**Long-Term Results of a Second-Generation, Small-Diameter, Metal-On-Metal Bearing in Primary Total Hip Arthroplasty at 14-Year Follow-Up**

**Tobias Reiner 1,\*, Matthias C. Klotz 1, Kirsten Seelmann 1, Fabian Hertzsch 1, Moritz M. Innmann <sup>1</sup> , Marcus R. Streit 1, Timo A. Nees <sup>1</sup> , Babak Moradi <sup>1</sup> , Christian Merle 1, Jan Philippe Kretzer <sup>2</sup> and Tobias Gotterbarm 1,3**


Received: 18 November 2019; Accepted: 20 January 2020; Published: 24 January 2020

**Abstract:** (1) Background: The objective of the present study was to review the clinical and radiological results of a small-head, MoM bearing in primary THA and to determine blood metal ion levels at long-term follow-up. (2) Methods: We retrospectively evaluated the clinical and radiological results of 284 small-diameter, MoM 28-mm Metasul THA at a mean follow-up of 14.5 years, and measured blood metal ion concentrations in 174 of these patients. (3) Results: After 14 years, survival free for revision due to any reason was 94%. Proximal femoral osteolysis was seen in 23% of hips, and MRI demonstrated ARMD in 27 of the 66 investigated hips (41%). Mean cobalt, chromium, and titanium ion concentrations were 0.82 μg/L (range 0.22–4.45), 1.51 μg/L (0.04–22.69), and 2.68 μg/L (0.26–19.56) in patients with unilateral THA, and 2.59 μg/L (0.43–24.75), 2.50 μg/L (0.26–16.75), and 3.76 μg/L (0.67–19.77), respectively in patients with bilateral THA. Twenty-nine percent of patients showed cobalt or chromium ion levels > 2 μg/L. (4) Conclusions: Despite good clinical long-term results, increased blood metal ion levels (cobalt or chromium > 2 μg/L) were found in approximately one-third of asymptomatic patients, and proximal femoral osteolysis and ARMD were frequently seen in this cohort. Blood metal ion analysis appears helpful in the long-term follow-up of these patients in order to identify individuals at risk. In accordance with contemporary consensus statements, symptomatic patients with elevated metal ion levels and/or progressive osteolysis should be considered for additional CT or MARS MRI to determine the extent of soft tissue affection prior to revision surgery. Further studies are necessary to investigate the clinical relevance of ARMD in asymptomatic patients with small-head, MoM THA.

**Keywords:** Metasul; 28 mm small head; metal-on-metal THA; cobalt; chromium; titanium; blood metal ions

#### **1. Introduction**

Second-generation, small-head, metal-on-metal (MoM) total hip replacements were reintroduced in 1988 by Weber [1], and initiated the rise of metal-on-metal hip arthroplasties at the beginning of this

century. Metal-on-metal bearings were commonly implanted in younger patients hoping to overcome the polyethylene-wear-related complications of periprosthetic osteolysis and aseptic implant loosening. In 2008, metal-on-metal articulations were used in approximately 35% of all hip replacements in the United States [2]. High early failure rates, especially in large-diameter, metal-on-metal total hip arthroplasties (THA), and the growing incidence of adverse local tissue reactions related to metal wear, led to a swift decrease in the use of those implants in the subsequent years [3–6]. Accumulating metal ions in the joint cavity, which are generated by corrosive degradation of metal wear products, are able to influence both bone metabolism and the immune system through different pathways, contributing to the pathogenesis of periprosthetic osteolysis and the formation of adverse local soft tissue reactions, also referred to as ARMD (adverse reaction to metal debris). Although MoM bearings are rarely used nowadays, the systematic follow-up of these patients will continue to be of clinical importance due to the large number of metal-on-metal articulations that were implanted in past decades, especially in younger patients [7]. Risk stratification algorithms for the management of patients with MoM bearings have been provided by different regulatory authorities [8–10], and published guidelines suggest that small-diameter (< 36 mm) MoM implants are at low risk of developing ARMD. In contrast to large-diameter MoM articulations, a systematic long-term follow-up comparable to conventional THA with routine follow-up intervals of 3 to 5 years in the long term is considered sufficient, and blood metal ion analysis is not recommended in the follow-up routine of patients with small-diameter, MoM articulations [10]. Although some authors have recently raised concerns about the late onset of ARMD associated with increased metal wear of small-diameter, MoM implants [11–13], the results of metal ion analyses in the long-term follow-up of these patients are not clear.

The objective of the present study was (i) to evaluate the clinical and radiological results of small-head, MoM THA at long-term follow-up, (ii) to determine blood metal ion concentrations in a large cohort of patients at a minimum follow-up of 10 years, and (iii) to investigate potential risk factors associated with elevated blood metal ion levels in these patients.

#### **2. Materials and Methods**

#### *2.1. Study Design and Patients*

In this cross-sectional study, we retrospectively evaluated a consecutive series of 262 patients (284 hips) following cementless THA with a 28-mm Metasul metal-on-metal articulation. The study was approved by the ethics committee of the Heidelberg school of medicine (No. S-365/2013), and informed written consent was obtained prior to inclusion of each patient. Surgery was performed consecutively between April 1995 and November 2001 at Heidelberg University Hospital using either a modified Watson-Jones or a transgluteal lateral approach. The indication for the use of a MoM bearing at that time was young patient age and a high expected physical activity level. The mean age of patients at time of surgery was 52 years (range 21 to 74 years). At a mean follow-up of 14.5 years, 44 patients (17%, 33 male, 11 female) had died and 14 (5%) were lost to follow-up, leaving 193 patients (211 hips) who were available for review (Figure 1). Up to the latest follow-up, fourteen hips (5%) had undergone revision surgery. Of the remaining cohort, 174 patients (189 hips) agreed to participate in blood metal ion analysis, which was performed at a mean follow-up duration of 14.5 years (range 10.3 to 18.8 years) after surgery. In order to eliminate other sources of cobalt or chromium ion release, eighteen patients (19 hips) were excluded due to additional metal implants such as total knee replacements [14], and seventeen patients (17 hips) had to be excluded because of femoral components made of cobalt-chromium-alloys. Of the remaining cohort, 113 patients with unilateral THA and 26 patients with bilateral THA were available for further statistical analysis.

