Lack of early dislocation for dual mobility vs. fixed bearing total hip arthroplasty: A multi-center analysis of comparable cohorts

Abstract

Background

Dislocation is a major cause of morbidity and revision surgery following total hip arthroplasty (THA). To address such issues, dual mobility (DM) bearings were introduced as a more stable alternative to fixed-bearing (FB) prostheses. As such, we compared DM and FB systems in a cohort study in terms of dislocations, readmissions, and revisions.

Methods

A 27 multi-center retrospective review was performed of 664 DM and 218 FB cases from the same manufacturer with mean follow-up of 2.09 years and 1.83 years, respectively. Patient reported outcome measures (PROMs) including Harris Hip Score (HHS), SF12, EQ5D, and Lower Extremity Activity Score (LEAS) were evaluated as well as dislocation rates, readmissions, and revisions rates. We also performed a survivorship analysis through Kaplan-Meier estimator. Students t-test was used for normally distributed continuous data and Fisher exact test (P < 0.05) was used for discrete data.

Results

There were 0 dislocations in the DM (0%) group and 2 dislocations in the FB (0.92%) group (p = 0.06). Latest follow up HHS revealed a significant difference between groups (91.44 DM and 87.81 FB; p = 0.006). In addition, there was significant difference between DM and FB on SF12 Physical Component Score (PCS) (46.83 and 44.55, respectively, p = 0.015). Also, readmission rates at 30, 60 and 90 days remained lower for DM than for FB at each time point (1.05% vs. 2.75%, 1.81% vs. 2.75%, and 1.81% vs. 2.75, respectively). Overall, DM had a lower revision rate at 1.51% compared to 2.29% for FB (p = 0.24). The revision breakdown for DM revealed 0 (0%) for both Anatomic Dual Mobility (ADM) and Modular Dual Mobility (MDM) due to the acetabular component.) There was a difference, 14 (87.5%) for ADM and 2 (12.5%) due to the femoral component. The survivorship analysis revealed no significance difference between DM and FB at 4 years (97.90% and 97.26%, respectively).

Conclusion

In comparison to patients who undergo FB THA, DM bearings have improved PROMs and a lower rate of dislocation, readmission, and revision.

1. Introduction

Hip instability remains one of the most common reasons for hip revision aftertotal hip arthroplasty (THA) in the United States.¹ According to a national review study, 22% of all revision cases were due to instability.² The cumulative risk of dislocation within the first preoperative month is 1% and within the first year is 2%. This risk increases by 1% per 5-year period and reaches approximately 7% after 25 years post-surgery.³

Biomechanical requirements enable optimal mechanical function and overall stability of the hip arthroplasty. Component malposition is the most common cause of hip stability, which is affected by surgical technique and enhanced by an anterolateral approach.⁴ Dislocations can occur due to i) contact between neck of the prosthesis and articular component, resulting in primary impingement and ii) contact between bony femur and bony pelvis, leading to secondary impingement.⁵ Often times, impingement and patients who are at a high risk of dislocation may result in hip instability without the influence of surgical technique.⁶ Risk factors can include neuromuscular diseases, cognitive dysfunction, those aged >75 years and those with previous surgical history to the hip.⁷

Dual mobility (DM) systems have been introduced in the United States as an effective design option to increase stability after total hip arthroplasty (THA). The concept originated as early as 1974 by Bousquet and Rambert in an attempt to combine the low friction principle of earlier and larger diameter femoral head that has shown improved stability.⁸ The Anatomic Dual Mobility (ADM) system and Modular Dual Mobility (MDM) polyethylene liners have been designed to increase impingement-free range of motion and decrease dislocation risk by increasing the effective head size of the implant.⁹ Stroh et al. showed dislocation rates of 0.1% for primary THA vs. 2%–7% for primary FB THA and 3.5% for revision THA using DM vs. 16% for revision FB THA.¹⁰ Additional measures are supported to improve construct stability, reduce impingement, and reduce wear. These include highly crossed linked polyethylene (HXLPE), larger heads, porous coating, and cementless prosthesis.¹¹

Previous literature has been limited to small retrospective studies that lack a control group or used a small matched cohort. In this study, we used age, BMI, and gender comparable cohorts of patients in DM and FB groups. The study was designed to retrospectively compare a series of DM hips with a comparable series of FB hips to examine (1) Patient Reported Outcome Measurements (PROMs) at 2.1 years in the DM group and 1.8 years in the FB group, and (2) incidence of dislocation, readmission, and revision as well as a survivorship analysis.

