10, 14).
In this study we report the rate of radiological pseudotumour recurrence and the clinical results of a cohort of 50 revisions of single-type stemmed modular large head metal-on-metal hip arthroplasties revised to a 28 mm femoral head.
Patients and Methods
From January 2005 to July 2010 a Bi-Metric porous-coated uncemented stem with a metal-on-metal M2a-Magnum femoral head and ReCap acetabular component (Biomet, Warsaw, Indiana) was used in our facility, mainly in young active patients. The head and acetabular components are made up of cobalt-chrome-molybdenum, high-carbon, as-cast (single-heated components). The system is modular with increasing head size and concomitant larger shell size and the option to adapt the neck length using different length taper adapters. The stem and taper adapter are made of a titanium-aluminum-vanadium alloy.
An extensive screening protocol for all MoM articulations was initiated starting July 2010 (4). This screening protocol consisted of a questionnaire about symptoms in the groin, buttock, thigh and leg such as pain, swelling, discomfort, numbness and sensations of subluxation and clicking. Patients were questioned about their use of vitamin supplements and possible history of allergies. Each patient underwent CT evaluation of the pelvis and knee to detect periarticular masses and measure anteversion of the femoral component relative to the retrocondylar axis of the knee and anteversion of the acetabular component. All patients underwent pelvic (pubis-centred) and hip (both hip- and axially-centred) radiographs. Laboratory analysis with atomic mass absorption spectrometry of whole blood levels of cobalt and chromium ions was performed preoperatively at the time of CT evaluation and repeated annually. Blood samples were collected according to accepted guidelines (15). Blood samples were collected in metal-free vacutainers and taken using a standardised technique to minimise the risk of contamination: the first 5 mL blood was discarded to eliminate metal contamination. The metal ion levels in whole blood were determined using an atomic mass absorption spectrometry. The accuracy and precision of the method was validated with other laboratories using either inductive-couple plasma spectrometer (ICP-MS) or atomic mass spectrometer. Serum levels of leucocytes, ESR, CRP, creatinine and urea were also analysed to screen for infection and renal impairment. Only patients who underwent this prospective extensive screening protocol were included in this study. We defined a pseudotumour as a (semi)-solid or cystic periprosthetic soft-tissue mass with a minimum diameter of 2 cm that could not be attributed to infection, malignancy, bursa or scar tissue (4). Patients received a preoperative CT-scan which was repeated one year after surgery.
Revision surgery was performed by two experienced orthopaedic surgeons in our hospital. All revisions were conducted by a posterior approach. When reactive tissue or pseudotumour was present, it was radically resected but to a degree that was considered sensible without compromising neurovascular structures (mainly on the anterior side) and abductor muscles. Bone impaction grafting of the acetabulum was used in all procedures. During the operation, findings such as metallosis, aspect and location of the pseudotumour, corrosion of the male taper, abductor damage and acetabular defects were recorded. A 2-stage revision procedure was performed in a preoperatively suspected infection.
In the postoperative period after the index revision surgery, x-rays and metal ion analysis were prospectively collected at 3, 6, 12 and 24 months and a CT scan was performed 1 year postoperatively in all cases to detect a recurrence. We also collected the postoperative functional scoring (HOOS) for clinical outcome and a patient-reported survey of patient health (SF-36) after 1 and 2 years.
We defined a complication as; all adverse events in the observation period, including readmission within the observation period.
We calculated the decline of metal ion levels. Values obtained post-revision were normalised to a percentage of the initial, individual pre-revision value of the unilaterally revised hips. The rate of decline was calculated by dividing the percentage of the initial value by the number of days between date of revision surgery and date of repeat blood draw for ion levels.
We obtained approval for the collection and analysis of our data from the regional ethics committee.
Results
From January 2005 to July 2010 a total of 723 MoM total hip arthroplasties were implanted in 643 patients. Only 15 arthroplasties were revised in the first 5 years (prior to our screening protocol). Because no preoperative CT-scan was performed, these 15 patients were excluded from the analysis. To date 79 hips (10.9%) have been revised. Figure 1 shows the revision free survival in years. In the current analysis we included the 50 revisions that have reached a minimum 2-year follow-up and thus were evaluated by our prospective extensive screening protocol. Median survivorship after implantation of the primary prosthesis was 44 (12-78) months. Data were analysed at a mean follow-up time of 37.3 (25-56) months after the revision procedure. Patient characteristics and operation characteristic are shown in Table I.