**Figure 1.** Flowchart summarizing clinical follow-up and patient selection for metal ion analysis.

A 28-mm Metasul (Zimmer, Winterthur, Switzerland) MoM articulation was used in all hips. The acetabular component of this implant consists of a forged, high-carbide (0.2–0.25%) cobalt-chromium alloy liner, which is embedded in a polyethylene insert. It was used in combination with a cementless, press-fit titanium acetabular shell; 95 hips received an Allofit acetabular cup (Zimmer, Winterthur, Switzerland) and 58 received a Fitmore acetabular component (Zimmer, Winterthur, Switzerland). An uncemented straight-tapered titanium stem with a standard 12/14 mm Euro taper was used in all hips for femoral reconstruction; 128 hips received a CLS Spotorno stem (Zimmer, Winterthur, Switzerland) and 25 received a G2 stem (Depuy Orthopaedics, Warsaw, Poland).

#### *2.2. Clinical and Radiographic Follow-up*

Clinical examination was performed using the Harris Hip Score. Standard pelvis anteroposterior and lateral radiographs of the hip were evaluated with regard to radiolucent lines and osteolysis. We defined periprosthetic osteolysis as a lucent zone absent of trabecular bone, which was not visible on the immediate postoperative radiograph [15]. Radiolucencies and osteolysis were evaluated according to the zones established by Gruen et al. [16] and the classification system of DeLee and Charnley [17]. Cup inclination angles were determined using the TraumaCad software (TraumaCad®, Voyant Health, Columbia, SC, USA), taking the inter-teardrop line as a fixed landmark [18]. In addition, cross-sectional imaging with metal artifact reduction sequence magnetic resonance imaging (MARS MRI) was available in 53 patients (66 hips) of the study cohort, which were retrospectively evaluated regarding ARMD formation. The indication to perform MARS MRI in these patients was blood cobalt or chromium ion level > 1 μg/L. A total of 107 patients in the study cohort fulfilled these inclusion criteria and were invited for MRI as part of a previously published study [13].

#### *2.3. Metal Ion Analysis*

Blood samples were taken using a blood collection system specific for trace metal ion analysis (Sarstedt, Nuembrecht, Germany; Refs. 58.1162.600 and 01.1604.400). The first 5ml of blood were discarded and blood samples were stored at −20 ◦C. Whole blood metal ion analysis was performed at the Geochemical Laboratories at Heidelberg University using high-resolution, inductively-coupled, plasma-mass spectrometry (HR-ICP-MS, Element 2, Thermo Fisher Scientific, Bremen, Germany). ICP-MS is currently considered one of the preferred techniques for blood metal ion measurement [10]. All samples were analyzed at the same time in order to minimize calibration errors arising from the spectrometer. Metal ion analysis was repeated three times in every sample and mean values were calculated. Detection limits of 0.005 μg/L for cobalt, 0.02 μg/L for chromium, and 0.06 μg/L for titanium were established for this method [19]. Additionally, the glomerular filtration rate (GFR) was calculated using the CKD-EPI formula based on the serum creatinine values of each patient.

#### *2.4. Statistical Methods*

Statistical analysis was performed using the software SPSS® for Windows® (version 22.0; SPSS IBM Corp., Chicago, IL, USA) and Graphpad Prism® (version 6.0, Graphpad Software, San Diego, CA, USA). Data were evaluated descriptively as arithmetic mean, standard deviation, median, minimum, and maximum. Demographic data and mean metal ion levels were compared between the bilateral and the unilateral group using the student's t-test. For comparison of categorical variables between the two groups, the chi-square test was used. Kaplan-Meier survivorship analysis was performed with revision for any reason as the endpoint. In the unilateral group, correlation analysis was performed using Spearman correlation coefficient and multivariate linear regression analysis in order to investigate the correlation between blood metal ion concentration and potential risk factors associated with elevated cobalt ion levels, which were defined as gender, cup inclination angle, body mass index, and follow-up length. Additionally, the relationship between periprosthetic osteolysis and blood metal ion concentrations of cobalt, chromium, and titanium was assessed using logistic regression analysis. Correlation was defined as poor (0.00 to 0.20), fair (0.21 to 0.40), moderate (0.41 to 0.60), good (0.61 to 0.80), or excellent (0.81 to 1.00). All tests were two-sided and a p-value < 0.05 was considered significant.

#### **3. Results**

#### *3.1. Survival Analysis*

The cumulative survival rate at 10 years, using revision for any reason as the endpoint, was 96% (95% confidence interval (CI); 92–98%; 235 hips at risk) and 94% (95% CI; 90–96%; 112 hips at risk) at a mean follow-up of 14 years (Figure 2). Of the 14 hips requiring revision surgery, four (1.4%) were revised for adverse reaction to metal debris (ARMD) and four (1.4%) were revised for aseptic loosening of either the femoral (n = 2) or acetabular component (n = 2). Another four hips (1.4%) were revised for infection, and two were revised for late periprosthetic fracture (0.7%). The mean time to revision surgery for ARMD was 10.5 years (range 7 to 15 years), and the mean time to revision surgery for aseptic loosening was 6.2 years (range 3.5 to 11 years).

**Figure 2.** Kaplan-Meier analysis showing the survival free of revision for any cause was 96% (95% CI 92–98%) at 10 years and 94% (95% CI; 90–96%) at a mean follow-up of 14 years.

