High Complication Rates Following Revision of Aseptic Metal-on-Metal Hips to Dual Mobility Bearings

Abstract

Background

Patients with metal-on-metal (MoM) bearings continue to experience complications requiring revision total hip arthroplasty (rTHA). Revision to dual mobility (DM) bearings is often debated given a persistent metal-metal interface. We sought to report on the outcomes of rTHA from MoM to DM constructs.

Methods

Patients who underwent rTHAs from 2010 to 2024 were retrospectively reviewed. Only patients with MoM constructs undergoing revision to DM were included. Demographic and perioperative data were collected, in addition to complications and survivorship free of rerevision.

Results

Thirty-two hips in 28 patients underwent rTHA to DM constructs. The mean time from MoM THA to rTHA was 9.4 years. All rTHAs were indicated for pain and concerning serum metal ion levels. Findings at the time of rTHA included fluid collections in 68.7% (n = 22) of cases, taper corrosion in 43.7% (n = 14), and acetabular component loosening in 18.75% (n = 6). Mean preoperative cobalt and chromium levels were 13.6 ug/L and 10.5 ug/L, respectively. The rerevision rate was 31.25% (n = 10 hips) at a mean follow-up of 6 years. The most common reasons for rerevision were dislocation (n = 4, 12.5%), followed by acetabular loosening and periprosthetic joint infection (PJI) (n = 3 each, 9.38%). At most recent follow-up, serum metal ions had decreased to mean cobalt of 0.98 ug/L and chromium of 5.15 ug/L.

Conclusions

Revision surgery of MoM THA is associated with high complication rates. Serum cobalt and chromium ions can be expected to steadily decline after revision to DM despite the metallic interface.

Level of Evidence

IV; retrospective case series.

Introduction

Metal-on-metal (MoM) primary total hip arthroplasty (THA) was heavily popularized in the early 2000s, with over half a million MoM THAs being completed from 2006 to 2012. [1] The rise of this articulation was predominantly fueled by the complications associated with metal-on-polyethylene particulate debris leading to osteolysis and aseptic loosening before highly crosslinked polyethylene was popularized. [2] However, MoM bearings were similarly found to produce smaller particulate debris in higher quantities, with an overall higher surface area available for wear. [[3], [4], [5], [6]] This ultimately led to the recall of several MoM implants due to unacceptable complication rates compared to metal-on-polyethylene bearings. The rate of adverse local tissue reaction (ALTR) with a mass in patients with MoM hips has been reported between 20% and 35%. [7,8] These masses, along with particulate debris derived from MoM bearings, have been associated with osteolysis, as well as ALTR and soft tissue necrosis, which has the potential to lead to catastrophic complications. [9]

In light of the failure rate of MoM constructs, revision surgery is often indicated, with reported revision rates as high as 37% for MoM bearings. [10] These revision surgeries are often more challenging and are associated with less predictable outcomes. Rerevision rates after these procedures have been reported between 11% and 20% at mid-term follow-up, with instability being one of the most common complications. [[11], [12], [13]] Due to this, dual mobility (DM) constructs are often utilized to increase the relative head size and jump distance, theoretically reducing the risk of dislocation. However, with DM constructs of most companies, there remains a metal junction between the liner and acetabular component which carries the potential risk of liner malseating and/or persistent elevation of serum metal ions. [[14], [15], [16], [17]] This has led many to question the clinical appropriateness of modular DM liners in complications of MoM bearings in order to eliminate metal junctions.

There are limited data in the literature to aid surgeons in clinical decision-making, and early studies have demonstrated that DM may help improve the risk of instability after MoM revision [18,19]. The purpose of this study was to review the clinical outcomes and postoperative complications and metal ions of our institution’s aseptic revisions of MoM THAs using DM constructs. We hypothesized our complication rate to vary between 10% and 37% based on previously cited studies.

