Performance of metal-on-metal (MOM) bearings was of great interest until recently. Major concerns emerged over high incidence of MOM-wear failures and initially there appeared greater risks with MOM total hip arthroplasty (THA) designs compared to resurfacing arthroplasty (RSA). Impingement of the metal neck against the THA cup was likely the differentiating risk. There is a major difference between RSA and THA in (i) size of femoral necks and (ii) risk of THA metal necks impinging on metal cups. For example, a 46mm THA with 12.5mm neck, a 3.68 head:neck (H/N) ratio, provides a suitably large range-of-motion (ROM). In contrast, an RSA patient with retained 31mm size of natural neck would only have H/N = 1.48, indicating even less ROM than a Charnley THA. However, the enigma is that RSA patients have as good or better ROM in majority of clinical studies. We studied this apparent RSA vs THA dilemma by examining MOM retrievals for signs of adverse impingement. We previously described CoCr stripe wear in failed THA bearings, notably alignment of polar and basal wear stripes coincident with the rim profiles of the cups (Clarke 2013). Our governing hypothesis was that RSA patients had to routinely sublux their hips to get ROM comparable to THA. Our THA impingement studies showed polar stripes within 15o of the polar axis in large heads. For the various RSA diameters, we calculated that wear stripes angled 40o from the femoral axis could indicate impingement with no subluxation, whereas smaller angles would indicate routine subluxation of RSA femoral-shell from cup. We compared explanted RSA (N=15) and THA (N=15) bearings representing three vendors (42–54mm diameters). Wear maps and head-stripes were ink-marked for visualization, photography, and analysis. Wear areas were calculated using spherical equations and wear-stripe angles measured by computer graphics.

The results showed that RSA femoral shells had wear areas circular in shape with areas varying 1,085- 3,121mm2. These averaged 14% larger than in matched THA heads but statistically significant difference was not proven. Polar stripes were readily identifiable on femoral components, 75% for RSA cases and 100% for THA. These contained identical linear scratches and all were sited within 30o of neck axis, confirming our hypothesis that RSA patients had to sublux their hips to achieve same motion as THA. Examination of cup wear areas revealed all showed ‘edge-loading’, but RSA cups had a significantly greater degree.

Retrieval studies are limited by uncontrolled case sources, varied brands, and small numbers. In this study, we were able to match RSA and THA cases by vendor and diameter. The RSA retrievals revealed polar stripes identical to THA by site, topography and inclination to femoral-neck axis. This confirmed our starting hypothesis and explained the large clinical ROM available in RSA patients. The larger wear areas on RSA femoral shells, although not statistically significant, and the larger ‘edge loading’ sites in RSA cups appeared as further support for routine subluxation of femoral-shells during hip impingement.

This study presents an unusual recurrent case of pigmented villonodular synovitis (PVNS) around a ceramic-on-metal (COM) hip retrieved at 9-years. PVNS literature relates to metal-polyethylene and ceramic-ceramic bearings. Amstutz reported 2 cases with MOM resurfacing and Xiaomei reported PVNS recurring at 14 years with metal-on-polyethylene THA. Friedman reported on PVNS recurrence in a ceramic THA. Ours may be the first reported case of recurrent PVNS of a ceramic-on-metal articulation.

This young female patient (now 38-years of age) had a total hip replacement in 2006 for PVNS in her left hip. In her initial work-up, this case was presumed to be a pseudotumor problem, typical of those related to CoCr debris with high metal-ion concentrations. She had an CoCr stem (AML), 36mm Biolox-delta head (Ceramtec), and a Pinnacle acetabular cup with CoCr liner (Ultramet, Depuy J&J). This patient had no concerns regarding subluxation, dislocation or squeaking. Three years ago she complained of mild to moderate groin and thigh pain in her left hip. This worsened in the past year. She noticed increased swelling now with an asymmetry to her right hip. She went to the emergency room in Dec-2014 and was referred to a plastic surgeon. In our consult we reviewed MARS-MRI and CT-scans that demonstrated multiple mass lesions surrounding the hip. Laboratory results presented Co=0.7, Cr=0.3 ESR=38 and Crp=0.3.