#### *3.2. Clinical and Radiographic Evaluation*

The mean Harris Hip Score of the cohort was 90 points (range 40 to 100) at the time of follow-up. The mean inclination angle of the acetabular component was 42 degrees (range 29–50 degrees). No femoral component showed radiographic signs of loosening. Radiographs demonstrated femoral osteolysis in 23% of the hips and radiolucent lines > 2 mm in 13% of hips. Osteolysis and radiolucent lines were predominantly located in the proximal Gruen zones. Their distribution is illustrated in Figure 3. Periacetabular osteolysis was rarely seen, with an overall frequency of 2%. MARS-MRI demonstrated pseudotumor formation in 27 of the 66 investigated patients (41%). ARMD were generally small and predominantly cystic in nature. More detailed results of this investigation were previously published in another study of this research group [13].

**Figure 3.** Results of the radiographic evaluation showing the distribution of radiolucent lines (RL) and osteolysis (OL), as seen on anteroposterior and lateral radiographs according to Gruen zones at a mean follow-up of 14 years.

#### *3.3. Metal Ion Analysis*

A total of 139 patients were eligible for blood metal ion analysis, with the Metasul bearing being the only known source for cobalt or chromium ion release (Figure 1). The demographic data of the study cohort are summarized in Table 1. The results of blood metal ion analysis are shown in Table 2 and Figure 4. Patients with bilateral THA showed higher mean cobalt and chromium levels; however, this difference was not statistically significant. Forty-one patients (29%) had either cobalt or chromium ion levels > 2 μg/L, and 23 (17%) showed cobalt or chromium ion levels > 3 μg/L. Ninety-four patients (68%) demonstrated titanium ion levels > 2 μg/L and 26 (19%) had titanium ion levels > 4 μg/L. Four patients showed radiological evidence of femoral neck impingement without disassociation of the acetabular liner as a possible source for increased metal wear, which was visible as a little notch at the femoral neck on the lateral radiograph. Three of the four patients were asymptomatic with a mean HHS of 98 points. Mean cobalt, chromium, and titanium ion levels were 3.23 μg/L, 2.84 μg/L, and 8.69 μg/L, respectively. All other patients with increased metal ion levels showed no evidence of mechanical failure or component loosening on plain radiographs. No patient in the study cohort showed severe chronic kidney disease (GFR < 30 ml/min). Univariate analysis revealed moderate correlation between cobalt and chromium ion concentrations ( = 0.465, *p* < 0.001), and fair correlation between chromium and titanium ion levels ( = 0.228, *p* = 0.015) and between body mass index and cobalt ion levels ( = −0.224, *p* = 0.017). However, in multivariate analysis, none of the tested variables was proven as a risk factor for elevated metal ion levels. Logistic regression analysis showed no association between the presence of periprosthetic osteolysis and blood metal ion levels of cobalt (odds ratio, 0.94; 95% CI, 0.50–1.77; *p* = 0.941), chromium (OR, 1.01; 95% CI, 0.85–1.21; *p* = 0.905), or titanium (OR, 0.88; 95% CI, 0.66–1.16; *p* = 0.362).


**Table 1.** Demographic data of the unilateral and bilateral group for metal ion analysis.

\* indicating statistically significant differences between the two groups


**Figure 4.** Box-and-whisker plots showing whole blood ion concentrations of cobalt, chromium, and titanium. The box marks the range between first and third quartile, with the band inside the box indicating the median and whiskers indicating minimum and maximum data respectively.

#### **4. Discussion**

Small-diameter, metal-on-metal implants are supposed to be at low risk of developing ARMD, and a systematic follow-up comparable to conventional THA is considered to be sufficient due to the good clinical mid- and long-term results reported in the literature [10]. Current guidelines recommend additional imaging using CT-scan or MARS-MRI to rule out potential ARMD in patients with blood

cobalt ion levels > 2 μg/L [10]. However, little is known about the metal ion exposure in patients with small-head, MoM THA at long-term follow-up. The aim of this study was to report clinical and radiological results and to investigate blood metal ion levels in a large cohort of patients with small-diameter, MoM THA at long-term follow-up. The results of this study show that despite good clinical results, radiological findings of femoral osteolysis and ARMD were frequently seen in this cohort of patients with well-functioning small-head, metal-on-metal THA, and 29% of patients demonstrated elevated cobalt or chromium ion levels, i.e., > 2 μg/L, at long term follow-up.

To our knowledge, the present study represents the largest cohort of patients following small-head, MoM THA investigated with blood metal ion analysis at long term follow-up. Metal ion levels may vary significantly depending on the medium (e.g., whole blood, serum, or erythrocytes) and the technique (AAS vs. ICP-MS) used for analysis, which limits comparison among published studies [20]. Migaud et al. investigated whole blood metal ion concentrations in 26 patients following small-diameter, Metasul, metal-on-metal THA at a mean follow-up of 12 years. They reported median cobalt and chromium levels of 0.95 μg/L (range 0.4–4.8 μg/L) and 1.2 μg/L (range 0.1–5.6 μg/L), respectively [21]. Comparable results were reported by Ayoub et al., with mean cobalt ion levels of 1.85 μg/L (range 0.35–13.6 μg/L) and chromium ion levels of 1.32 μg/L (range 0.1–7.9 μg/L) at a mean follow-up of 15.9 years [22]. Our results of metal ion analysis at a mean follow-up period of 14 years are consistent with these findings, with mean metal ion levels being within the range of < 2 μg/L. However, approximately one-third of patients in our cohort demonstrated metal ion levels above 2 μg/L, and therefore, should undergo further imaging with ultrasound, CT-scan, and/or MARS-MRI in order to rule out ARMD, according to current guidelines [10]. In the study of Ayoub et al., only three patients demonstrated cobalt ion levels > 3 μg/L, and no ARMD was seen in this group of 42 female patients using ultrasound assessment. We presume that the higher proportion of patients with elevated cobalt and chromium ion concentrations seen in our study might be attributed to the larger patient cohort. The prevalence of ARMD in asymptomatic patients with small-diameter, MoM THA at long-term follow-up still is not clear, and larger cohort studies using CT or MRI should be performed to address that question. In accordance with our findings, a study by Hwang et al. investigated the prevalence of ARMD in patients following 28-mm Metasul MoM THA using computed tomography, and found ARMD to be present in 20% of the hips at a mean follow-up of 15 years [11].