Material and methods

Study design and patient selection

Following institutional review board (IRB: Pro00110001) approval, a retrospective single-center cohort study was performed evaluating patients who underwent revision THA (rTHA) between January 2010 and December 2024 (Fig. 1). The electronic medical records were queried to identify all patients who underwent revision of MoM THAs during the study period. Cases were screened manually to confirm indication and implant construct.

Figure 1.

STROBE flow diagram of cohort identification. STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

Inclusion criteria comprised patients who underwent:

• •Revision of a MoM acetabular construct
• •Elective revision for ALTR, metal wear disease, or elevated metal ion levels
• •Conversion to a modular DM articulation (either isolated liner exchange or full acetabular component revision)
• •Availability of preoperative serum cobalt and chromium levels

Exclusion criteria were:

• •Preoperative or intraoperative diagnosis of periprosthetic joint infection
• •Revision to constrained liners or conventional polyethylene liners
• •Absence of preoperative metal ion data

This resulted in 32 THA revisions in 28 patients. The study cohort included both monoblock and modular MoM acetabular components. The Strengthening the Reporting of Observational Studies in Epidemiology guidelines were followed.

Data collection

Demographic variables collected included age, sex, body mass index, and laterality. Operative details included type of primary MoM construct (monoblock vs modular), revision strategy (liner exchange vs acetabular revision), intraoperative soft tissue findings, and presence of abductor deficiency. Time from primary MoM arthroplasty to revision was recorded.

Postoperative outcomes included:

• •Complications (including dislocation, infection, rerevision)
• •Reoperation
• •Mortality
• •Final clinical follow-up duration

Metal ion assessment

Serum cobalt and chromium levels were obtained as part of institutional surveillance for MoM arthroplasties. Preoperative metal ion levels were defined as the most recent measurement prior to revision surgery.

Postoperative levels were recorded at final available follow-up. Values reported as <1 μg/mL (institutional lower detection threshold) were recorded as 0.9 μg/mL for calculation of means. Values reported as “none detected” were recorded as 0 μg/mL.

In patients with bilateral MoM arthroplasties, metal ion values were included until revision of the contralateral hip. Final postoperative levels were recorded after bilateral revision when applicable. Due to the retrospective design, postoperative metal ion measurements were not uniformly available; therefore, metal ion analyses were performed on complete cases only.

Statistical analysis

Data are presented using descriptive statistics, including means with standard deviations (SD) for continuous variables and counts with percentages for categorical variables. Survivorship free of any reoperation was evaluated using Kaplan-Meier survival analysis. Survival analysis was performed using SPSS (Version 23; IBM, Armonk, NY). All other statistical analyses were performed using Microsoft Excel (Version 16.89; Microsoft Corporation, Redmond, WA). Given the limited sample size and absence of a comparison cohort, analyses were exploratory and descriptive in nature.

Results

Table 1 presents the study demographics including age, gender, and indications for surgery. Most patients carried a pre-THA diagnosis of degenerative joint disease and were over the age of 50. Table 2 presents data on the primary components used at the time of MoM THA. Table 3 presents data on the demographics and timing between MoM and rTHA. Prior to rTHA, metal levels were registered at a mean of 17 months before surgery (range, 0-100; SD, 32.42). Cobalt levels ranged from <1 to 32.6 ug/l with a mean of 13.6 ug/l (SD, 14.4) and chromium 0.7 to 21.2 ug/l, mean 10.5 (SD, 18.9).

Table 1.

Characteristics and demographics of patients undergoing primary metal-on-metal total hip arthroplasty.

Primary MoM THA	Total Number of THA (n = 32)	Percentage of Cohort
Mean age	54
Min	28
Max	78
Standard deviation	9.77
Females	15	47%
Males	17	53%
Indication of primary THA
Osteoarthritis	25	78.1%
Avascular necrosis	5	15.6%
Unknown	2	6.3%

Table 2.

Study cohort implant type.

