Introduction

Use of large-diameter metal-on-metal (MoM) articulations in THA increased, at least in part, because of the possibility of achieving improved joint stability and excellent wear characteristics in vitro [11, 12, 28, 39, 55, 57]. Enhanced stability is afforded by greater ROM to impingement and increased jump distance, that is the distance required for the head to pull out from the acetabular shell to subluxate [39, 52]. Previous studies from our center with a MoM device with a 38-mm fixed-diameter head and cobalt-chromium (CoCr) monoblock shell reported no dislocations in the early 3-month postoperative period compared with an early incidence of 3% with 28-mm heads in a MoM device [12, 57]. More recently we reported on dislocation rates in primary THA with large heads, > 36 mm in diameter, in several material combinations, including 1635 large-diameter MoM THAs [39]. With a mean followup of 3 years, only one dislocation occurred. Recently there have been concerning reports of adverse reactions to metal debris (ARMD), pseudotumors, and systemic complications to metal ions [1, 9, 15, 19, 20, 35, 46, 47, 53, 54, 56, 63, 65]. Three large-diameter MoM devices have been voluntarily recalled by the manufacturers [64]. However, reports indicate considerable variation in ARMD frequency and outcomes between different devices and sometimes even for the same device [9, 16, 26, 29, 32, 34, 44, 46, 47, 60]. For example, although some researchers have reported good survival and a low incidence of ARMD with the Magnum™ device in THA (Biomet, Inc, Warsaw, IN, USA) [29, 44, 60], two studies have reported a high incidence of ARMD with pseudotumor formation when a comprehensive intensive screening protocol was implemented, revealing rates of 39% at 3.6 years in one study [9] and 54% definite, probable, or possible ARMD at a mean followup of 6 years [46].

There is much debate about and several unresolved issues with respect to large-diameter MoM THAs, including the appropriate indications for these devices, whether any large-diameter MoM THA devices are safe to use in some patient populations, what risk factors are for ARMD and device failure, what the long-term patient-reported outcomes and survival are of large-diameter MoM bearings in THA, and what is the best practice for monitoring patients who have these devices. It is estimated that more than 1,000,000 MoM bearings were used worldwide since 1996, and national registry data have indicated failure rates two to three times higher with these devices in THA compared with non-MoM bearings [30]. Likewise, our center reported a nearly twofold higher failure rate with MoM bearings in THA compared with metal-on-improved polyethylene bearings (4% versus 2%, respectively) [48]. Our early experience with MoM THA was very promising, and we became confident that we were providing the most durable, functional articulation possible for our patients. Then we began seeing some puzzling failures with unexplained pain, normal radiographic appearance, well-fixed components at the time of revision, and in some cases alarming soft tissue changes. Failures became more frequent and we learned more about ARMD as well as other modes of failure with these devices. We ceased using all metal-on-metal bearings several years ago but must continue to treat and care for our patients who have had MoM THA. Certainly further study of these devices and continued monitoring of patients with these devices are warranted to determine what the frequency of failure is for our center and determine how best to manage our patients who have large-head MoM THA.

This study therefore attempts to determine at a minimum of 2 years’ followup (1) the proportion of patients who experienced a dislocation in a large series of large-diameter MoM THAs involving two designs by a single manufacturer; (2) the short-term survivorship obtained with these implants; (3) the causes of failure and the proportion of patients who developed ARMD in this series; and (4) whether there were any identifiable risk factors for revision such as sex, disease profile, procedure type, age, height, weight, body mass index (BMI), preoperative clinical scores, cup diameter, cup angle, or stem diameter.

Patients and Methods

Study Design and Setting

We identified all patients who underwent THA at our center and reviewed those performed with MoM articulations with head diameters ≥ 38 mm and having minimum 2-year followup. All available clinical records and radiographs were reviewed to determine the frequency of dislocation and failure and the reason for failure. Patients who underwent revision of the acetabular component for any reason were compared with patients who were not revised to determine risk factors for failure.