Revision free survival in years.
Patient characteristics
#Used components: Cemented polyethylene FAL cup (Waldemar link), Cemented SPII stem (Waldemar link), MP reconstruction hip stem (Waldemar link), Avantage dual-mobility cup (Biomet), Polar dual-mobility cup (Smith and Nephew).
All cups were placed after bone impaction grafting.
Two-stage revision.
Most patients (38) were revised for ARMD and all were symptomatic. During revision there was mainly acetabular bone loss in 44 hips. All hips were revised to a 28 mm metal or ceramic head with titanium sleeve, in 44 cases combined with a polyethylene cemented acetabular cup. Due to instability with insufficient femoral anteversion and/or insufficient tension, 4 hips were revised to a dual-mobility system. 2 hips were revised in a 2-stage procedure because of a suspected infection. In 3 patients it was not possible to disconnect the head from the femoral stem during the revision. In these cases we also revised the stem. Figure 2 shows a scatterplot of the acetabular component orientation.
Scatterplot showing acetabular component orientation prior to revision.
Complications
There was a complication in 12 (24%) revisions (Tab. II). 4 patients (all in the ARMD group) needed a reoperation. Re-revision to a dual-mobility articulation was indicated in 2 patients due to recurrent dislocations, resulting in stable hips at latest follow-up. In total 1 of 3 patients with a perioperative fracture during stem revision underwent a re-revision due to dislocation of the revised hip. This patient ultimately developed an infection, leading to removal of the prosthesis. Three patients had more than 1 complication at follow-up. In 6 patients there was visible significant damage of the gluteal medial muscle by the pseudotumour during surgery; 3 of these patients had a dislocation after the revision. In 24 of the revisions there was visible taper corrosion after removal of the head.
Complications after revision procedure grouped by reason for revision
Recurrence of the pseudotumour
A total of 6 patients had a discrete residual mass at CT evaluation after 1 year. The mass was commonly located on the anterior side relative to the hip and corresponded in all cases to a mass present on the preoperative CT scan. In 1 patient there was a residual symptomatic mass compromising the neurovascular bundle at the anterior side. This pseudotumour was subsequently resected by a vascular surgeon through an anterior approach, resulting in resolution of complaints.
Metal ions
Mean cobalt in whole blood decreased from 6.7 µg/L (0.59-211) preoperatively to 1.05 µg/L (0.59-58.8) 1 year postoperatively in patients without a contralateral metal articulation (Fig. 3). These data are expressed as percentage decline per day (Figs. 4 and 5).
Whole blood cobalt levels (µg/L) pre- and post-revision in unilateral and bilateral MoM.
Decline rate (in %) of cobalt after revision.
Decline rate (in %) in years of chromium after revisions.
Clinical follow-up
The mean HOOS score after 1 year follow-up was 58.8. In Table III the HOOS score is grouped by reason for revisions. Mean HOOS score 2 years post-revision was 61.9. Mean scores after 1 year follow-up for the SF-36 were 34.7 for the physical component summary and 52.2 for the mental component summary. After 2 years 35.2 and 53 respectively. A total of 40 patients were willing to undergo the revision procedure again, 6 patients were not, 3 patients refused cooperation and 1 patient had a recent re-revision. This results in a worst-case satisfaction rate of 80%, and best case 88%. Figure 6 shows the revision free survival after revision in months.
Hoos score grouped by reason for revision * 1 year/2 years after revision
received HOOS scores per group: ARMD 23, Loosening 3, Instability 2.
Revision free survival after revision in months.
Discussion
Our study reveals no recurrence of pseudotumour and resolution of metal ion elevation in patients revised from large-head MoM to a conventional bearing system with a small femoral head. Although we were able to successfully treat metal debris-related adverse reactions, the complication rate was high and clinical results poor.
We treated a cohort of patients with a single type of prosthesis after application of a uniform extensive screening protocol. We also followed a uniform surgical approach for revision, always using bone impaction grafting and if possible a conventional polyethylene cup with a 28 mm head This was initially a metal head and in later cases we used a ceramic head with titanium taper sleeve.
Our study has a number of limitations. Firstly, the small numbers, short term follow-up and various reasons for revision precluded a statistical analysis of complications. Secondly, we did not include a control group in this series. Therefore, is was not possible to compare clinical outcome and complications between groups.