Primary cups
Articular surface replacement (ASR)	11	34.4%
Pinnacle	14	43.8%
M2a Magnum	4	12.5%
Trident	1	3.1%
X3	1	3.1%
Conserve	1	3.1%

Primary cup sizes
Min	50
Max	60
No data	10

Primary head sizes
Min	36
Max	53
No data	6

Primary stems
Summit	18	56.3%
S-ROM	2	6.3%
AML	1	3.1%
Taperloc	7	21.9%
Rejuvenate	1	3.1%
Corail	1	3.1%
Wright Microport	1	3.1%
Accolade 1	1	3.1%

Table 3.

Demographic and characteristics of patients undergoing conversion of total hip arthroplasty to dual mobility.

Age at conversion THA to DM
Average	64
Min	38
Max	89
Standard deviation	10.50
Most recent BMI
Mean	29.1
Min	19.6
Max	42.1
Standard deviation	6.02
Timing from MoM THA to revision THA
Mean (mo)	113
Min	10
Max	230
SD	59.10

BMI, body mass index.

Imaging findings

Preoperative magnetic resonance imaging (MRI) demonstrated imaging abnormalities consistent with ALTR in 9 of 12 hips (75%) with available MRI reports. Periarticular fluid collections were identified in 9 hips (75%), with sizes ranging from approximately 2 cm to 20 cm in maximal craniocaudal dimension. Large ALTR collections (>10 cm) were present in 3 hips (25%). Most fluid collections were located posterolaterally and demonstrated communication with the hip joint and/or pseudocapsule in 7 hips (58%).

Extension into adjacent soft tissues was common, including involvement of the greater trochanteric bursa or iliotibial band in 6 hips (50%) and iliopsoas bursal or intramuscular extension in 4 hips (33%). Intrapelvic extension was observed in 1 hip (8%).

Complex imaging characteristics were also frequently observed, including thickened pseudocapsule and heterogeneous internal debris in 6 hips (50%). T1 hyperintense components suggestive of proteinaceous or metal-laden fluid were present in 5 hips (42%). Solid synovitic components or nodular pseudocapsular thickening were noted in 4 hips (33%). Osseous involvement, including focal erosion or scalloping of the greater trochanter or acetabulum, was identified in 3 hips (25%).

Three hips (25%) did not demonstrate a discrete periarticular fluid collection; however, one of these cases demonstrated heterogeneous soft tissue signal changes consistent with ALTR particle disease without a defined mass.

Overall, MRI findings demonstrated a spectrum of ALTR severity, with the majority of hips exhibiting moderate to extensive periarticular fluid collections and soft tissue involvement.

Median cobalt levels were higher in hips with ALTR with a mass compared to those without (20 μg/L vs 9.2 μg/L). Chromium levels demonstrated a similar but less pronounced trend (median 8 μg/L vs 6.05 μg/L). While elevated metal ion levels were associated with the presence of ALTR, no consistent linear relationship was observed between metal ion concentration and maximal ALTR volume on MRI. Given the limited sample size available, we did not attempt to perform statistical comparison.

At the time of rTHA, 68.7% of patients were noted to have an intraoperative fluid collection adjacent to the hip. Taper corrosion and loose acetabular components were found in less than half the patients (n = 14, 43.7% and n = 6, 18.75% respectively). None of the femoral components were found to be loose. The mean time from MoM to rTHA was 113 months (range, 14-230) with the mean age being 62 (range, 38-89). Most recent body mass index was 29.1 (range, 19.6-42.1).

Table 4 presents operative details of the acetabular components used in rTHA. All of the inner diameter heads of the femoral construct were 28 mm. These were made of ceramic most commonly (n = 27, 84.37%) followed by metal (N = 4) and one of oxidized zirconium-coated zirconium alloy (Oxinium, Smith & Nephew; Memphis, TN). The mean follow-up for the rTHA patients was 75.77 months (range, 4-156; SD, 57). One patient in this study has since passed away from unrelated medical causes.

Table 4.

Operative details of patients undergoing revision total hip arthroplasty.

