Summary

Corrosion and fretting damage at the head-neck interface of artificial hip joints is more severe with larger head sizes. This is a concern, as the release of metal particles and ions can cause adverse tissue reactions, similar to those observed high wear metal-on-metal articulations.

Introduction

This retrieval analysis study consisted of two goals. The first goal was to determine if there was a difference in the corrosion and fretting damage along the taper interface between large femoral heads in comparison to monopolar hemiarthroplasty heads. The second goal was to examine if the diameter of monopolar hemiarthroplasty heads can influence corrosion and fretting damage along the taper interface.

The wear particles released from the polyethylene (PE) tibial insert of modular total knee replacements (TKRs) have been shown to cause wear particle induced osteolysis, which may necessitate revision surgery [1]. Wear occurs at the backside surface of the PE insert of modular TKRs, resulting from the relative movement between the PE insert and the tibial tray [2]. Wear particles generated from the backside surface of the PE insert have been shown to be smaller in size than those originating from the articular surface [1], and may therefore have increased biological activity and osteolytic potential [3-4]. The ability to predict backside micromotion and contact pressure by finite element simulation has previously been demonstrated by O'Brien et al. [6-7]. Although the effect of insert thickness on articular surface contact pressure has been investigated [5], the effects of insert thickness on backside contact pressures, backside micromotion, and wear has not received adequate attention. Brandt et al. [2] has suggested that increased insert thickness was associated with increased backside damage (Fig. 1). In the present study, finite element simulations were conducted using the Sigma - Press Fit Condylar TKR (Sigma-PFC®, DePuy Orthopedics Inc., Warsaw, IN) with inserts of different insert thickness ranging between 5, 10, 15, 20 and 25 mm. The TKRs were simulated under ISO 14343-2 [8]. A non-linear PE material model was implemented by means of the J2-plasticity theory [6] and the effects of insert thickness on backside micromotion and contact pressure were analyzed. At the peak loading of the simulated gait cycle (time=13%), the 5 mm thick PE insert showed a greater backside peak contact pressure than the 25 mm thickness PE insert. Increasing insert thickness from 5 mm to 25 mm lead to approximately 15% greater peak micromotion at the modular interface (Fig. 2). This effect may be attributed to the ability of the PE material to distribute the load more evenly through deformation at the modular interface and reduce micromotion for thinner inserts. It is suggested that increased insert thickness results in increased moments at the modular interface that could lead to increased backside wear in silico. Although an increase in PE insert thickness was only associated with a moderate increase in backside micromotion in the present study, it was deemed likely that backside micromotion could be accelerated for thicker inserts in vivo as the PE locking mechanism has been shown to degrade after extended implantation periods.

Femoral components with an oxidized zirconium-niobium (OxZr) gradient ceramic surface (Oxinium, Smith & Nephew, Memphis, TN) were introduced as an alternative to cobalt-chromium (CoCr) alloy femoral components for the purpose of PE wear reduction in total knee replacements [1]. In the present study, the surface damage and clinical performance of both CoCr alloy and OxZr femoral components were investigated. By matching CoCr alloy and OxZr femoral components for clinical factors, as done by Heyse et al. [2], the surface damage on retrieved CoCr alloy and OxZr femoral component was assessed. Twenty-six retrieved cobalt-chromium (CoCr) alloy femoral components were matched with twenty-six retrieved oxidized zirconium (OxZr) femoral components for implantation period, body-mass index, patient gender, implant type (cruciate ligament retaining/substituting), and polyethylene insert thickness. Detailed surface profilometry was performed on retrieved femoral condyles in areas that had not been damaged by gouging [3] with the specific purpose of investigating the in vivo wear behaviour of undamaged OxZr surface. In addition, the cumulative survivorships were calculated for patients who had received CoCr alloy or OxZr femoral components from our orthopaedic database. In order to identify factors that affect the clinical performance of CoCr alloy and OxZr femoral components, the findings from the retrieval analysis and the survivorship analysis were combined. The Rp, Rpm, and Rpk-values for the retrieved CoCr alloy femoral components were found significantly higher than the Rp, Rpm, and Rpk-values for the retrieved OxZr femoral components (p ≤ 0.031). The roughness parameters values (Ra, Rq, Rz, Rp, Rpm, Rpk, Rv, and Rsk) for the retrieved CoCr alloy femoral components were found significantly higher than the values of the new, never implanted CoCr alloy femoral components (p ≥ 0.001). The surface roughness was higher on the medial condyles than the lateral condyles of the retrieved CoCr alloy femoral components; such a difference was not observed on the retrieved OxZr femoral components. The OxZr bearing surface appeared to protect the femoral components from abrasive wear in vivo. At 8.5-years follow up, the cumulative survivorship for the CoCr alloy femoral components (98%) was not found to be statistically significantly different (p = 0.343, Breslow test) from the OxZr femoral components (97.5%). Therefore, OxZr femoral components appeared to possess low wear characteristics and could be particularly suitable for younger, heavier patients to ensure long-term durability.