2. Materials and methods

We retrospectively queried the registry of 27 affiliated centers for patients who underwent THA with a DM or FB system by the same manufacturer. A statistician then performed a comparative analysis to include age, BMI, and gender comparable cohorts. An initial search was performed for all patients undergoing a primary THA with a DM or FB implant, non-resurfacing, and non-metal-on-metal implant to determine PROMs. A second search was performed with the same inclusion criteria to determine postoperative complications and revisions. We identified the cohort of patients and found them to be comparable for age, body mass index, and gender.

There were 664 patients in the DM group and 218 patients in the FB group (Table 1). DM and FB THAs were performed from July 2011 through April 2018. There was no significant difference in age, gender, and body mass index (Table 1). Median follow-up was 2.09 in the DM group and 1.83 years in the FB group.

Table 1.

Demographics of the comparable cohort.

	DM (N = 664)	FB (N = 218)	P value
Age, Mean ± SD, y	61.73, ±9.16	61.68, ±9.02	0.9508
Gender, F:M, n (%)	308(46.39%):356(53.61%)	102(46.79%):116(53.21%)	0.9175
BMI, mean± SD, kg/m2	29.48, ±4.92	29.70, ±4.86	0.5666
Follow-up, median (range), y	2.09 (0.07–7.03)	1.83 (0.10–5.40)	0.0187

2.1. Eligibility and inclusion criteria

All patients that received a DM acetabular component and FB component were found from the institutional registry system at Stryker Orthopaedics Mahwah, NJ, USA from previous prospective trials. Both the DM and FB THAs were performed over the course of 8 years, from July 2011 through April 2018. We found the cohort of patients to be comparable for gender, age, and body mass index. The surgeries were performed at 27 affiliated centers that underwent a uniform posterior approach to THA. The inclusion criteria remained consistent across the groups. Inclusionary diagnoses included: degenerative arthritis, post-traumatic arthritis, osteonecrosis, inflammatory arthritis, and degenerative arthritis.

2.2. Surgery

All THAs procedures were performed with a uniform posterior approach from July 2011–April 2018. All patients received the fixed-bearing component, MDM dual mobility prosthesis (Fig. 3A) or ADM dual mobility prosthesis (Fig. 3B). ADM combines an anatomic shell that offers a large diameter with low friction. The ADM cup is a monoblock cobalt chrome alloy cup with a titanium surface and hydroxyapatite. MDM provides the surgeon the choice of fixation surfaces and screwhole configuration for THA. The DM systems are designed to enhanced stability and jump distance, which can increase range of motion in patients compared to the FB prosthesis. All of the DM prostheses were cementless, which offers better long-term bond and eliminates the possibility of the breakdown of cement.

Fig. 3.

(A). Modular dual mobility (MDM X3; stryker orthopaedics, Mahwah, NJ). (B). Anatomic dual mobility (ADM X3; stryker orthopaedics, Mahwah, NJ).

2.3. Fixed bearings

Based on the time period, from July 2011–April 2018, FB acetabular components were used and implemented according to the posterior approach. The implants were noncemented, press-fit titanium shells with either HA coating or porous metal.

2.4. Clinical assessment

Harris Hip Score (HHS), Short Form Health Survery (SF-12), Lower Extremity Activity Scale (LEAS), and EQ-5D scores were evaluated at 2.1 years follow-up in the DM group and 1.8 years follow-up in the FB group.

2.5. Statistical analysis

Normally distributed continuous data was compared using the Students t-test data while the discrete data were compared using the Fisher exact test. In addition, Chi-square test and Wilcoxon tests were used. A p value < .05 was determined to be statistically significant.

