Clinical outcomes of ceramicized ball heads in total hip replacement bearings: a literature review

Titanium nitride: clinical results in THR bearings

TiN coatings are applied to hip, knee, ankle and shoulder replacements, but in spite of their extensive use, little exists in the literature concerning their clinical outcomes. Notwithstanding that TiN has long been in clinical use, there are only a few reports about the clinical outcomes for these bearings in total hip arthroplasty. The comprehensive review of TiN coatings by van Hove et al (29) reports only 7 clinical studies on TiN-coated implants since 1992, (2 on ankle replacements, 2 on THR, 2 on total knee replacements (TKR), and 1 on hip resurfacing). Sparse reports are putting in question the damage resistance of TiN coatings in clinical applications (30). The analysis of retrieved THR heads (31) shows extensive abrasion of the coating: in the case of a high-demand patient, the damage was extended to about 70% of the articulating surface.

The first clinical study of TiN in THR was by Massoud et al (32), who reviewed a cohort of 76 patients who underwent hip arthroplasty using the Capital THR. This implant – now withdrawn from clinical use – was a modification of the Charnley low friction arthroplasty. TiN-coated ball heads were used in 16 implants, while the remaining 60 received CoCr-alloy heads, all coupled with UHMWPE cups. At a mean 26 months of follow-up (range 10-37 months), 44% of the stems with TiN-coated heads were loose, while loosening was observed in 21.5% of the stems with a CoCr-alloy ball head.

Buechel et al (33) reported the clinical results for 130 consecutive hip arthroplasties performed in 11,777 patients from 1990 to 1999. All of the implants were cementless devices with the femoral head made of titanium alloy coated with a polished 6- to 10-µm-thick layer of TiN applied by the PVD process. Survivorship of the implants was 95.5% at 11.2 years. The analysis was carried out using as endpoints: (a) revision of any implant component for any reason, (b) clinically poor hip score, (c) radiographic evidence of failure (radiolucency >2 mm in all zones surrounding an implant) or (d) gross implant migration. Although radiolucent lines were observed in Gruen zones 1 to 14, there was no osteolysis around acetabular or femoral components of the implants. One patient showed asymptomatic femoral stem subsidence, and one acetabular component was revised for reported loosening.

Less favorable results were reported by Barouk et al (34), who investigated a series of 52 titanium alloy ball heads, TiN coated by ion implantation. Thirteen prostheses were revised for aseptic loosening, while revision for the same reason was needed for only one in the 118 prostheses making use of different types of ball heads (stainless steel, chromium-cobalt, alumina or zirconia). Besides aseptic loosening, the implants with TiN-coated ball heads showed a high incidence of polyethylene wear, metallosis and modifications of the head surface (disappearance of the coating, scratch marks etc.). Table I summarizes the clinical results of TiN coatings reported in the literature thus far.

Zirconium and chromium nitrides

Extensive hip simulator studies were completed for chromium nitride (CrN) and chromium carbon nitride (CrCN)–coated THR bearings (35), but these devices were not introduced into clinical use.

Wear simulator tests performed on a unicompartmental knee implant with zirconium nitride (ZrN) surface were reported by Affatato et al (36). The coating of this device was obtained under nitrogen atmosphere by a proprietary PVD process making use of a series of sputtering targets. The result was a coating formed by a thin Cr bond layer, 5 layers of CrN–CrCN and a final layer of ZrN, with total thickness between 3.5 and 5 μm. The aim of the gradient layers of CrN–CrCN is to bridge the high difference in hardness and stiffness between the mild CoCrMo substrate and the final ZrN coating, assuring in this way the mechanical integrity of the coating system. In addition, this system forms a barrier for ion diffusion from the substrate to the surrounding tissue (37). ZrN is used clinically in at least 1 knee replacement device, but there are no reports on its clinical outcomes so far.

Silicon nitride coatings

Silicon nitride (Si₃N₄) ceramics have long been used clinically in interbody fusion devices for spinal surgery (38). Their application in neurosurgical and maxillofacial devices is especially attractive for the lack of artefacts in nuclear magnetic resonance imaging (39). The potential of bulk Si₃N₄ in the production of ball heads and inserts for THRs has been investigated for some time (40, 41), and a clinical trial of Si₃N₄ ball heads manufactured by a US company is presently in progress (42). For an exhaustive review of the application of Si₃N₄ ceramics, the reader may refer to the work of Sonny Bal and Rahaman (43).