For hip resurfacing and large-head, metal-on-metal THA, different risk factors for implant failure and elevated metal ion levels could be identified, such as high cup inclination angles or female sex [23,24]. Sidaginamale et al. [25] found a correlation between elevated ion levels and abnormal wear patterns in retrievals of resurfacing components. Langton et al. [24] analyzed 278 asymptomatic patients with hip resurfacing devices, and found elevated cobalt ion concentrations and female sex to be associated with early implant failure secondary to ARMD. Hart et al. [26] showed that increased blood metal ion concentrations were associated with implant failure in patients after hip resurfacing and large-diameter, metal-on-metal THA. In accordance with the findings of Lass et al. [27], we could not identify any risk factors associated with elevated blood metal ion levels in this cohort of patients with small-head, metal-on-metal implants. Impingement between the femoral neck and the Metasul liner is a known phenomenon, which can lead to increased metal wear or disassociation of the acetabular liner [28,29]. Four patients in our cohort showed radiological signs of impingement, with a visible notch at the femoral neck of the titanium stem on the lateral radiographs; titanium ion levels were increased in these patients. Therefore, titanium ion analysis can be beneficial to detect excessive wear due to impingement, in particular because it can be difficult to diagnose acetabular impingement on plain radiographs in some cases.

We found an acceptable clinical outcome for this bearing type according to the NICE recommendations [30], with a cumulative rate of implant survival of 96% with revision for any reason as the end point at 10 years. The mean patient age of 52 years at time of surgery was relatively young in this cohort. This was mainly attributed to the fact that the indication for THA in combination with a small-head, metal-on-metal bearing at that time was advanced osteoarthritis in young patients

with a high activity level, which can be considered a potential selection bias when comparing our data to other reports on implant survival. Comparable long-term results for the Metasul bearing have been reported by Lass et al. [27] (survival rate of 87% at 18.8 years) and Hwang et al. [31] (survival rate of 97.8% at 18.4 years for acetabular cup revision for any reason). However, there is concern about the high rate of proximal femoral osteolysis and ARMD, as well as the high prevalence of elevated metal ion concentrations found in this cohort. We abandoned the use of metal-on-metal articulations in favor of alternative bearings such as ceramic on highly cross-linked polyethylene, as the local and systemic long-term effects associated with metal debris and metal ion release are still not fully understood [15].

There are some limitations to our study. Five percent of patients were lost to follow-up and a further 10% declined to participate in blood metal ion analysis. Also, no CT-scans were carried out in order to avoid additional radiation exposure. As a consequence, the rate of osteolysis could have been underestimated, especially around the acetabular components. Furthermore, MRI was only performed in 66 of the 189 hips (35%) that were available for clinical and radiological assessment, which could have resulted in a selection bias regarding the prevalence of ARMD. The fact that only 66 of the 107 advised patients with elevated metal ion levels agreed to participate in MRI assessment was mainly attributed to long travel distances and/or the absence of symptoms [13]. Further studies with larger patient cohorts using CT or MRI should be performed to investigate the prevalence of ARMD in asymptomatic patients with small-diameter, MoM THA at long-term follow-up. In addition, due to the cross-sectional study design, metal ion analysis was performed at a single time point, with a mean follow-up of 14.4 years after surgery; no sequential analysis was performed for each patient. However, longitudinal studies showed that blood metal ion levels in patients with well-functioning small-head, metal-on-metal bearings did not tend to increase over time [32–34].

#### **5. Conclusions**

The present study demonstrates good clinical results for cementless, 28 mm, MoM, THA at long-term follow-up, with a cumulative survival rate of 94% after 14 years. However, increased blood metal ion levels (cobalt or chromium > 2 μg/L) were found in approximately one-third of asymptomatic patients, and proximal femoral osteolysis and ARMD were frequently seen in this cohort. Blood metal ion analysis appears helpful in the long-term follow-up of these patients in order to identify individuals at risk. In accordance with contemporary consensus statements [10], symptomatic patients with elevated metal ion levels and/or progressive osteolysis should be considered for additional CT or MARS MRI to determine the extent of soft tissue affection prior to revision surgery. Further studies are necessary to investigate the clinical relevance of ARMD in asymptomatic patients with small-head, MoM THA.

**Author Contributions:** Conceptualization, T.R. and M.C.K.; Methodology, T.R. and M.C.K.; Formal Analysis, T.R.; Investigation, T.R., M.C.K., K.S., F.H.; Resources, T.R., M.C.K., J.P.K., T.G., C.M.; Writing—Original Draft Preparation, T.R.; Writing—Review and Editing, T.R., T.G., M.M.I., M.R.S., T.A.N., B.M., J.P.K., C.M.; Visualization, T.R.; Supervision, T.R., M.C.K., T.G.; Project Administration, T.R. and T.G.; Funding Acquisition, T.R. and T.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the non-commercial research fund of Stiftung Endoprothetik (Hamburg, Germany), grant number 57,000 €, and by the Baden-Württemberg Ministry of Science, Research and the Arts and by Ruprecht-Karls-Universität Heidelberg.