3. Results

There were significant differences in clinical outcomes between the DM group and FB group (Table 2). HHS at latest follow-up was significantly different for DM and FB, 91.44 ± 12.76 and 87.81 ± 16.52, respectively (p = .0187, Fig. 1). In addition, SF12 PCS was significantly different for DM and FB as well, 46.83 ± 10.42 and 44.55 ± 12.18, respectively (p = 0.015, Fig. 1). In the DM group, there were no dislocations and no intra-prosthetic dislocations at 2-year mean follow-up. In the FB group, there were 2 dislocations (0.92%). Readmission trends were similar for 30 days, 60 days and 90 days. The percentage of readmission for FB was the same for all three time periods at 2.75% while the percentage for DM is was 0.05% at 30 days and 1.81% at both 60 days and 90 days. There were 10 revisions (1.51%) in the DM group and 6 revisions (2.75%) in the FB group. Specifically, the most common cause for revision in the DM group was Periprosthetic femoral fracture (0.75%). In the FB group, the most common cause was deep joint infection (1.83%) (Table 3). In regards to the DM revisions due to the femoral component, 14/16 (87.5%) were attributed to ADM option compared to 2/16 (12.5%) from the MDM option. However, there was 0 (0%) revisions due to the acetabular component in both cohorts (Table 4). The survivorship analysis revealed no significant difference between DM and FB groups at 4 years but DM was higher at 97.90% with 95% CI (95.86%, 98.94%) compared to 97.26% for FB with 95% CI (93.46%, 98.86%) (Fig. 2).

Table 2.

Patient Reported Outcomes Measurements in DM and FB cohorts.

	DM (N = 664)	FB (N = 218)	P value
HHS at preop, Mean ± SD	54.03, ±12.94	52.08, ±14.97	0.0873
HHS at latest, Mean ± SD	91.44, ±12.76	87.81, ±16.52	0.006
Change of HHS from preop to latest(Delta), Mean ± SD	37.49±16.24	34.52± 19.89	0.0645
SF12 PCS at Preop, Mean ± SD	31.89, ±8.93	29.69, ±8.69	0.0016
SF12 PCS at the latest, Mean ± SD	46.83, ±10.42	44.55, ±12.18	0.015
Change of SF12 PCS from preop to latest(Delta), Mean ± SD	14.99, ±12.02	14.78, ±13.01	0.8308
SF12 MCS at Preop, Mean ± SD	52.15, ±11.75	52.70, ±11.64	0.5501
SF12 MCS at the latest, Mean ± SD	54.38, ±8.35	54.67, ±9.18	0.6641
Change of SF12 MCS from preop to latest(Delta), Mean ± SD	2.25, ±10.59	1.91, ±11.37	0.6898
EQ5D TTO at Preop, Mean ± SD	0.61, ±0.20	0.60, ±0.20	0.758
EQ5D TTO at Latest, Mean ± SD	0.88, ±0.15	0.84, ±0.19	0.0099
Change of EQ5D TTO from preop to latest(Delta), Mean ± SD	0.27, ±0.20	0.24, ±0.24	0.0879
EQ5D VAS at Preop, Mean ± SD	68.58, ±18.85	64.35, ±23.46	0.0183
EQ5D VAS at Latest, Mean ± SD	82.78, ±14.77	79.55, ±17.65	0.0198
Change of EQ5D VAS from preop to latest(Delta), Mean ± SD	14.20, ±19.25	15.34, ±26.17	0.5671
LEAS at Preop, Mean ± SD	8.98, ±3.00	8.52, ±2.74	0.0421
LEAS at Latest, Mean ± SD	11.33, ±3.21	10.86, ±3.38	0.0676
Change of LEAS from preop to latest(Delta), Mean ± SD	2.33, ±3.62	2.31, ±3.53	0.9348

Fig. 1.

Mean clinical outcomes between DM and FB at postop.

Table 3.

Postoperative complications and revisions.