At revision surgery, the joint fluid was hemorrhagic/bloody with hemosiderin staining the soft tissues. Multiple large 4–5×5cm nodules were present in anterior aspect of the hip as well as multiple nodules surrounding posterior capsule and sciatic nerve. Pathology demonstrated a very cellular matrix with hemosiderin-stained tissue and multiple giant cells, which was judged consistent with PVNS. The trunnion showed no fretting, no contamination and no discoloration. The superior neck showed impingement due to low-inclination cup. There was minimal evidence of metal-debris staining the tissues. There was a large metallic-like stripe across the ceramic head.

This is a particularly interesting case and may be the first reported recurrent PVNS around a ceramic-on-metal bearing (COM). Data is scant regarding clinical results of COM bearings and here we have a nine-year result in a young and active female patient. She was believed to have a metalosis-related pseudotumor yet her metal-ion levels were not alarmingly high and there was no particular evidence of implant damage or gross wear products. In addition, the CoCr trunnion appeared pristine. Our work-up continues with analyses of wear and histopath-evidence. This case may demonstrate the need for a broadening of the differential diagnosis when dealing with hip failures.

Metal-on-metal retrieval studies indicated that MOM wear-rates could rise as high as 60–70mm3/year in short-term failures (Morlock, 2008). In contrast, some MOM and ceramic-on-ceramic (COC) devices of 1970's era performed admirably over 2–3 decades (Schmalzreid, 1996; Shishido, 2003). While technology has aided analysis of short-term MOM and COC failures (Morlock 2008; Lord 2011), information on successful THA remains scant. Lack of long-term data creates difficulties in setting benchmarks for simulator studies and establishing guidelines for use in standards. In this study we compared clinical and wear histories for a 30-year MOM and a 32-year COC to establish such long-term, wear-rates.

The McKeeTM retrieval was cemented and made 100% of CoCr alloy (Fig. 1a). This patient had a right femoral fracture at 47 years of age, treated by internal-fixation, which failed. Her revision with a Judet implant also failed, leaving her right hip as a Girdlestone. At the age of 68, she had a McKee THA implanted in left hip, and used it until almost 98 years of age (Campbell, 2003). The COC case was a press-fit AutophorTM THA, head and cup made of alumina ceramic, with the only metal being the CoCr stem (Fig. 1c). This was implanted in a female patient 17-years of age active in sports (water-skiing). This modular THA was revised 32-years later due to hip pain from cup migration. Wear on these implants was identified by stereomicroscopy and stained red for photography (Fig. 1). Cup-to-neck impingement was denoted by circumferential neck notching, roughness was assessed by interferometry, and wear determined by CMM (Lord, 2011).

McKee head wear covered 1092mm2 area (Figs. 1a, 2: hemi-area ratio 58%). There was no stripe wear and head roughness was 36nm (Ra). Cup wear covered an area of 1790mm2 (hemi-area 63%). Circumferential damage was noted on the supero-posterior femoral neck with scuff marks also on posterior collar (Fig. 2c). Head and cup wear amounted to 37.7 and 25.2mm3, respectively. Total MOM wear was 62.9mm3, indicating a wear-rate of 2.1mm3/year.

Ceramic head wear consisted of two circular patterns (Fig. 1c), the major one of area 1790mm2 (hemi-area 79%). No wear stripes were identified. Non-worn and extensively worn surfaces had roughness (Ra) 17nm and 123nm, respectively. The cup showed 360o circumferential arc of rim wear with a small, non-wear zone inferiorly (Fig. 1c). Gray metallic transfer was evident, EDS revealing Co and Cr (Fig. 3a). Head and cup wear volumes were 77.2 and 54mm3, respectively. Total COC wear amounted to 131.2mm3 indicating a wear-rate of 4.1mm3/year.