Thirdly, 3 different acetabular and 2 different femoral implants were used in the revision operation. Although we aimed to preserve the stem and revise the acetabular component with a conventional polyethylene cemented cup, several factors such as component orientation, abductor function and available bone stock urged the use of dual-mobility designs. In later cases we opted to avoid cobalt chrome components and switched to a modular titanium sleeve and ceramic head in the remaining patients.
Fourthly, we had no CT immediately postoperatively as a reference for later imaging. Therefore we cannot be certain that the postoperative residual masses did not represent a recurrence. However all postoperative masses were present on the preoperative CT scan. During surgery we conservatively resected the anterior part of the swelling through a posterior approach, which was hard to reach and often in close proximity to femoral vessels and nerve. Therefore we consider the postoperatively found swellings to be most likely residual than recurrent.
Fifthly, we had insufficient preoperative functional scores for analysis and a further 6 patients refused submission of patient-reported outcome scores for legal reasons. Legal issues seem to influence our response rate in clinical scores, as lawyers actively advice patients not to submit clinical scores. Legal issues in the literature have also been reported to negatively influence the results of patient-related outcome scores (PROMS) (16). Finally, these are short-term results, possibly too brief to record progression of a potential residual pseudotumour.
The most suitable revision articulation to reduce the risk of dislocation is a matter of ongoing debate. Use of constrained liners raise concerns about impingement and early failure, especially in young active patients (17). Dual-mobility devices have been proposed as a solution (18), but these lack sufficient long-term follow-up and suitable registry data and in our opinion should be used with caution. Additionally, the use of large femoral heads might potentially increase the risk of taper-related problems (18).
In 6 patients there was a small residual mass at follow-up; there was no progression of this mass in the subsequent follow-up and there were no recurrences. A recurrence rate of up to 20% after revisions has been reported. It is hypothesised that remaining wear debris after revision is responsible for this recurrence (10, 14). Another hypothesis is that new wear debris can occur at the taper junction after revision, especially with the use of large femoral heads. The low recurrence rate in our study with the uniform use of 28 mm heads seems in accordance with the latter hypothesis (19, 20).
The outcome of the PROMS using the HOOS score and the SF-36 were poor, especially in the ARMD subgroup. One explanation for these inferior clinical results is significant and sometimes massive tissue destruction, which was observed intraoperatively in the ARMD group, and associated extensive resection of soft tissue. In patients revised for loosening, requiring much less of a debridement, the clinical scores seem better (Tab. III) A high complication rate is also responsible for the poor clinical results. One specific problem is the inability to remove the femoral head during surgery due to corrosion of the taper. In 3 patients it was necessary to perform a stem revision because of this problem; this was associated with 2 femoral fractures during attempts to remove the head. Both patients had a poor clinical result. To tackle this problem a ‘pulley puller’ device that matched the prosthesis was developed in close cooperation with the manufacturer, after which the problem became less frequent.
The HOOS score found in our study compares to scores reported after a two-stage revision procedure of an infected hip arthroplasty (21). Revision for infection has many similarities to revision for ARMD, such as presence of an inflammatory reaction, the need for extensive debridement, and sometimes multiple operations.
In our study, whole blood ion levels generally decreased after the revision, generally reaching normal levels one year postoperatively in unilateral cases.
We found a relatively high number (n = 12) of non-pseudotumour-related traditional revisions. Before our recall we were unaware of the problems in these patients, as they only surfaced during our extensive screening program effective since May 2010. These revisions should therefore be considered as a ‘recall effect’. It is our experience that our screening protocol revealed clear indications for revision that were missed or at least delayed by the traditional routine follow-up program after primary total hip arthroplasty. Only patients with symptomatic pseudotumours were scheduled for revision surgery.
Apart from some residual swellings located on the difficult-to-reach anterior side of the hip with a posterior approach, no pseudotumour recurrence was found and metal ions returned to normal, at least in the unilateral cases. But predominantly due to the resection of large quantities of periarticular tissue, the damage from local inflammation and the revision of well-fixed components, a high complication rate and poor clinical results were encountered. Revision of a MoM arthroplasty should therefore be decided with caution, especially in patients with mild symptoms and no evident abductor muscle damage.
Ultimately, revising a metal-on-metal hip arthroplasty to a conventional cemented polyethylene or dual-mobility cup with bone impaction grafting and a 28 mm head can adequately treat the high ion levels and probably the disease of ARMD, but at a cost.
Footnotes
Financial support: None.
Conflict of interest: None.