Revision cup size	Total Number of THA (n = 32)
Mean	58.29
Min	52
Max	70
Cups left in place (n)	4
Acetabular components
Trident	21
G7	2
Bimentum	1
Redapt	4
Microport	1
Left in place	4
Dual mobility heads
Min	42
max	60

Serum metal levels decreased after rTHA in the subset of patients with long-term metal level screening (Table 5). Fifteen hips, 46.87% of rTHAs, had metal ion levels examined after rTHA to DM. The latest levels after rTHA for cobalt ranged from 0 to 3.6 ug/L with a mean of 0.98 and SD 1.04. Chromium levels also decreased after rTHA with values ranging from 0.6 ug/L to 11.2 ug/L with a mean of 5.15 (SD, 6.54). These values were collected at a mean of 45 months after rTHA.

Table 5.

Metal ion levels.

Before rTHA	Months before revision THA	Cobalt (ug/l)a	Chrome (ug/l)a
Min	0	<1	0.7
Max	100	32.6	21.2
Mean	17	13.6	10.5
SD	32.42	14.4	18.9
After rTHA to DM
Min	6	None detected	0.60
Max	153	3.60	11.20
Mean	45	0.98	5.15
SD	41.71	1.04	6.54

^aFifteen hips had postoperative values with 2 patients having bilateral MoM THAs that underwent a rTHA at a mean of 45 months (SD: 41.71, range: 6-153).

There were 10 revisions later performed in the patients who underwent rTHA. Figure 2 demonstrates the revision distribution with almost equal distribution of PJI, instability, and acetabular component loosening. None of these patients have had further revisions during the study period, with a mean follow-up of 75.77 months (SD, 57; range, 4-153). The Kaplan-Meier curve for all-cause rerevisions is shown in Figure 3, and the hazard ratios with 95% confidence intervals are shown in Table 6.

Figure 2.

Indications for additional revision total hip arthroplasty.

Figure 3.

The Kaplan-Meier curve for all-cause rerevisions.

Table 6.

Hazard ratio analysis of rerevision risks.

Complication-specific hazard ratio versus no complications	Hazard ratio (B)	95% Confidence interval lower	95% Confidence interval upper
PJI	21.554	3.405	136.429
Aseptic loosening	3.305	0.851	12.844
Dislocation	9.491	2.314	38.932

Discussion

Revision THA associated with MoM articulations often presents a complex problem for surgeons, secondary to the associated bony and soft tissue destruction, and often yields inferior outcomes, with instability remaining a large problem. The DM dual articulation design was proposed to increase jump distance and maximize impingement-free arc of motion. [20] Results have been promising, showing dislocation rates between 1.5% and 10.4% in the revision setting, as well as decreased risk of rerevision compared to standard articulations. [[21], [22], [23]] However, there has been a paucity of literature looking at the effectiveness of DM constructs specifically in the setting of prior failed MoM articulations. One study published in 2023 evaluated a case series of 24 patients who underwent revision THA to an off-label ADM/MDMX3 DM liner (Stryker, Mahwah, NJ, USA) and followed 18 patients for a mean follow-up of 53.2 months. [19] Additionally, they found that at a minimum of 2 years (30-95 months), metal ions dropped to 0.8 μg/. Here the authors reported on one revision for persistent pain leading to rerevision to a standard head and liner. Another case series that evaluated revision THAs from MoM to DM reported good outcomes on 11 revised hips with only 1 rerevision. [24]

It has been well established that cobalt and chromium serum levels may increase secondary to wear of metal-on-metal articulations. There has been a lack of consensus on the exact threshold of metal ions that puts patients at a high risk of requiring revision surgery, with safe upper limits being reported between 4 and 10 μg/L. [25,26] These levels have been shown to significantly decrease after revision surgery using ceramic-on-polyethylene and metal-on-polyethylene articulations. [27] However, given the modular nature of modern DM constructs, there has been concern for possible secondary metal particle generation due to a new metal-metal interface. Small, short-term studies have shown that cobalt and chromium ions remain largely normal using these modular DM constructs, with approximately 5% having elevated levels greater than 1 ug/L. [[15], [16], [17]] The preoperative mean cobalt and chromium levels in our cohort, 13.6 ug/L and 10.5 ug/L respectively, were well above the literature reported safe upper limits. We also had a substantial decrease in both metal ion levels, similar to previously reported literature, even with the use of modular DM constructs. Although the postoperative mean cobalt in our study was within reference range of <1 ug/L (0.98 ug/L), the mean chromium level did not regress to normal levels, which supports longitudinal surveillance of patients revised to a modular DM construct.