Ceramic-on-ceramic (C-C) total hip replacements (THRs) are an attractive option for young, active patients [1, 2]. However, more clinical data is necessary to establish the reasons of failure of contemporary C-C THRs in vivo. The objective of the present study was to assess the surface damage on retrieved C-C THRs and determine possible influential factors that may explain their in vivo performance. Thirty-five C-C retrievals of material type Biolox® forte (n=28) and Biolox® delta (n=7) (CeramTec AG, Plochingen, Germany) were collected after a mean of 3.7 ± 3.2 years in vivo. Semi-quantitative surface damage assessment [3] was performed on all retrievals to obtain both a damage score (DS) (Fig. 1). Contact profilometry was performed on the retrieved femoral heads to characterize the type of surface damage (metal transfer, stripe wear). Scanning electron microscope (SEM) images were obtained from two femoral heads displaying areas of typical surface damage. The implantation period correlated with the damage score (DS) of the femoral heads (R=0.573, p<0.001) and the acetabular cups (R=0.592, p<0.001). However, the metal transfer DS of the femoral heads did not correlate with implantation period (R=0.185, p=0.29). Surface roughness of metal-transfer areas were positively skewed (additive metal transfer) while the stripe damage areas were negatively skewed (grain removal), as evidenced by SEM analysis. Stripe damage was observed on both the Biolox® forte and Biolox® delta retrieved femoral heads; however, the extent of grain removal appeared less severe on the Biolox® delta retrieved femoral heads due to their overall smaller grain size (Fig. 2). Inclination angles > 45° was associated with a greater DS rate [DS/time of implantation], which had also been suggested elsewhere [4]. Four patients reported squeaking in their C-C THRs; one of which was a 54 yr-old male patient who completed three full marathons with his implant. In this his case, the DS for this retrieval was below average, with metal-transfer being the only macroscopic damage feature. Fracture of the acetabular liner occurred in three patients, all of which had malpositioned components. Metal-transfer on the ceramic surface could possibly cause a local break down of the fluid film and may facilitate, in addition to an increased inclination angle, stripe damage via an adhesive wear mechanism. Therefore, direct contact between the Ti-alloy acetabular shell and the ceramic femoral head should be avoided at primary surgery. C-C THRs remain an attractive option for young, active patients, but care must be taken during implantation to appropriately position the acetabular cup and to avoid unwanted metal-transfer as such alteration at the bearing interface may change implant tribology.

Revision of fractured ceramic-on-ceramic total hip replacements with a cobalt-chromium (CoCr) alloy-on-polyethylene articulation can facilitate metallosis and require further expensive revision surgery [1–3]. In the present study, a fifty-two year old male patient suffered from fatal cardiomyopathy after undergoing revision total hip arthroplasty. The patient had received a polyethylene-ceramic acetabular liner and a ceramic femoral head as his primary total hip replacement. The polyethylene-ceramic sandwich acetabular liner fractured in vivo after 58 months and the patient underwent his first revision surgery where he received a Vitamin E stabilized acetabular Polyethylene (PE) liner and a CoCr alloy femoral head with documented synovectomy at that time. After 15 months, the patient was admitted to hospital in cardiogenic shock, with retrieval of the bearing components. Before the second revision surgery, peak serum cobalt levels measured 6,521 μg/L, 78-times greater than serum cobalt levels of 83μg/L associated with cobalt poisoning [4]. Serum titanium levels found in the patient measured 17.5 μg/L) normal, healthy range 0–1.4 μg/L). The retrieved CoCr alloy femoral head had lost a total of 28.3g (24% or an estimated amount of 102 × 10⁻⁹ wear particles (∼2 μm diameter) [1]) within 16 months of in vivo service. Despite initiating a cobalt chelating therapy, the patients' cardiac left ventricular ejection fraction remained reduced at 6%. This was followed by multi-organ failure, and ultimately the patient passed away shortly after being taken off life support. Embedded ceramic particles were found on the backside and articular surfaces of the Vitamin E-stabilized PE acetabular liner. Evidence of fretting wear on the titanium (Ti) alloy acetabular shell was present, possibly explaining the increased serum Ti levels. Scanning electron microscopy and energy dispersive X-ray analyses confirmed Ti alloy transfer on the embedded ceramic particles on the backside PE liner surface and CoCr alloy transfer on the embedded ceramic particles on the articular PE liner surface. A fractured ceramic-on-ceramic total hip replacement should not be revised to a CoCr alloy-on-polyethylene articulation irrespective of concurrent synovectomy [5] as it can cause severe, third-body wear to the CoCr alloy femoral head that can lead to metallosis with fatal, systemic consequences.