Amorphous or nanocrystalline Si₃N₄ coatings can be obtained by PVD or chemical vapor deposition (CVD) (44, 45). The coating composition can be tailored in terms of silicon to nitrogen ratio and in concentration of additives such as carbon, oxygen or fluorine. Si₃N₄ coatings on cobalt-chromium alloy are presently under development in the framework of an EU-funded project (46, 47) for application in hip or knee prosthesis bearings. Si₃N₄ ball heads and liners are also used as a substrate for nanocrystalline diamond (NCD) coating (48-49-50).

Zirconium dioxide: clinical results in THR bearings

Studies performed in hip simulators under standard conditions on artificially damaged head surfaces show that OxZr – in comparison with CoCr components – is superior, or at least equivalent, in its wear behavior (54-55-56-57). Nevertheless, such in vitro results are not confirmed by a growing body of clinical evidence (58-59-60-61-62-63-64-65-66) that is putting into question the advantages of OxZr over CoCr ball heads in hip replacements. For a summary of these studies, see Table II.

A study by Kadar et al (60) may be taken as an example of the doubts raised about the benefits of OxZr femoral heads in the clinical setting. This randomized study evaluated the wear and migration of 5 different cemented THRs. They enrolled 150 patients who received (a) a Charnley femoral stem with a 22.2-mm head coupled to a γ-sterilized Ogee cup; or (b) a Spectron EF femoral stem with a 28-mm head used with either EtO-sterilized polyethylene (Reflection All-Poly) or highly cross-linked (Reflection All-Poly HXLPE) cups, combined with either CoCr or Oxinium femoral heads. After 2 years, in the Spectron group, there was no significant difference in wear for OxZr versus CoCr ball heads. The EtO-sterilized UHMWPE cups had more than 4 times higher proximal penetration than the highly cross-linked polyethylene (HXLPE) cups, irrespective of the head used. The Charnley/Ogee system “was not outperformed” in the trial.

The largest study of OxZr bearings so far was by Jassim et al (62). They enrolled 400 patients undergoing hip arthroplasty in 4 hospitals, comparing wear, functional outcomes and complications. Randomized groups of patients received (a) a CoCr femoral head and XLPE liner, (b) OxZr femoral head and XLPE liner, or (c) OxZr femoral head and conventional UHMWPE liner. There were no statistically significant differences in clinical scores of the 368 patients available at 5-year follow-up. Penetration was significantly higher in the UHMWPE liner group compared with both XLPE liner groups, but no significant difference was noted between CoCr and OxZr articulating with XLPE.

Also the results presented by Zaoui et al (63) show that in the clinical setting, XLPE wear appears to be independent from the nature of surface of the metallic the ball head. One hundred patients were randomized to receive cemented Charnley stems (22.25-mm bearings) with either OxZr or stainless steel (SS; M30NW) femoral heads. UHMWPE was used in the first 50 patients, whereas XLPE was used in the next 50 patients. All other parameters were identical in both groups. At a median 6-year follow-up (range 4-8 years) Merle d’Aubigné scores among the 4 groups were similar. Apart from the SS-UHMWPE group, all of the other groups had a steady-state penetration rate well below the osteolysis threshold set by Dumbleton et al (67).

A study by Garvin et al (59) is presently the one with the longest follow-up for OxZr (mean follow-up 12 years, range 9-14 years). The study was carried out retrospectively on the records of 87 patients 50 years old or younger (96 hips) who had undergone hip arthroplasty in the authors’ hospital from 1999 to 2005. All of the implants had the same XLPE acetabular component coupled to 2 different stems (tapered press fit or porous coated). Femoral heads were made out CoCr (n = 19), ceramic (n = 34) or OxZr (n = 43), and different bearing diameters were used (φ26 mm, n = 14; φ28 mm, n = 71; φ32 mm, n = 11). At the follow-up, none of the hip radiographs showed evidence of loosening or osteolysis. The authors reported that linear and volumetric wear of XLPE coupled with 28-mm ceramic heads were much lower than with OxZr and on CoCr, notwithstanding the differences in pathology among patients and in stem types, bearing materials and diameters among implants.

A further body of literature reports cases of damage to OxZr ball heads following dislocation of the joint during handling in the operating room. Cases of severe OxZr scale damage have been reported since 2007 (68-69-70-71-72-73-74). The thin, hard oxide surface scale applied onto the soft substrate can be severely scratched in the case of contact with a hard surface (e.g., intraoperatively), in contacts with the socket rim in cases of joint instability or by interposition of third bodies in the bearing gap.