Between October 2001 and February 2010, 1451 patients (1695 hips) underwent large-diameter MoM primary THA at our institution by two surgeons (AVL, KRB) using two devices from a single manufacturer (Biomet, Inc), which represents 48% (1695 of 3567) of the primary THAs performed during that period. Of the 1695 hips, 55 nonrevised hips were in 51 patients who either died within 2 years of surgery or had not returned for 2-year followup before their death. In addition, 200 nonrevised hips were in 180 patients who have been lost to contact before completing 2-year followup, leaving a cohort of 1440 hips (85%) in 1235 patients with minimum 2-year followup available for review. A monoblock CoCr acetabular component with a fixed 38-mm inner diameter, increasing outer diameter, and applied porous plasma-sprayed titanium (Ti-PPS) fixation surface, the M2a-38 (Biomet, Inc) (Fig. 1), was used until September 2005 in 636 THAs. A resurfacing style monoblock CoCr acetabular component coated with Ti-PPS, the Magnum™ (Biomet, Inc) (Fig. 2), features a 3-mm fixed shell thickness, inner diameters up to 60 mm, and articulates with a modular CoCr head component, 6 mm smaller in diameter than the cup, that mates with a titanium insert taper adapter, and was used beginning in November 2004 in 804 THAs. Clinical records and radiographs were reviewed to determine incidence and etiology of acetabular revision. Patients whose THAs were revised were compared with patients with surviving THAs to identify risk factors.

Fig. 1
figure 1

The M2a-38 design (Biomet, Inc, Warsaw, IN, USA) was introduced at our center in October 2001. It is a monoblock CoCr acetabular component with a fixed 38-mm inner diameter, increasing outer diameter, and applied Ti-PPS fixation surface.

Full size image
Fig. 2
figure 2

The Magnum™ design (Biomet, Inc, Warsaw, IN, USA) was used in our center beginning in November 2004 through February 2010. It is a resurfacing style monoblock CoCr acetabular component coated with Ti-PPS and features a 3-mm fixed shell thickness, inner diameters up to 60 mm, and outer diameters up to 66 mm. It articulates with a modular CoCr head component that is effectively 6 mm smaller in diameter than the cup, which is mated with a titanium insert taper adapter.

Full size image

Participants/Study Subjects

The main indications for the use of large-diameter MoM THAs during the study period were THAs in patients who were younger, more active, and more high demand, and who were perceived to have the potential for benefit from an alternate bearing. We also used them in patients who were considered at a higher risk of instability and dislocation postoperatively and in patients with adequate bone stock to achieve stable fixation without the use of adjunct screw fixation. A contraindication for the use of large-diameter MoM devices was in patients with a known history of renal insufficiency.

Description of Experiment, Treatment, or Surgery

The surgical approach for the majority of patients was direct lateral, used in 94% of hips (1351 of 1440). Of those, a less invasive modification introduced in 2003 was used in 902 hips. A muscle-sparing anterior supine intermuscular approach was used in 6% (88 of 1440), and a less invasive posterior approach was used in one. All femoral components implanted were made by the same manufacturer as the MoM acetabular components used (Biomet, Inc): two were cemented, CoCr primary components and two were cementless, titanium revision components; all others were cementless, titanium primary stems with 78% Mallory-Head Porous (1121 of 1440), 21% TaperLoc Microplasty (301 of 1440), and 14 TaperLoc. All devices used in this study were cleared by the US Food and Drug Administration and were used according to labeling provided.

Aftercare

Patients were allowed immediate full weightbearing with the assistance of ambulatory aids. Ambulatory aids were discontinued because the patient walked with a minimal or no limp and without pain. All patients were instructed to use the same postoperative hip precautions for 6 weeks, including to sleep on their back, use an elevated toilet seat, use a cushion in all low chairs, not to flex at the waist past 90°, and to avoid excessive adduction such as crossing one leg over the other.