These two THA functioned successfully over 3 decades. The McKee retrieval had minor signs of impingement but no adverse “stripe wear”. This MOM performed satisfactorily due to good positioning and patient's advanced age (68 to 98Yrs of age). The COC patient was 17 years of age at index surgery and active. The ceramic cup showed 360o of edge wear, CoCr transfer and a COC wear-rate double that of the MOM retrieval. Thus the high ceramic wear-resistance protected this youthful patient.

The purpose of this study was to determine the survivorship for a MOM implant series performed by a single community surgeon followed using a practical clinical model. A retrospective cohort of 104 primary MOM THA procedures (94 patients) were all performed by one surgeon at three local hospitals now with 10–13 years follow-up. Sixteen patients are deceased and 16 patients have been lost to follow-up. In the remaining 62 patients, 8 are bilateral providing a total of 70 THA for study. The clinical follow-up model included: hip scores, X-rays, ultrasound, and metal ion concentrations (Co, Cr, Ti). Due to the diversity of patient location, a variety of clinical labs were utilized for metal ions. Statistical methods included Kaplan-Meier survival curve and One-way ANOVA. Hip scores were available for 70 THA and of these 61 had a hip score (HHS) above 80 (87%). X-rays were available for 49 hips and of these 38 (78%) had lateral/version angles in the safe zone (Fig 1: inclination ≤ 55 and anteversion ≤ 35). Thirty-eight ultrasound exams were performed and of these three yielded fluid collections (8%). Metal ion concentrations were documented in 39 of 62 patients (63%, either serum or whole blood). Six outliers were identified with high concentrations of metal ions (Fig 2); Co 0.3–143.9 ppb (median 3.6), Cr 0.2–200.3 ppb (median 2.2) and Ti 2-110 ppb (median 54). Six patients were revised by the original surgeon. Three of six with elevated ions were documented as wear problems and the other three were revised for infection, femur fracture and metal-ion sensitivity. The survivorship of 92.5% at 10 years (Fig. 3) may be partly due to the exclusive use of antero-lateral approach performed by one surgeon with 78% of cups well placed and the MOM design used exclusively.

Retrieval studies of metal-on-metal (MOM) resurfaced hips revealed cup “edge wear” as a common failure mechanism [Morlock-2008]. Retrieval analysis of total hip arthroplasty (THA) also demonstrated extensive rim wear (Fig. 1: 190–220o arcs), typically across the superior cup [Clarke-2013]. Such wear patterns have not been demonstrated in hip simulator studies. The simulator “steep cup” models typically had motion arcs (flexion, etc.) input via the femoral head [Leslie-2008, Angadji-2009]. With fixed-inclination cups this produces constant loading of cup rim against the head (Fig. 2a). This is unlikely to be the physiological norm, unless patients walk constantly on the rims of mal-positioned cups. More likely the patients produce edge-wear intermittently due to functional and postural variations. Therefore a novel simulator model is proposed in which the cup undergoes edge-wear intermittently at one extreme of flexion (Fig. 2a). Our study objective using this new simulator model (Fig. 2a, b) was to (i) demonstrate MOM wear-rates and wear patches as a function of these dynamic-inclinations (40 o, 50 o, 70o), and (ii) compare the simulator data to MOM retrievals (Fig. 1).