Symptomatic cobalt toxicity from a failed total hip replacement is a rare but devastating complication. It has been reported following revision of fractured ceramic components, as well as in patients with failed metal-on-metal articulations. Potential clinical findings include fatigue, weakness, hypothyroidism, cardiomyopathy, polycythaemia, visual and hearing impairment, cognitive dysfunction, and neuropathy. We report a case of an otherwise healthy 46-year-old patient, who developed progressively worsening symptoms of cobalt toxicity beginning approximately six months following synovectomy and revision of a fractured ceramic-on-ceramic total hip replacement to a metal-on-polyethylene bearing. The whole blood cobalt levels peaked at 6521 µg/l. The patient died from cobalt-induced cardiomyopathy. Implant retrieval analysis confirmed a loss of 28.3 g mass of the cobalt–chromium femoral head as a result of severe abrasive wear by ceramic particles embedded in the revision polyethylene liner. Autopsy findings were consistent with heavy metal-induced cardiomyopathy.

We recommend using new ceramics at revision to minimise the risk of wear-related cobalt toxicity following breakage of ceramic components.

Cite this article: Bone Joint J 2013;95-B:31–7.

Wear of the polyethylene (PE) insert in total knee replacements can lead to wear-particle and fluid-pressure induced osteolysis. One major factor affecting the wear behaviour of the PE insert in-vivo is the surface characteristics of the articulating femoral components. Contemporary femoral components available in Canada are either made of cast Cobalt Chromium (CoCr) alloy or have an oxidized zirconium surface (Oxinium). The latter type of femoral components have shown to have increased abrasive wear resistance and increased surface wettability, thus leading to reduced PE wear in-vitro compared with conventional cast CoCr components. Although surface damage has been reported on femoral components in general, there have been no reports in the literature as to what extent the recommended operating techniques affect the surface tribology of either type of femoral component.

Twenty-two retrieved total knee replacements were identified with profound surface damage on the posterior aspect of the femoral condyles. The femoral components were of three different knee systems: five retrievals from the NexGen(r) total knee system (Zimmer Inc., Warsaw, IN), twelve retrievals from the Genesis II(r) total knee system (CoCr alloy or Oxinium; Smith & Nephew Inc., Memphis, TN), and five retrievals from the Duracon(r) total knee system (Stryker Inc., Mahwah, NJ). Reasons for revision were all non-wear-related and included aseptic loosening in two cases, painful flexion instability, and chronic infection. All retrieved femoral components showed evidence of surface damage on the condyles, at an average of 99° flexion (range, 43° – 135° flexion). Titanium (Ti) alloy transfer and abrasive surface damage were evident on all retrieved CoCr alloy femoral components that came in contact with Ti alloy tibial trays. Surface damage on the retrieved Oxinium femoral components was gouging, associated with the removal and cracking of the oxide and exposure of the zirconium alloy substrate material. CoCr alloy femoral components that had unintended contact with CoCr alloy tibial trays also showed evidence of gouging and abrasive wear.

All femoral components showed severe surface damage in the posterior aspect of the condyles. The femoral surface was heavily scratched and the oxidized zirconium coating surface appeared removed. The surface analysis suggested that the surface damage most likely occurred during the time of initial implantation. In particular, it appeared that the femoral condyles were resting on the posterior aspect of the tibial tray in flexion, thus scratching the femoral components. Such scratches could potentially lead to accelerated PE insert wear and reduced implant longevity, thus making expensive revisions surgery necessary. The authors strongly suggest a revision of the current operating techniques recommended by the implant manufacturer to prevent this type of surface damage from occurring.

Purpose: This study develops and validates a technique to quantify polyethylene wear in tibial inserts using micro-computed tomography (micro-CT), a nondestructive high resolution imaging technique that provides detailed images of surface geometry in addition to volumetric measurements.

Method: Six unworn and six wear-simulated Anatomic Modular Knee (DePuy Inc, Warsaw, IN) tibial inserts were evaluated. Each insert was scanned three times using micro-CT at a resolution of 50 μm. The insert surface was reconstructed for each scan through automatic segmentation and the insert volume was calculated. Gravimetric analysis was also performed for all inserts, and the micro-CT and gravimetric volumes were compared to determine accuracy. The utility of surface deviation maps derived from micro-CT was demonstrated by co-registering a worn and unworn insert. 3D deviations were measured continuously across the entire insert surface, including the articular and backside surfaces.