Cobalt and chromium follow different metabolic elimination pathways, which result in different postoperative decline patterns after revision of MoM implants. Several studies have demonstrated that metal ion levels decrease at variable rates after revision and that this process is multifactorial. Durrani et al. evaluated 16 patients undergoing revision for painful MoM hips and observed a rapid decline in cobalt levels with a slower reduction in chromium. [28] Similarly, Nicolli et al. reported similar trends in their cohort of 7 patients. [29] These differences are thought to relate to variations in metal ion storage and distribution. Cobalt typically circulates as a free ion and is cleared relatively quickly through renal excretion, whereas chromium has greater binding to plasma proteins and erythrocytes and may be sequestered in periprosthetic tissues and macrophages. As a result, chromium levels may remain elevated for longer periods after revision.

Our study reports that 10 of the 32 hip (31.25%) that underwent revision to a DM construct ultimately required rerevision. This is similar to previous literature, with rerevision rates in patients with MoM articulations being reported as high as 28% by Munro et al. and 26% in the Australian Joint Registry. [30,31] Prosthetic joint infection accounted for 3 of the 10 patients, in which 2 were successfully managed with a debridement, antibiotics, and implant retention procedure and 1 required a 2-stage exchange. It has been reported that there may be a higher incidence of PJI associated with MoM hips, with rates between 5.6 and 6.7% being reported in the literature, much higher than the reported rate of approximately 1% for primary THA. [[32], [33], [34]] The hypothesis behind this proposed increased rate of PJI is centered around the increased inflammatory response associated with ALVAL and preponderance of necrotic tissue from metal wear but has yet to be definitively confirmed. Of note, most of the hips in this study also had fluid collections around the hip which is a gross sign of ALVAL.

Despite the use of DM in our study population, 4 of the 32 patients (12.5%) sustained dislocations that necessitated rerevision. All of these patients were successfully managed with an increased head ball length and retention of the DM construct, with no further dislocations reported at a mean follow-up of 6 years. The rate of instability in patients who have undergone revision of a MoM articulation has been reported as high as 22%. [35] It is thought the local soft tissue destruction and necrosis mediated by mass formation can lead to significant abductor deficiency, leading to increased rates of instability. Our cohort did have a high rate of ALTR presence, approximately 69%, which could have contributed to the predilection toward instability in our population. In our case, 3 out of 4 (75%) of patients with dislocations had ALTR present. However, it should be noted that Klemt et al. examined outcomes of revision DM constructs in a broad range of patients with abductor deficiency and noted no dislocations and 87% implant survivorship at 3 years, but less than half of their cohort of 42 patients receiving DM constructs had original MoM articulations. [36]

Meriem et al. evaluated a similar population of patients who received DM constructs for failed MoM bearings and reported a total of 6 rerevisions in their cohort of 64 patients, including 2 for instability, 2 for prosthetic joint infection, and 2 for aseptic loosening. [37] Colacchio et al. reported 2 total rerevisions, 1 for instability and 1 for periprosthetic fracture, in their cohort of 29 patients also assessing DM constructs in MoM revisions. [18] Both of these studies had results similar to ours, in which we had 4 rerevisions for instability, 3 for prosthetic joint infection, and 3 for aseptic loosening of the acetabular component.