Two simulator studies were run, both using 60mm MOM. Four bearings were run to 1-million cycles (1Mc) with cups peaking at 40 and 50° dynamic-inclinations, thus providing control data with no edge-wear. In 2nd study, 4 MOM were run with cups given a dynamic-inclination of 70° to produce edge-wear effects. In study-2 currently at 2.5Mc duration, the femoral heads showed the two classical wear phases with run-in at 1.7mm³/Mc and steady-state at 0.084mm³/Mc (Fig. 3a). Wear-rate for cups at 2.34mm³/Mc was 40% higher than heads and continued to rise linearly with time (Fig. 3a). At 2.5Mc, cup wear averaged ×5.7 greater than heads and resulting wear-patterns extended 85°−225° around cup rim (Fig. 3b: average 151°). In study-1, wear patches in cups with 40° dynamic-inclination approached within 12.4mm of the cup rim as denoted by circumferential grooves. This margin-of-safety (MOS) represented a 24°angle. The cup wear-patch averaged area of 1,760mm2. With cups run at 70o dynamic-inclination, the wear patches were transferred an additional 30o towards the rim thereby representing a 6° transfer across the rim.

This is the 1st wear study to use the new dynamic-inclination test mode to better simulate cup function in vivo. It was particularly satisfying to see the similarity in wear-patterns between retrieval (Fig. 1) and simulator cups (Fig. 3b). It is also the 1st study to monitor sites and magnitudes of cup wear areas and to purposely produce “edge wear”. The cups with 40° and 50° dynamic-inclinations had large margins of safety. With 70° dynamic-inclination the margin of safety was lost - effectively there was a 6° transfer of the wear patch across the cup rim. Even this apparently small effect at one location in each gait cycle sufficiently perturbed MOM performance that wear increased by an order of magnitude. Notably this was all cup wear and not by femoral head participation. The study continues but at 2.5Mc duration the cups revealed 5-fold greater wear than heads.

Metal-on-metal (MOM) retrieval studies have demonstrated that CoCr bearings used in total hip arthroplasty (THA) and resurfacing (RSA) featured stripe wear damage on heads, likely created by rim impact with CoCr cups.^1-3 Such subluxation damage may release quantities of large CoCr particles that would provoke aggressive 3^rd-body wear. With RSA, the natural femoral neck reduces the head-neck ratio but avoids risk of metal-to-metal impingement (Fig. 1).⁴ For this study, twelve retrieved RSA were compared to 12 THA (Table 1), evaluating, (i) patterns of habitual wear, (ii) stripe-wear damage and (iii) 3^rd-body abrasive scratches. Considering RSA have head/neck ratios much lower than large-diameter THA, any impingement damage should be uniquely positioned on the heads.

Twelve RSA and THA retrievals were selected with respect to similar diameter range and vendors with follow-up ranging typically 1–6 years (Table 1). Patterns of habitual wear were mapped to determine position in vivo. Stripe damage was mapped at three sites: polar, equatorial and basal. Wear patterns were examined using SEM and white light interferometry (WLI). Graphical models characterized the complex geometry of the natural femoral neck in coronal and sagittal planes and provided RSA head-neck ratios.⁴

Normal area patterns of habitual wear were similar on RSA and THA bearings. The wear patterns showing cup rim-breakout proved larger for RSA cups than THA. Polar stripes presented in juxtaposition to the polar axis in both RSA and THA (Fig. 1). As anticipated, basal stripes on RSA occurred at steeper cup-impingement angles (CIA) than THA. The micro-topography of stripe damage was similar on both RSA and THA heads. Some scratches were illustrative of 3^rd-body wear featuring raised lips, punctuated terminuses, and crater-like depressions (Fig. 2).

Neck narrowing observed following RSA procedures may be a consequence of impingement and subluxation due to the small head-neck ratios. However, lacking a metal femoral neck, such RSA impingement would not result in metal debris being released. Nevertheless it has been suggested that cup-to-head impingement produced large CoCr particles and also cup “edge wear” as the head orbits the cup rim.⁴ Our study showed that impingement had occurred as evidenced by the polar stripes and 3^rd-body wear by large hard particles as evidenced by the wide scratches with raised lips. We can therefore agree with the prior study, that 2-body and 3rd-body wear mechanisms were present in both RSA and THA retrievals.