**Acknowledgments:** We like to thank Thomas Bruckner PhD, for his assistance with the statistical analysis and Stefan Rheinberger for his help with blood metal ion analysis.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**

1. Weber, B.G. Experience with the Metasul total hip bearing system. *Clin. Orthop. Relat. Res.* **1996**, *329*, S69–S77. [CrossRef] [PubMed]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Titanium Acetabular Component Deformation under Cyclic Loading**

#### **Nicholas A. Beckmann 1,2,\*, Rudi G. Bitsch 3, Theresa Bormann 4, Ste**ff**en Braun <sup>4</sup> and Sebastian Jaeger <sup>4</sup>**


Received: 19 November 2019; Accepted: 17 December 2019; Published: 20 December 2019 -

**Abstract:** Acetabular cup deformation may affect liner/cup congruency, clearance and/or osseointegration. It is unclear, whether deformation of the acetabular components occurs during load and to what extent. To evaluate this, revision multi-hole cups were implanted into six cadaver hemipelvises in two scenarios: without acetabular defect (ND); with a large acetabular defect (LD) that was treated with an augment. In the LD scenario, the cup and augment were attached to the bone and each other with screws. Subsequently, the implanted hemipelvises were loaded under a physiologic partial-weight-bearing modality. The deformation of the acetabular components was determined using a best-fit algorithm. The statistical evaluation involved repeated-measures ANOVA. The mean elastic distension of the ND cup was 292.9 μm (SD 12.2 μm); in the LD scenario, 43.7 μm (SD 11.2 μm); the mean maximal augment distension was 79.6 μm (SD 21.6 μm). A significant difference between the maximal distension of the cups in both scenarios was noted (F(1, 10) = 11.404; *p* = 0.007). No significant difference was noted between the compression of the ND and LD cups, nor between LD cups and LD augments. The LD cup displayed significantly lower elastic distension than the ND cup, most likely due to increased stiffness from the affixed augment and screw fixation.

**Keywords:** total hip arthroplasty; implant deformation; acetabulum

#### **1. Introduction**

Total hip arthroplasty (THA) is considered to be one of the most successful operations performed in orthopedic surgery and the treatment of choice for end-stage osteoarthritis of the hip [1]. Consequently, the frequency of hip joint arthroplasty continues to increase worldwide. In 2007 Kurtz et al. estimated that the demand for primary total hip arthroplasty in the USA will increase by 174% to 572,000 by 2030 [2]. Concurrently they projected that the demand for revision THA would increase between 2005 and 2030 from 40,800 to 96,700 procedures representing an increase of 137% [2]. Generally, the longevity of revision THA is less than that of primary THA [3]. Lie et al. in 2004 analyzed 4762 revisions reported to the Norwegian Arthroplasty Register with a mean follow up of 3.2 years and found a 26% risk of failure after 10 years for cases without prior infection [3].

Cementless acetabular components have achieved widespread acceptance in THA as a result of their improved and reliable long-term results [4]. Primary stability is achieved through press-fit fixation that requires 1–3 mm under-reaming of the acetabular cavity and forceful impaction [5]. The forces utilized are substantial and generally result in some degree of deformation of the metal cup [6,7], particularly when the cup is thin-walled and of large diameter [8]. Meding et al. have shown that implant deformation resulting from the implantation process is non-uniform and results in diametrical pinching close to the implant rim, which is ascribed to impact pressure against the ischial and ileal columns of the acetabulum [8]. Prior studies have focused on assessing the pattern and degree of deformation resulting from the implantation process [7,9–14]. A few studies have assessed deformation after press-fit implantation and load application [15,16], however, these studies did not investigate cup deformation during load on the cup, as would be the case once the patient moves his/her hip.

The goal of our study was to evaluate and compare the elastic deformation of in-vitro titanium press-fit cups in two revision THA scenarios during subjection to cyclic gait loading for varying time periods. One sample group consisted of cadaver bone implanted with only a revision Gription cup (Pinnacle Multihole with Gription coating, Depuy/Synthes) (GC), whereas the second group was implanted with a revision Gription cup (GCS) and Gription augment (GAS) construct (Depuy/Synthes) both fixed to the pelvis with screws. We considered the diametrical change of the component rim during cyclic loading to relate to and reflect elastic deformation of the component that could potentially influence stable seating of the implant and the bone/implant apposition. This could contribute to micromotion on the implant-bone interface and between the cup and modular liner, respectively, which in turn might affect the osseointegration of the implant or influence the backside wear of the liner.

#### **2. Materials and Methods**

This study was approved by the local ethics committee (Ethikkommission Medizinische Fakultät Heidelberg, S309/2011).

Six fresh frozen cadaver hemipelvises were thawed at room temperature, dissected free of soft tissue. BMD measurements were done on attached femoral neck fragments prior to their removal from the hemipelvises. BMD was evaluated using Dual X-ray Absorptiometry (DXA) (QDR-2000 DXA densitometer; Hologic Inc, Waltham, MA, USA) and AP radiographs were also obtained on all specimens for pre-operative planning, and to exclude relevant pathology.

The average donor age was 78 years (range 51 to 96 years); the average donor mass was 66 kg (range 49 to 86 kg); and the average donor height was 171 cm (range 147.3 to 177.8 cm) resulting in a mean BMI of 23 kg/m<sup>2</sup> (range 19.0 to 28.9 kg/m2). The mean bone mineral density, BMD, was 0.8 g/cm<sup>2</sup> (range 0.641 to 0.924 g/cm2). Cadaver pelvises with a BMD below 0.6 g/cm<sup>2</sup> were excluded.

We compared two component scenarios that are frequently used in revision surgery.