Results: The mean percent volume difference between the micro-CT and gravimetric techniques was 0.04% for the unworn inserts and 0.03% for the worn inserts. No significant difference was found between the micro-CT and gravimetric volumes for the unworn or worn inserts (P = 0.237 and P = 0.135, respectively). The mean coefficient of variation for volume between scans was 0.07% for both unworn and worn inserts. The map of surface deviations between the worn and unworn insert revealed focal deviations exceeding 750 μm due to wear.

Conclusion: Micro-CT provides precise and accurate volumetric measurements of polyethylene tibial inserts. Quantifiable 3D articular and backside surface deviation maps can be created from the detailed geometry provided by the technique. Compared to coordinate mapping, micro-CT provides 10 times greater surface sampling resolution (50 μm vs 500 μm) across the entire insert surface. Micro-CT is a useful analysis tool for wear simulator and retrieval studies of the polyethylene components used in total knee replacement.

Attempts have been made to develop standardise guidelines for knee implant wear testing of polyethylene (PE). The current ISO-14243-3 standard recommends the use “calf serum”, without giving ranges on the specific protein constituents and characteristics. In the present study, three types of frequently used calf sera with various protein constituents (albumin, globulins) were utilised. The effect of osmolality and hyaluronic acid (HA) was also assessed. An attempt was made to identify synovial fluid (SF) characteristics that may be responsible for the boundary lubrication in the joint with the goal to develop a more clinically relevant lubricant.

Twenty samples of SF were drawn from twenty patients and analyzed. Specific protein constituents and osmolality were then compared to three calf sera used for wear testing. Test One (six million cycles (Mc)): Bovine calf serum (BCS), newborn calf serum (NCS) and alpha-calf serum (ACS) were diluted with distilled water (DW). Test two (5.5Mc): ACS with an osmolality of 312 ± 1.00 mmol/kg (closest to clinical osmolality; diluted with phosphate buffered saline, PBS) and 145 ± 2.00 mmol/kg (diluted with DW) were consecutively tested. HA was added at a concentration of 1.5g/l. Modular total knee replacements of cruciate retaining design (GUR 1050, 10mm PE insert) were used.

ACS diluted with PBS appeared to be of closest specific protein constituents and osmolality when compared to SF. The wear rate for BCS was 21.81 ± 2.48 mg/Mc, 17.05 ± 3.25 mg/Mc for NCS, and 13.44 ± 0.79 mg/Mc for ACS (p < 0.016). Decreased osmolality amplified the PE wear by a factor of 2.3 (p = 0.020). Adding HA increased the PE wear by a factor of two (p = 0.002).

There was significant difference in PE wear rates between the three calf-sera. BCS and NBC did not have clinically relevant levels of specific protein constituents. This study strongly suggests that current standards for total knee wear testing should be revised to enable more controlled wear testing under more clinically relevant conditions. It is suggested to be of particular importance when new bearing materials, such as cross-linked PE’s, are evaluated and proposed for clinical application.

Purpose: To assess the clinical utility of a computer-based program (discrete dynamic contour method of boundary refinement; Matlab®) to quantify the location and extent of periprosthetic osteolysis around cemented total knee arthroplasties.

Methods: The radiographs of 171 primary Anatomical Modular Knee arthroplasties were reviewed by three independent, blinded raters using Matlab® software program. The location, size (mm2), polyethylene sterilization technique, alignment, presence or absence of femoral osteolysis and patient characteristics was recorded.

Results: The mean duration of follow-up was 8.47 ± 1.10 years with minimum in vivo wear interval= 2.0 years. The overall incidence of radiographically apparent osteolysis was estimated at 30.41%. Furthermore, the average intra-class correlation coefficient for this measurement technique for three independent observers was found to be significant for medially and laterally located zones of osteolysis (I.C.C. value = .7801; 95% C.I. = .7161–.8316; (p< .05). There was a statistically significant trend towards higher rates of osteolysis occurring within the series of polyethylene inserts gamma irradiated in air. Lowest rates of osteolysis were measured in the series sterilized in gas plasma (27.09 mm2 per year vs. 16.24 mm2 per year respectively; p< .001).

Conclusions: The Discrete Dynamic Contour method of boundary refinement represents an acceptably reliable means by which one can quantify the location and extent of osteolysis based upon digitalized radiographs of total knee replacements.