	DM (N = 664)	FB (N = 218)	P value
Re-admission within 30 days, n(%)	7(1.05%)	6(2.75%)	0.0992
Re-admission within 60 days, n(%)	12(1.81%)	6(2.75%)	0.4098
Re-admission within 90 days, n(%)	12(1.81%)	6(2.75%)	0.4098
Dislocation, n(%)	0 (0%)	2 (0.92%)	0.0609
Revision, Deep joint Infection, n (%)	2 (0.30%)	4 (1.83%)	NS
Revision, Femoral Component Loosening, n (%)	1 (0.15%)		NS
Revision, Femoral Component Subsidence, n (%)	1 (0.15%)		NS
Revision, Perioprosthetic Femoral Fracture (%)	5 (0.075%)	2 (0.92%)	NS
Revision, Femoral Impingement, n (%)	1 (0.15%)		NS
Revision, Acetaular (n%)	0(0%)	0(0%)	NS
Revision, Femoral n(%)	10(1.51%)	6(2.75%)	0.2452
Mortality	8(1.20%)	5(2.29%)	0.2471

Table 4.

Revision breakdown for Dual Mobility.

ADM (%)	MDM (%)	Reason
1 (6.25%)	1 (6.25%)	Deep Joint Infection
1 (6.25%)	0 (0%)	Femoral Component Loosening
1 (6.25%)	1 (6.25%)	Femoral Component Subsidence
4 (25.0%)	0 (0%)	Perisprosthetic Femoral Fracture
1 (6.25%)	0 (0%)	Femoral Impingement
0 (0%)	0 (0%)	Total, Acetabular Component
14 (87.5%)	2 (12.5%)	Total, Femoral Component

Fig. 2.

Kaplan-Meier Survivorship graph.

4. Discussion

DM implants have historically been advantageous over FB implants in terms of cost-effectiveness, reduced rate of dislocation, and improved survivorship.¹² Epinette et al. showed that the potential maximum cost-savings of THA-DM over THA-FB was more than 100 million Euros per year in France.¹³ Rowan et al. and Romagnoli et al. both showed the benefit of using DM cup in preventing dislocation over the FB implants with a significant difference in both studies, in a 100 patient age and BMI matched cohort (p < .01) and a systematic review that included 2408 hips (p < .00001).^14,15 Darrith et al. revealed that DM implants have a low rate of instability at a mean follow-up of 8.5 years and a good overall survivorship in primary THA at a mean of 1.3 years.¹⁶ The risk of dislocation of uncomplicated primary THA has decreased from 5% to 1% in centers because of the anterior surgical approach, larger femoral heads (32 mm–36 mm), and improved repair of the posterior tissues. This was clearly supported within the analysis from the dislocation, revision, and readmission rates favoring DM over the FB implants (Table 3).

Our study has built on the analysis by comparing PROMs in a comparable cohort analysis over a 2.1 year period for the DM group and 1.8 years period for the FB group. Rowan et al. showed no difference in postoperative Modified Harris Hip Score between the DM (87.2 ±16.6) and the FB cohort (87.9 ± 13.67, P = .78).¹⁴ Our study revealed a significant difference between the groups (Table 2) given larger cohorts, n = 664 for DM and n = 218 for FB. Ko et al. maintained similar results by finding that DM cups increased the ROM before impingements, adding 30.5° in flexion, 15.4° in adduction, and 22.4° in external rotation.¹⁷

Polyethylene wear and cup size have significant implications in DM systems for intraprosthetic stability. Polyethylene wear >2 mm has been identified as a risk factor for late dislocation. However, advances in polyethylene characteristics, such as the increased cross-linking for DM systems, resulted in lower femoral head penetration rates. Femoral head size ≥36 mm has been showed to reduce dislocation rates. Specifically, a larger head-neck ratio and metal head diameters contribute to better range of motion and stability, respectively. In a similar fashion, our study revealed similar dislocation results to the existing literature. Rowan et al. and Prudhon et al. found a difference in dislocations between DM and FB (7 vs. 0, p = .01) and 41 for DM dislocation compared to 11 for FB dislocation, p < .001.^14,18 Darrith et al. showed a similar mean dislocation rate of 0.46%, showing the stability of DM prostheses. Particularly, the large diameter of DM cups can assist in stability by altering the head-neck ratio.¹⁶

Our study revealed a lower readmission rate at all time points for DM compared to FB (Table 3). De Martino et al. stated that instability after THA accounts for 32.4% of readmissions, which accounted for some of the difference between the groups.¹⁹ Epinette et al. stressed the increased costs associated with hospital readmissions.¹³ Schairer et al. had higher readmission rates of 4% at 30 days and 7% at 90 days and even higher readmission rate (10%) at 90 days for revision patients.²⁰ To minimize cost and revision readmission, DM showed in a reduction readmission rates.