#### *2.1. Scenario 1: Revision Cup without Substantial Bony Defect (ND)*

In the first scenario, a revision Gription cup (GC) (Pinnacle® Multihole with Gription® coating, DePuy Synthes Companies, Warsaw, IN, USA) was implanted press-fit in six hemipelvises without bone defects according to manufacturer recommendations as described in previous studies [17,18]. The acetabula were reamed in a concentric fashion in 2 mm increments removing all cartilage. The last reamer used was 1 mm smaller than the corresponding cup size, allowing a press-fit insertion of the cup. The cups were implanted at 45◦ of inclination and 15◦ of anteversion. Cup sizes of 50, 52, 54, 56, and 58 mm (58 mm cup was implanted in two cases) were used to correspond to the respective acetabular cavity. No additional screws were used as the press-fit allows instant stability of the cup in the bone.

#### *2.2. Scenario 2: Revision Cup with Large Bony Defect Treated Using an Augment (LD)*

In the second scenario, the same six hemipelvises received a Paprosky 2b defect of 10 mm depth, which was created in a standardized manner at the posterolateral aspect of the acetabulum. 30% of the circumference of the rim was involved and the edge of the defect bordered on the anterior inferior iliac spine. The defects were subsequently covered with a 10 mm Gription augment (GAS), which

was fixed to the host bone with two 5.5 mm × 30 mm screws, the so-called "augment-first technique". A Gription cup (GCS, Gription Cup with Screws) was then fixed to the host bone using three 6.5 mm × 30 mm screws and to the augment with one 6.5 mm × 15 mm screw. Cup diameters referred to 50, 52, 54, 56, 58 and 58 mm. No cement was used. Marathon® cross-linked polyethylene liners (DePuy Synthes Companies, Warsaw, IN, USA) corresponding to the respective cup and a 28 mm diameter metal head were used in all cases. THA's were performed by a highly trained and experienced surgeon (RGB). Post-operative radiographs were obtained and confirmed the positioning of the implants and exclude fractures in all cases.

The implanted hemi-pelvises were fixated in a container using polyurethane foam (RenCast FC 53 A/B, Goessel + Pfaff GmbH, Karlskron/Brautlach, Germany) and integrated into a custom-made multi-axial testing machine (TD, testing device). This customized multi-axial testing machine enabled us to apply the changing loads and force vectors generated during a normal gait cycle as described in prior studies [17,18].

Loads applied were taken from the Bergmann et al. OrthoLoad data set [19]. Bergmann et al. [20] divided the gait cycle into phases and for each phase load components in the x, y and z axes were given according to a defined coordinate system. Using the load data (*Fx, Fy, Fz*) and their respective angles, the orientation of the resultant force vectors during normal gait could be replicated. Using this data in our testing machine, the magnitude and direction of the force vectors were controlled by the MTS regulator (MTS headquarters, 14000 Technology Drive, Eden Prairie, MN, USA).

Utilizing our TD, our specimens were subjected to our adjusted loads in a cyclic physiologic manner that mimicked the normal gait cycle with the difference that we limited the applied load to 30% of that experienced in the normal gait cycle. We chose 30% of the normal load as an estimate of the partial weight-bearing allowed in patients during the immediate postoperative phase of revision hip arthroplasty. The loads we applied varied from a maximum of 69.93% to a minimum of 8.71% of body weight compared to the respective values at 233.1% and 29.02%, which have been determined with full weight bearing in the normal gait cycle [20,21]. One thousand load cycles were carried out at 1 Hz.

Marker points (size 0.8 mm, GOM Gmbh, Braunschweig, Germany) trackable by an optical measuring system were placed around the circumference of the cup rim. They were also placed along the rim of the augment that constituted 30% of the entire cup circumference (see Figure 1).

**Figure 1.** Photograph of acetabular cup, augment and liner. Note the attached optical marker points along the rim of the cup and augment.

The measurements were taken at 0, 50, 100, 200, 400, 600, 800, and 1000 cycles using a frame rate of 15Hz. For each set of load cycles, we used the optical readings from the cup rim and augment marker points, respectively, to calculate best-fit circles by utilizing the Best Fit Algorithm that then enabled us to measure the maximal and minimal circle diameters at each set of load cycles (see Figure 2) [22,23]. We considered elastic deformation to be the change in diameter of these "best-fit" circles from the circle diameter measured initially after implantation without load, which was calculated from the positional changes of the rim markers. Mean maximum and minimum diameter changes (mean maximum and minimum deformation) were calculated for both GC (measurement 1), and for GCS and GAS (measurement 2) construct groups for each of 0–1000 cycle sets. The results for the three groups were then compared.

**Figure 2.** Representation of best-fit circle, denoted by the dotted lines around the rim of the acetabular cup, which was calculated by the relative motion of the optical markers.

#### **3. Statistics**

The data were evaluated descriptively using the arithmetic mean, standard deviation, minimum and maximum. A repeated-measures analysis of variance (ANOVA) was performed to test for significant differences for the parameter of deformation (primary cup vs. cup with screws and cup with screws vs. augment). Prior to data analysis, the normal distribution of the data was evaluated using a Shapiro–Wilk test, which was chosen over other statistical tests of normal distribution since a prior study has shown that this test provides more power given a known significance than other tests of normal distribution [24]. Subsequently, the homogeneity of variance was verified using the Levene test, which is a prerequisite for ANOVA. The results allowed for the use of the ANOVA test. The Greenhouse–Geisser adjustment was used to correct for violations of sphericity. The data were analyzed using SPSS 25 (IBM, Armonk, New York, NY, USA).

#### **4. Results**

The results are displayed below in Table 1 and Figure 3.

**Table 1.** Mean elastic deformation (μm) for scenario 1 (GC) and scenario 2 (GCS and GAS) during all cycle sets.


**Figure 3.** Bar graph of mean elastic deformation (distension and distension) of GC (no bony defect; scenario 1) and GCS and GAS (with bony defect; scenario 2).