Variables, Outcome Measures, Data Sources, and Bias

Patients returned to our clinic for followup in the immediate postoperative period at approximately 6 weeks and then were instructed to return yearly thereafter or sooner if a problem arose. Patients were assessed at each followup time using the Harris hip score (HHS) [18] and, beginning in 2011, the UCLA activity score [3]. Patients presenting postoperatively with a painful THA were evaluated in a similar fashion to patients with painful THA of any bearing type including clinical history, physical examination, radiographs, and laboratory testing for infection, searching for possible extrinsic as well as intrinsic causes. Using a risk stratification process, if the THA is suspicious for ARMD, further testing may be performed including hip aspiration with manual cell count, whole body bone scan, electromyography, serum metal ion testing, ultrasound, CT scan, and/or magnetic artifact reduction sequence (MARS) MRI studies [30, 36]. Postoperative records were reviewed to determine the frequency of dislocation and the frequency and reasons for component revision. Our understanding of ARMD evolved over the course of the study and continues to evolve. In five patients revised before 2007 for unexplained pain, determination of ARMD was based mainly on intraoperative findings. Since that time, as better means of diagnosis have become more widely available, ARMD has been better identified before revision. Patients with components that were revised for any reason were compared with patients not revised in terms of preoperative demographics including age at surgery, sex, underlying disease profile, height, weight, BMI, preoperative HHS, and perioperative factors including procedure type, surgical approach, cup outer diameter, head diameter, neck length, and stem diameter. Radiographic assessment was performed to measure the inclination (abduction) angle of the acetabular component on the AP pelvis view at 6 weeks postoperatively. Measurements were done by a blinded observer (JBA) on digital radiographs using Intelerad software (Montreal, Quebec, Canada). The angle between the inferior aspect of the ischium and the face of the acetabular component was measured using a Cobb angle tool.

Statistical Analysis and Study Size

Analysis of device survival/failure times were displayed using Kaplan-Meier curves and significance testing was done using proportional hazards models. An initial exploratory analysis was performed looking at several potential risk factors (gender, BMI, activity level, stem diameter, age, etc) individually. Those factors that were associated with p values < 0.2 were used in a multivariable proportional hazards regression model and significance was assessed at the 0.05 level. Adjusted risk ratios and their 95% confidence intervals (CIs) are reported for significant factors. All analyses were carried out using JMP/11 Pro ® software (SAS Institute, Cary, NC, USA).

Demographics and Description of Study Population

Diagnosis by hip for the majority of patients was osteoarthritis (81% [1161 of 1440]), mean patient age at surgery was 58 years (SD 10), and 55% of patients were males (677 of 1235, 787 of 1440 hips) (Table 1). Activity level was categorized as light labor for the majority of patients (51%; 740 of 1440) followed by moderate manual labor in 28% (401 of 1440).

Table 1 Demographic and perioperative variables for all patients with minimum 2-year followup
Full size table

Accounting for All Patients and Study Subjects

All patients had minimum 2-year followup (mean, 7 years; range, 2–12 years). No patients were recalled specifically for this study; all data were obtained from medical records and radiographs.

Results

Dislocation occurred in one hip of 1440 (< 1%) 2 years after primary THA, and the patient was treated with acetabular revision because of recurrent dislocation after closed reduction and cast-bracing. Unfortunately, the patient, although initially alert and oriented after the operative procedure, experienced a cerebrovascular accident in the afternoon on the day of surgery and died a few days later.

Kaplan-Meier analysis revealed survival free of component revision was 87% at 12 years (95% CI, 84%–90%) (Fig. 3). There was no difference in survival between the two acetabular devices used (Fig. 4). A total of 108 hips in 102 patients have been revised at a mean time of 4 years postoperative (range, same day to 11 years; SD 3). Ninety-six of these revisions have been the result of aseptic causes.

Fig. 3
figure 3

Kaplan-Meier analysis revealed survival free of component revision was 87% at 12 years for the 1235 patients with 1440 large-diameter MoM THA in our study.

Full size image
Fig. 4
figure 4

Kaplan-Meier analysis for the two acetabular devices in our study showed no significant difference in survival, with 87% survival at 12 years for the M2a-38 (Biomet, Inc, Warsaw, IN, USA) and 88% survival at 10 years for the Magnum™ (Biomet, Inc).

Full size image

Indications for revision were dislocation in 1% (one of 108), acetabular aseptic loosening or failure of ingrowth in 31% (34 of 108) of revised hips, well-fixed acetabular components revised unrelated to the articulation in 9% (10 of 108), infection treated with two-stage exchange in 11% (12 of 108), acetabular malposition revised the same day in one hip, periprosthetic femoral fracture treated with stem exchange only in 3% (three of 108), and ARMD with elevated serum metal ions, pseudotumor formation, and/or soft tissue damage present on revision in 44% (47 of 108). Complete revision operative records were available for 45 of 47 patients revised for ARMD. Corrosion at the neck/head junction was present in 40% (18 of 45) and pseudotumor was noted in 49% (22 of 45) of ARMD revised patients.