Revision surgery can take place as a result of implant wear and loosening, infection, dislocation, reaction to metal ions or periprosthetic fracture. In terms of revision rates, DM showed a lower revision rate throughout the literature to support our difference, 1.51% for DM and 2.75% for FB. Darrith et al. showed a similar mean revision rate of 2.0% for any revision of the DM bearing.¹⁶ Rowan et al. found a similar revision breakdown with 0 revisions for instability in the DM cohort and 2 revisions for instability in the FB cohort.¹⁴ The analysis revealed the stability of both DM cups. This is important since there is a lack of large data on DM constructs throughout literature. This experience had similar results to a multi-centre study, in supporting the use of a DM cup over FB prosthesis.²¹ This concept is also supported by another institution that valued the benefits of reducing the rates of complications even in light of a theoretical concern of an adverse local tissue reaction.²²

The ADM option is a monoblock deisgn without any modularity, reducing the concern regarding the potential for elevated metal ion levels. Highly crosslinked polyethylene and optimized head/neck ratios have helped mitigate the early failure of ADM bearings.²³ Jones et al. demonstrated such results in lack of dislocations using ADM implants, even in high risk patients, age>75 yrs, female gender, prior sugery, and ASA grade>3.²⁴ However, Waddell et al. showed the risk for intraprosthetic dislocation remains, as seen in the damage of the cobalt chrome acetabular shell. It resolved that elevating the polyethylene liner can avoid the “catching” caused by the direct axial traction.²⁵ Although we did not observe any dislocations, the ADM cup showed stability in having 0 (0%) revisions due to the acetabular component and 6/12 (50%) of DM revisions were attributed to femoral loosening, cracks, or fractures (Table 4).

The MDM option provides a large head size (up to 58 mm) and a thick PE outer liner. These features addressed the concerns that both Haughom et al. and Amstutz et al. and had regarding a small head size (36 mm) and a thin PE implanted liner.^26,27 MDM creates a mechanical block to improve range of motion and decrease wear and loosening seen in constrained liners. Two studies, Anderson et al. and Berend et al., showed the limitations of constrained components in finding dislocations rates of 29% and 28% respectively.^28,29 Particularly, the screw holes allowed for visualization of the acetabular floor during impaction and improve fixation. We anticipate reduced wear of the MDM option and would be a valuable study in the future. However, this study showed equal acetabular stability in the MDM and ADM options(0% for both).

There were limitations in the study. We performed a retrospective cohort study given that DM is only system that is currently used. We can aim for a prospective study that includes a MDM and ADM comparison to determine their respective efficacies and PROMs differences. We addressed the short follow-up period by matching the cohorts and maintaining high patients’ numbers that are not prevalent in the literature. We hope to gain insight into alternative surgical techniques in the future but the homogeneity of FB implants and surgical technique help to mitigate this issue. Our study benefited in terms of using 27 centers to obtain PROMs in order to improve the overall generalizability of the results. While the indications between the centers may vary, general DM hips are used in higher risk patients. In addition, since 70% of all dislocations occur within the first year, the absence of dislocation at shorter follow-up is very significant. We chose not to match the cohorts to avoid limiting efficiency or statistical power. The combination of a multi-center approach and a BMI, age, and gender comparable cohort analysis solidifies the importance of this study.

The difference in PROMs between DM and FB acts in a unique way in the literature, given the size of the DM cohort (n = 664). Chughtai et al. used a cohort that was 453 patients and obtained a similar analysis of dislocation in a cementless stem.³⁰ Several studies, Philippot et al. and Boyer et al. benefited from larger follow-up time periods (17 years and 22 years, respectively) but 77% of dislocations occur within the first year.^7,31 This study represents one of the largest dual-mobility studies, in terms of patient numbers and maintains the lack of dislocations within a 2.1 years follow-up period for DM and 1.8 years follow-up period for FB. It also shows superior clinical outcomes in the DM group compared to the FB group that isn't examined thoroughly in current literature.

Funding

There is no funding source.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Declaration of competing interest

GW is a Stryker paid consultant.