Comparison of the compression and distension deformation between primary and revision cup:

The cyclical loading showed no statistically significant effect under the standardized loading conditions with regards to the compression deformation, F(3.027, 30.266) = 0.484, *p* = 0.698. There was no statistically significant difference for the compression cup deformation between the primary and revision groups, F(1, 10) = 1.740, *p* = 0.217, yet a tendency for higher elastic deformation of the ND cup (GC) can be found.

The duration of load (i.e., the later load cycles) had no significant influence on the distension deformation, F(1.718, 17.176) = 0.368, *p* = 0.666. The repeated measures ANOVA with Greenhouse-Geisser correction determined that the difference in distension deformation of the cups in scenario 1 (No Defect) and scenario 2 (Large Defect) was statistically significant, F(1, 10) = 11.404, *p* = 0.007.

Comparison of the compression and distension deformation between the revision cup and augment:

Under the standardized loading conditions the cyclical loading showed no statistically significant effect with regards to the compression deformation, F(1.889, 18.892) = 1.048, *p* = 0.367. There was no statistically significant difference for the compression deformation between the revision cup and augment, F(1, 10) = 0.015, *p* = 0.904.

The duration of load (the later load cycles) had no significant influence on the distension deformation, F(1.515, 15.152) = 1.104, *p* = 0.340. The repeated measures ANOVA with Greenhouse-Geisser correction determined that there was no statistically significant difference between LD cup (GCS) and augment (GAS) with regards to the distension deformation of each, F(1, 10) = 0.389, *p* = 0.547.

The results are displayed in Table 1 and Figure 3.

#### **5. Discussion**

A large body of existing research has played an important role in the current success of both primary and revision THA, and resulted in a steadily increasing number of procedures and a decreasing age of patients [25,26]. Loosening and dislocation have been identified as the major causes of implant failure [26–28] and have been the focus of most research. Micro-motion at the bone/implant interface with subsequent particle production, tissue reaction, and osteolysis has been well documented [17,29–32]. Primary stability of the implant is recognized as critically important for ultimate surgical success [33–35]. Early osseointegration of all porous metal implants requires only minimal relative motion between implant and host bone. Micro-motion or bone/implant gap size of up to 50 μm has been shown to result in successful osseointegration, but above 150 μm there is attachment by fibrous tissue [33,36–38].

In contrast to studies of micromotion and other contributing factors leading to implant failure, implant deformation has received relatively little research attention. Existing studies have focused almost exclusively on deformation changes that occur during the process of implantation, and have disregarded deformation occurring afterward. The deformation of press-fit acetabular cups into 1–3 mm under-reamed sockets has been documented in several prior studies [8,37]. Prior researchers noted that the acetabular bone is most dense at the anterosuperior and posteroinferior margins (ileal and ischial columns) and the pubic area, constituting 3-point support [28]. The ileal and ischial columns are the most unyielding foci of the acetabular rim during the implantation process [12,28], and post-implantation provide the most bone/implant contact, foci of greatest load transfer and ultimately the most support for the implant [6,28]. Studies have shown that contact at the pelvic rim/implant area constitutes 25–50% of the total host bone/ implant apposition [9,36]. The host bone implant apposition decreases from rim to pole [39]. It has been found that even under ideal circumstances, total bone/implant apposition is never achieved and is unnecessary for successful osseointegration [9]. Cadaver studies of previously well-functioning components have shown that bone/implant contact areas have varied widely in extent [36]. The pelvic rim is also the area of greatest implant support during weight-bearing/gait loading [28].

During forceful press-fit implantation (previously measured at 400 N with porous titanium acetabular cups in-vivo) [8,37], the ileal and ischial columns exert a pinching effect on the cup component [7,9,12,14,35,37]. This causes the cup to assume a hemi-elliptical rather than a hemispherical shape that can lead to incongruity, diminished apposition and increased gap areas at the bone/implant interface [9,11,35]. These gaps, if excessive, can be associated with a number of negative consequences, including improper depth and angle of implant seating with subsequent potential dislocation, increased micro-motion at the bone/implant interface [35], facilitation of particle accumulation and increased tissue fluid, impaired liner insertion secondary to distortion of the cup locking mechanism and diminished clearance that adversely affects joint lubrication and liner wear [8,9,11,35,37,40]. Other factors that have been shown to influence the degree of cup deformation are the reaming process prior to implantation and the geometry of the acetabular cavity, characteristics of the implanted cup, bone density/hardness, the force applied during implantation and the seating of the cup [8,9,37,41,42]. Prior studies have shown that manual reaming results in a cavity that is usually slightly larger than the last reaming instrument and is hemi-elliptical in shape as a result of the varying degrees of stiffness throughout the host acetabular cavity [7,9,14,42]. Lin et al. found large errors in hand-reamed cavities [34]. Therefore, careful reaming of the acetabular cavity and accurate cup seating have been identified as significant modifiable factors in reducing the degree of implant deformation [9] that occurs during the implantation process, and we hypothesize that they may also influence the degree of elastic deformation that is the focus of our study. Characteristics of the implant that have been found to influence the potential to deform are the type of metal used, and diameter and thickness of the cup. Meding et al. found that titanium cups deform more than those of CoCr, and increasing the diameter and thinness of the wall are associated with increased potential to deform [8,35].

The focus and methodology of our current study differ in several respects from all prior studies of deformation. Our study does not evaluate deformation occurring from the process of implantation, but looks at elastic deformation occurring as a result of cyclic loading applied several days post-implantation. We utilized cadaver bone rather than the synthetic bone to permit the most realistic testing scenario, and measurements were made several days after implantation to minimize the elliptical distortion of the implantation process since it has been demonstrated that for several days post-implantation there is a visco-elastic relaxation of the pelvic bone and implant that reduces the deformation of titanium components [7]. We focused on the rim since it has been shown to be the site of the most bone/implant apposition and the region of maximal loading with gait [6,28]. Since our methodology utilized the derivation of a "best-fit circle" based on an assumption of symmetry of the

acetabular implant, our results are not direct measurements of implant deformation or of gap size and can only be used for comparison of our two study groups.