This problem was shown in the retrieval analysis of OxZr femoral heads performed by Moussa et al (75) in the Hospital for Special Surgery, New York, USA. They analyzed 59 OxZr retrieved ball heads explanted for recurrent dislocation. Eighteen heads showed gross visual evidence of substantial surface damage, disruption of the OxZr layer exposing the metal alloy underneath. The case reported by Cassar-Gheiti et al (76) shows the relevance of the problem. They treated a patient 10 years postoperatively who at the age of 35 had received a bilateral THR, with OxZr/XLPE bearings. After revision due to pain, both implants showed radiographic evidence of XLPE wear and extensive osteolysis. Magnetic resonance imaging (MRI) and computerized tomography showed catastrophic wear of the ball head and of the socket (XLPE insert and metal shell), cortical expansion and large cystic lesions. The patient underwent 2-stage revision surgery, with the implantation of a spacer with antibiotics, followed by the implantation of a tumor prosthesis.

Diamond-like carbon (DLC) coatings

Diamond-like carbon (DLC) refers to a class of carbon-based materials with process-dependent characteristics and behavior. The structure of DLC films may be described as a dispersion of diamond “nodules” in a graphite matrix. It is inherently metastable, and may transform to graphite at high temperatures. Their characteristics are often described by the ratio sp3/sp2 – for example, the concentration of diamond-type bonds vs. the graphite ones, a ratio that depends on the film deposition conditions. The biocompatibility of DLC coatings is a well-documented behavior (77-78-79-80). As in other coatings, the main problem with DLC coatings as well is the adhesion to the substrate, as the interfacial residual stresses may reach very high values within the 1- to 2-µm-thick coating layer. Several comprehensive reviews (81-82-83-84) were published in recent years about the potential application of DLCs in medical devices, and the development of nanocrystallyne diamond coatings (NCD) may offer new opportunities.

A number of experimental works have aimed to assess the wear behavior of DLC with regard to its use in THR bearings. The results have been confounded because these studies have been carried out on a number of different materials (different process parameters and different manufacturers). In addition, wear tests have been made with different experimental configurations and different environments. The results of hip simulator tests were controversial: e.g. Mishra and Davidson (83) measured wear rates of UHMWPE cups similar for alumina and DLC coated heads. Nevertheless, the same cups had lower wear rates if coupled with CoCr ball heads (83). Affatato et al (36) measured similar wear in UHMWPE cups against DLC-coated and uncoated CoCr heads, both values higher than the wear of the same cups against alumina heads.

Only a few attempts at clinical use of DLCs in joint replacement bearings have been reported so far, and at least 3 cases of clinical failure are described in the literature. The one involving the most patients was that of the knee replacement Diamond Rota Gliding, implanted in about 200 patients in Switzerland, which was withdrawn from the market in 2001 due to some cases of failure shortly after implantation, for delamination of the coating and high polyethylene wear (85). A French company made a further attempt with DLC coated components for shoulder replacements and ankle joints, but soon left the market (86). Taeger et al (87) reported the results of a clinical prospective study comparing the survivorship of 202 hip athroplasties with DLC-coated/UHMWPE either alumina/UHMWPE bearings, carried out by the same surgeonover a 2 years period. Survivorship analysis for aseptic loosening at a mean 7.5 years postoperatively demonstrated a significant difference between the 2 groups (54 % survival for DLC/UHMWPE vs. 88 % for Al₂O₃/UHMWPE bearings). No correlation with variables such as age, sex or bodyweight could be detected. Scanning electron microscopy (SEM) analysis detected a number of small pits in the diamond-carbon layers on the surfaces of 19 DLC heads retrieved in revision for aseptic loosening. Debris due to delamination of the DLC layer caused high UHMWPE wear and - in some cases - the metal heads too were worn out. Failures of DLC coated THR ball heads were also reported by de Thomasson et al (88).

An interesting approach now under development consists of the application of NCD coating on Si₃N₄ (89). NCD coatings can be obtained by CVD from hydrocarbon gases like methane at temperatures in the range 500°C-800°C. Si₃N₄ is the preferred substrate, because of the chemical bonds assuring the good adhesion of NCD coatings. Wear tests of ceramic–ceramic bearings with NCD coatings applied on Si₃N₄ ball heads and liners gave promising results (90, 91). Also, testing of multilayer NCD coatings on titanium or CoCr alloys is reported in the literature (92, 93), but information about clinical studies are missing so far.