Introduction

Over 40-years the dominant form of implant fixation has been bone cement (PMMA). However the presence of circulating PMMA debris represents a 3rd-body wear mechanism for metal-on-polyethylene (MPE). Wear studies using PMMA slurries represent tests of clinical relevance (Table 1). Cup designs now use many varieties of highly-crosslinked polyethylene (HXPE) of improved wear resistance. However there appears to be no adverse wear studies of vitamin-E blended cups.1–4 The addition of vitamin E as an anti-oxidant is the currently preferred method to preserve mechanical properties and ageing resistance of HXPE. Therefore the present study examined the response of vitamin-E blended liners to PMMA abrasion combined with CoCr and ceramic heads. The hip simulator wear study was run in two phases to compare wear with, (i) clean lubricants and (ii) PMMA slurries.

Introduction

Looking for optimal solutions to wear risks evident in total hip arthroplasty (THA), silicon nitride ceramic bearings (Si₃N₄) are noted for demanding high-temperature applications such as diesel engines and aerospace bearings. As high-strength ceramic for orthopedic applications, Si₃N₄ offers improved fracture toughness and fracture strength over contemporary aluminas (Al₂O₃). Our pilot studies of Si₃N₄ in 28mm diameter THA showed promising results at ISTA meeting of 2007.¹ In this simulator study, we compared the wear resistance of 40mm to 28mm diameter Si₃N₄ bearings.

The 28mm and 40mm bearings (Fig. 1) were fabricated from Si₃N₄ powder (Amedica Inc, Salt Lake City, UT).¹ Wear tests run were run at 3kN peak load in an orbital hip simulator (SWM, Monrovia, CA) and. The lubricant was standard bovine serum (Hyclone: diluted to 17 mg/ml protein concentration). Wear was measured by gravimetric method and wear-rates calculated by linear regression. SEM and interferometic microscopic was performed at 3.5-million cycles (3.5Mc) to 12Mc.

The simulator was run to 3.5Mc duration with no consistent weight-loss trends. The bearings could show either small positive or negative weight fluctuations in an unpredictable manner (Fig. 2). Surface analysis showed protein layers up to 3μm thick, furrowed due to abrasion by small particles (Fig. 3). The low ceramic wear was camouflaged by protein contaminants alternatively forming and shedding. From 3.5 to 12.8Mc duration we experimented with various detergents and wash-procedures, all to no avail. Protein coatings were also more prevalent on 44 mm heads, likely due to frictional heating by the larger diameter effect. Selected heads were washed with a mild acid solution - the cumulative effect appeared to be removal of some protein layers, but not in a predictable manner.

The Si₃N₄ ceramic is used in demanding industrial applications and it is therefore unfortunate that we are yet not able to quantify the actual wear performance of Si₃N₄/ Si₃N₄ bearings (COC). The contaminating protein layers combined with low-wearing silicon nitride obscured the actual wear data. This has also been a problem in prior studies with alumina and zirconia bearings. Considerable challenges still stand in the way of the optimal biomaterials choices that will result in reduced risk of failure while providing extended lifetimes. Thus important issues remain unsolved and call for innovative solutions. Searching for a more effective ‘wear-measurement’ remedy, we noted that abrasive slurries of bone cement (PMMA) used in contemporary simulator studies were effective in promoting adverse wear in polyethylene bearings. These investigations also revealed that PMMA debris did not damage CoCr heads^2,3, alumina heads^4,5 or diffusion-hardened zirconia heads (ZrDH).⁶ We can therefore speculate at this ISTA meeting of 2014 that future ceramic wear tests should incorporate PMMA slurries. Here a new hypothesis can be formulated, that PMMA particulates will provide a continual and beneficial removal of contaminating proteins from the ceramic surfaces (see Fig. 3) and thereby aid definition of low-wearing COC bearings such as Si₃N₄.

The application of non-oxide ceramics such as silicon nitride presented here may become a viable alternative for THA designs of next decade.