Our results showed that all our study specimens displayed some elastic deformation during cyclic loading, but there was a distinct difference between the two groups. We found statistically significantly more elastic deformation in Scenario 1 (revision cup only, lack of bony defect) than in Scenario 2 (revision cup plus augment and screws implanted into a Paprosky 2b defect). Values for the revision cup with screws and values for the augment with screws were not significantly different from each other, indicating that cup plus augment and screws tend to function as a unit.

For all specimens, maximal elastic deformation (compression and distension) was reached after approximately 50 cycles of loading with no significant additional deformation noted with increased cycles.

We hypothesize that the addition of screws and screws plus augment effectively increases the rigidity of the construct, which is consistent with findings of prior studies on deformation that occurred secondary to implantation, which showed that rigid CoCr cups deformed less than titanium cups [8]. Additionally, the compression of the cup and augment in scenario 2 proved more substantial than the distension.

Our results are clinically significant in several ways. The use of ancillary screws decreased the degree of elastic deformation of the implant rim, allowing better bone/implant apposition, reduced gap areas, and potentially improved osseointegration. The presence of an augment did not negatively impact the degree of elastic deformation, and the cup/augment construct effectively functioned as a unit. Our findings also indicate that cyclic loading that mimics normal gait is associated with increased elastic deformation and validates prior findings [9,35]. We propose that the elastic deformation we identified, and the deformation of implantation may both be potential influencing factors in the occurrence of backside wear of modular liners. Furthermore, acetabular component deformation during load may affect clearance at the articulating interface and may play a role in the rare, albeit relevant liner dislocation. One case series reported on 23 liner dissociations after Pinnacle implantation [43], which may be a result of overly elastic cups.

Cup diameter and wall thickness have repeatedly been shown to affect cup deformation. Large (jumbo) porous cups are an alternative option to cup/augment combination in revision THA associated with deficient bone stock and acetabular defects, and their large diameter and thin walls predispose to increased deformation, although ancillary screw fixation may limit this deformation. Oversized porous cups have been reported to have an increased risk of dislocation of multifactorial cause [6,8,44]. Deformation of the cups has not been considered a contributing factor but may be worthy of further consideration.

#### **6. Limitations**

Our study has several limitations. The cohorts were small, and only two revision constructs were assessed, the first without substantial osseous defect, and the second using a substantial (Paprosky 2b) bony defect.

The optical markers covered only half the implant rim, since the remaining rim was concealed from the optical cameras by the moving prosthesis neck—we assumed that the cup rim was symmetrical. Our methodology utilized the optical marker data to derive a "best-fit circle" and the change of circle diameter was considered to reflect elastic deformation. Therefore, our results are not a direct measure of deformation or bone/implant gap regions, but serve only as an indirect indicator, and can only be used as a means of comparison between our two study groups. This is the first study to utilize this methodology and evaluate deformation during loading, and there are therefore no currently existing comparable data for verification.

We utilized cadaver bone since the polyurethane models were validated only for deformation occurring during implantation [8] and we considered cadaver bone to be more physiologic. However, it lacks some of the viscoelastic properties of live bone, and may not reflect the clinical scenario [12,36].

There is tentative evidence that hardness of bone, related to BMD, necessitates increased compressive force for implantation that may affect results. Our cadaver sample was from older patients with a mean BMD of 0.8g/cm<sup>2</sup> (range 0.641 to 0.924 g/cm2) and values beyond this range may have different results.

In our experimental set-up, only Paprosky 2b defects were examined, and defects of other grades may have different results.

Multi-hole cups have been shown to present more deformation than single-hole cups during implantation [16], which allows one to assume that a similar effect may be seen under dynamic loading after implantation. For the purpose of comparability, a multi-hole cup was used in both scenarios, although this type of cup would tend not to be used clinically in a scenario without a bony defect.

Variations in surgical technique cannot be excluded although the same highly experienced surgeon (RGB) performed all implantations with supplied manufacturers tools.

In conclusion, our in-vitro study utilizing revision constructs in cadaver bone is the first to compare the elastic deformation occurring at the rim of two different implant constructs during cyclic loading that replicates the limited loading of normal gait as experienced in the early postoperative period. Our results show that the use of adjunctive screws significantly decreases the degree of elastic deformation under these conditions, and the inclusion of an augment does not adversely impact the degree of elastic deformation.

**Author Contributions:** Conceptualization, N.A.B., R.G.B. and S.J.; Data curation, N.A.B., R.G.B., S.B. and S.J.; Formal analysis, N.A.B., T.B. and S.J.; Funding acquisition, S.J.; Investigation, N.A.B., S.B. and S.J.; Methodology, N.A.B., R.G.B. and S.J.; Project administration, N.A.B. and S.J.; Supervision, S.J.; Validation, N.A.B. and S.J.; Writing—original draft, N.A.B.; Writing—review & editing, N.A.B., R.G.B., T.B., S.B. and S.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** Funding for this project was provided by the University of Heidelberg, and the Laboratory of Biomechanics and Implant Research.

**Conflicts of Interest:** The authors declare no pertinent conflict of interest. N.A.B. reports research grants from Johnson & Johnson DePuy Synthes that are not related to the current study. R.G.B. and S.J. report grants from B Braun Aesculap, Johnson & Johnson Depuy Synthes, Heraeus Medical, Waldemar Link, and Zimmer Biomet that are not related to the current study. In addition, S.J. reports grants from Peter Brehm GmbH, which is also unrelated to the current project. Both T.B. and S.B. report no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