Age at surgery, cup angle of inclination, and sex were all risk factors for component revision after controlling for potentially confounding variables (Table 2). Increasing cup angle and female sex were associated with poor survivorship, whereas older patients had better outcomes. Height, weight, BMI, underlying diagnosis, preoperative HHS, activity level, surgical approach, cup type and diameter, and stem type and diameter were not risk factors for revision.

Table 2 Risk factors for component failure
Full size table

Discussion

MoM bearings were reintroduced in THAs in the 1990 s in response to the shortcoming of articulations with polyethylene, namely osteolysis and loosening secondary to polyethylene wear, and high rates of dislocation. Early results with MoM bearings certainly were promising [14, 37, 38]. Based on the success of large-diameter MoM bearings in hip resurfacing arthroplasty and a desire for enhanced stability and durability, MoM designs for THA evolved to larger diameter head sizes. Although dislocation rates have decreased with the move to larger head diameters, failure rates have increased (Table 3). Amid numerous reports of high failure rates and concerning reports with ARMD, pseudotumors, and systemic complications to metal ions [1, 6, 9, 16, 19, 20, 24, 31, 35, 40, 46, 47, 53, 54, 56, 63, 65], three large-diameter MoM devices have been voluntarily recalled by the manufacturers: the Articular Surface Replacement (ASR) by DePuy (Warsaw, IN, USA), the Durom by Zimmer (Warsaw, IN, USA), and the R3 by Smith & Nephew (Memphis, TN, USA) [64]. Published studies have reported considerable variation in frequency of ARMD and failure for large-diameter devices [9, 16, 26, 29, 32, 34, 44, 46, 47, 60]. Our center previously reported a higher frequency of revision in patients with MoM THA compared with metal-on-improved polyethylene THA [48]. In an effort to monitor the results of the treatments we perform and strive for the best possible care of our patients, continued followup of our patients with MoM THA is necessary. The present study therefore sought to evaluate a large group of these patients at a minimum followup of 2 years, specifically with respect to dislocation, implant survivorship, the causes of failure (and the frequency of ARMD), and risk factors for revision.

Table 3 Published results with large-diameter (≥ 36 mm) metal-on-metal THA
Full size table

The first limitation of our study is that it was retrospective and therefore may be subject to selection bias. We tended to use large-diameter MoM devices in our younger, more active, more high-demand patients, and perhaps the higher activity levels in these patients resulted in higher rates of wear. Another limitation resulting from the retrospective nature is that 104 patients (126 hips) died during the study period, and 51 of those patients (55 hips) had not been seen for a 2-year clinical followup visit. Only 13 of the patients (14 hips) died before reaching 2 years postoperative. We know that eight of the patients (nine hips) had revisions before death. The other 96 patients (117 hips) had no known complications or revisions at the time of last followup. Another weakness of the study is that in addition to the 51 patients (55 hips) who died before a 2-year clinical assessment, minimum followup was not available for 200 hips in 180 presumed living patients. The Social Security Death Index was searched for all patients. Attempts were made to contact the patient at their last known address and phone numbers, by contacting referring and family physicians listed, and by searching available free Internet services. However, minimum 2-year clinical followup was available for 85% of patients.

Previous studies from our center with the M2a-38 reported no dislocations in the early followup period compared with an early proportion of 3% with 28-mm heads in a MoM device [12, 57]. We have reported on dislocation rates at our center in primary THA with large heads, > 36 mm in diameter, in several material combinations, including 1635 large-diameter MoM THAs [39]. With mean followup of 3 years, only one dislocation occurred for a frequency of 0.06% compared with an earlier experience involving a primary direct lateral approach THA with small heads (≤ 32 mm) in which the dislocation rate was 1% (12 of 1518) [41]. Likewise, in a report on 8059 cementless THAs with large-diameter MoM articulations from the Finnish Arthroplasty Register, the authors report 11 revisions resulting from dislocation for a rate of 0.1% compared with 175 dislocations in a series of 16,798 cemented metal-on-polyethylene THAs for a rate of 1% [47]. In another report from the Finnish Arthroplasty Register assessing risk of revision for dislocation by head diameter, the risk with head diameters > 36 mm (10,444 hips) was 0.09 compared with 28-mm-sized implants [28].