Introduction

Hip simulators proved to be valuable, pre-clinical tests for assessing wear. Preferred implant positioning has been with cup mounted above head, i.e. ‘Anatomical’ (Figs. 1a-c) ^1,2 while the ‘Inverted’ test (cup below head) was typically preferred in debris studies (Figs. 1d-f).^3,4 In an Anatomical study, wear patterns on cups and heads averaged 442 and 1668 mm² area, respectively, representing 8% and 30% of available hemi-surface (Table 1), i.e. the head pattern was ×3.8 times larger than cup. This concept of wear patterns is illustrated well in the ‘pin-on-disk’ test (Fig. 1) in which the oscillating pin has the ‘contained’ wear area (CWP) and the large wear track on the disk is the ‘distributed’ pattern (DWP). Hip simulators also create CWP and DWP patterns, site dependant on whether Anatomical (Fig. 1a-c) or ‘Inverted’ (Fig. 1d-f) test. However there is scant foundation as to clinical merits of either test mode. Retrieval studies of MOM bearings have indicated that cups have the larger wear patterns, i.e. contrary to simulator tests running Anatomical cups (Table 1).⁵ Therefore we compared Anatomical and Inverted cup modes using 38mm and 40mm MOM in two 5-million cycle simulator studies.

A 35-year-old female (age 35Yrs) had primary MOM total hip arthroplasty (THA) in 2008. At 8 months this patient postoperatively developed headaches, memory loss, vertigo, and aura-like symptoms that progressed to seizures. At 18 months review, she complained of progressive hip pain, a popping sensation and crepitus with joint motion. This patient weighed 284lbs with BMI of 38.5. Radiographs revealed the cup had 55° inclination, 39° anteversion (Fig. 1). Metal ion concentrations were high (blood: Co=126 mcg/L, Cr= 64mcg/L). Revision was performed in November 2010 A dark, serous fluid was observed, along with synovitis. The implants were well fixed and the femoral head could not be removed; thus the stem was removed by femoral osteotomy. With the head fused on this femoral stem, for the 1^st time it was possible to precisely determine the habitual patterns of MOM wear relative to her in-vivo function. We investigated (1) size and location of wear patterns and (2) signs of cup-stem impingement to help explain her symptoms developed over 32 months follow-up.

The retrieved MOM was a Magnum™ with head diameter 50mm and 50×56mm cup (Biomet). This was mounted on a Taperloc™ lateralized porous-coated stem. Components were examined visually and wear damage mapped by stereo-microscopy, interferometry, CMM, SEM, and EDS. Main-wear zone (MWZ) areas were calculated using standard spherical equations¹ and centroidal vectors determined.

The head-cup mismatch was 427um with the cup revealing a form factor of 228um. The cup showed wear area of 1275mm² that extended up to the cup rim over 150°arc. The cup rim was worn thin over a 90° arc with loss of cup bevel. The head showed an elliptical wear area of 2200mm² located centrally on the superior-medial surface (ellipsoidal ratio ×1.2). Compared to the hemispherical surface (50mm: hemi-area = 3927mm²), the worn area represented hemi-area ratio of 56%. The centroidal vectors measured 8° anterior and 24° superior to the head's polar axis (Fig. 2). Stripe wear damage revealed multiple impingement sites. SEM and EDS revealed stripes were contaminated by metal transfer from the stainless-steel instruments used at revision. The main impingement position was identified (Fig. 3) indicating the site of repetitive subluxations whereby the subluxing head thinned the cup, i.e. “edge wear”.

Cup and head wear patterns corresponded well, reinforcing our definition of the MWZ locations in vivo. The femoral MWZ was centrally located superiorly and medially with respect to the polar axis of the femoral neck and head. The noted impingement position indicated this patient had experienced repetitive subclinical subluxations (RSS).² The taper inside the fused head may also have been a contributory factor that we cannot ignore. Nevertheless her excessive cup thinning was likely a result of a steep cup and considerable anteversion allowing the femoral head to sublux over the cup rim, thus thinning the cup and wearing the rim bevel, and producing MOM wear debris.