Limitations

There are multiple limitations of our study that should be acknowledged. First and foremost, this is a retrospective review study, which carries inherent limitations related to the study design. Second, our sample size of 32 hips is small and difficult to draw externally valid conclusions, but our sample size is similar to the few prior studies looking to report on outcomes of a relatively rare clinical scenario. Our study is also looking at a single cohort in those that received a DM construct with revision of a MoM bearing, whereas including a control group could have added strength to our results. Nonetheless, our sample size was so limited that the power needed to achieve a comparison group of patients who were treated with different revision constructs such as constrained liners would require a great sample size. Furthermore, only 15 of the 32 patients in our cohort had postoperative serum metal ions measured, further limiting the generalizability of the serum metal ion results. This combined with the lack of standardize guidelines regarding optimal timing of metal ion level measurement induced variability in our retrospective findings. Lastly, obtaining metal ion labs from patients who have not followed up recently is challenging as some of these patients have moved.

The size of ALTR damage to the surroundings tissues may also have an effect of the outcomes of these surgeries, yet pathological grading and tissue measurements may not correlate with results. Other factors that may also affect dislocation rates like acetabular cup placement were also not evaluated. Additionally, acetabular cup alignment, a potential contributor to postoperative instability, was not systematically evaluated in this series. Bone defect classification was not consistently available. This represents an important limitation of our study as this likely may have influenced outcomes. In addition, in our cohort, 46% of patients were female. While this may suggest an increased risk of bone loss or smaller native bone stock compared to males, we did not have standardized measures of bone defect size to formally assess its contribution to outcomes. Lastly, our mean follow-up of 75 months allows us to comment on mid-term results, but long-term follow-up is needed given concerns of wear persist throughout the lifespan of a prosthetic articulation.

Conclusions

MoM THA is associated with a localized inflammatory response that can lead to soft tissue and bony destruction that often leads to a high rate of complications, often necessitating revision surgery. Our study showed that the use of a DM construct in this patient population had an instability rate below to some prior described incidences and serum cobalt and chromium levels significantly decreased postrevision, but there still remains a relatively high rate of rerevision. However, our results are comparable to the limited literature available on this specific topic, and a DM construct may still provide protective benefits for these patients. Failed MoM hip arthroplasty remains a difficult problem to solve, and comparative studies need to be completed in order to optimize the care of these patients. It is important to note that these findings should be interpreted in the context of a highly complex patient population, in which extensive bone loss and soft tissue compromise likely contributed to the elevated complication rates. Our results, therefore, may reflect the challenges inherent to this cohort rather than suggesting inferiority of DM constructs.

CRediT authorship contribution statement

Samuel Rosas: Writing – review & editing, Writing – original draft, Visualization, Validation, Project administration, Formal analysis, Data curation, Conceptualization. J. Scott Holmes: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Kevin A. Wu: Writing – review & editing, Writing – original draft, Visualization, Validation, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Andrew M. Schwartz: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition. Michael P. Bolognesi: Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Investigation, Formal analysis. Samuel S. Wellman: Writing – original draft, Validation, Supervision, Resources, Project administration, Methodology, Funding acquisition. Sean P. Ryan: Writing – original draft, Supervision, Software, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Conflict of interest

M.P. Bolognesi is a paid consultant for Total Joint Orthopaedics; owns stock or stock options in Amedica; received research support from Smith & Nephew, Zimmer Biomet, and DePuy, A Johnson & Johnson Company; is a member of the medical/orthopaedic publications editorial/governing board at Arthroplasty Today and Journal of Arthroplasty; and is a board member/committee appointments for American Association of Hip and Knee Surgeons, Eastern Orthopaedic Association, and Orthopaedic Research and Education Foundation. S. Ryan received research support from Smith & Nephew and Zimmer. A. Schwartz is a board member/committee appointments for American Association of Hip and Knee Surgeons and Musculoskeletal Infection Society. S. Wellman is a paid consultant for Total Joint Orthopaedics; owns stock or stock options in Joint Development; received research support from Smith & Nephew, Zimmer Biomet, DePuy, A Johnson & Johnson Company, Medacta, and Stryker; is a member of the medical/orthopaedic publications editorial/governing board at Journal of Arthroplasty; and is a board member/committee appointments for Hip Society; all other authors declare no potential conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2026.102010.