The Kaplan-Meier survival of 87% at 12 years in the current series of large-diameter MoM THAs with these components is somewhat lower than rates reported in combined data from Australia, England and Wales, and New Zealand [16], data from the Finnish Arthroplasty Register [47], and three independent series (Table 3) [29, 44, 60]. Component survival in the current study was lower than our earlier reported experience with 779 primary THAs with metal-on-improved polyethylene in which we observed 98% survival at 4 years [48]. Likewise, in the 2013 report of the National Joint Registry for England, Wales, and Northern Ireland, the effect of the bearing with a single widely used cup/stem combination, the Pinnacle/Corail (DePuy), was analyzed [49]. They observed that at 8 years postoperatively, the chance of a first revision with a MoM bearing was 11% compared with 2% with a ceramic-on-polyethylene bearing.

The predominant indications for component revision in the current series were ARMD, observed in 48% of hips revised (47 of 108), and aseptic loosening or failure of ingrowth, observed in 35% (38 of 108). There is considerable variation in published reports as to the risk of revision and ARMD incidence between different large-diameter MoM THA devices as well as for the devices used in the current study (Table 3) [9, 16, 26, 29, 32, 34, 44, 46, 47, 60]. Several studies have reported good survival and a low incidence of ARMD with the devices we used in this report [12, 29, 32, 34, 44, 47, 51, 57, 60, 61]. Two concerning studies involving the Magnum™ device in large-diameter MoM THA have reported a high incidence of ARMD with pseudotumor formation when a comprehensive screening protocol was implemented [9, 46]. Bosker et al. [9] performed CT scans in patients with Magnum™ THA and verified pseudotumors in 39% at mean 4-year followup. They observed a fourfold increased risk of pseudotumor formation when serum cobalt levels were > 5 µg/L and recommended close monitoring of all patients with MoM hip arthroplasty. Mokka et al. [46] conducted MARS MRI screening and serum metal ion testing of a group of patients with Magnum™ THA with mean 6-year followup and identified a 54% incidence of definite, probable, or possible ARMD. In a recent report involving patients with modular Pinnacle MoM THA (DePuy), screening with MARS MRI confirmed a 31% incidence of ARMD asymptomatic patients, raising concern that the need for routine cross-sectional imaging for all asymptomatic patients with MoM implants warrants further study [15].

We previously reported on our experience with all MoM articulations in primary THA through 2006 at our center and observed a higher incidence of failure in females than males [33]. In the current study with additional patients and further followup, we again observed a greater frequency of component failure in females than males (Table 2). Register data from Australia have shown higher risk revision for females than males who have undergone primary MoM THA with head sizes > 32 mm [4]. However, there was no appreciable difference in revision risk between males and females with head sizes ≤ 32 mm. Younger age in our study was a risk factor for revision (Table 2) with risk multiplied by 0.978 (decreased by approximately 2%) for each increased year of age at surgery. The Australian National Joint Replacement Registry also reported poorer results in younger patients with a cumulative percent revision for primary MoM THA of 21.8 at 12 years in patients < 55 years old and 12.1 in patients ≥ 75 years old [4]. Increased cup angle of inclination was a risk factor revision in the current study (Table 2) with an increased risk of 3% for each increase of one degree. Several studies of MoM resurfacing hip arthroplasty have correlated excessive angle of inclination with elevated serum and joint fluid levels of metal ions and increased wear secondary to edge loading [30]. However, in our study only 3% of cups overall (41 of 1440) and only 5% of cups (five of 108) in patients who underwent component revision had an angle of inclination > 55°. Height, weight, BMI, underlying diagnosis, preoperative HHS, activity level, surgical approach, cup type and diameter, and stem type and diameter did not correlate with risk of revision in the current study.

Large-diameter MoM THAs are associated with a very low dislocation rate. However, the revision rate with these devices was higher than expected. Failure secondary to ARMD or lack of ingrowth has been frequent. Taper corrosion may represent an additional source of metal debris. Risk factors for revision were younger age, female sex, and increased angle of cup inclination. In our practice we have discontinued the use of MoM devices. Patients with MoM devices, like all patients with total joint arthroplasties, should be encouraged to return for clinical and radiographic followup. Closer monitoring of patients with MoM implants has been recommended by several regulatory agencies and consensus groups of physicians [2, 16, 17, 21, 30, 36, 43, 62]. As we have previously stated, there should be a low threshold to perform a systematic evaluation of patients with MoM THA because early recognition and diagnosis will facilitate the initiation of appropriate treatment before significant adverse biological reactions occur [30, 36].