**2. Ceramic-on-ceramic hip prosthesis performance**

The performance of ceramic-on-ceramic hip joints has been evaluated using data obtained from joints operating well within the body, prostheses retrieved due to joint failure and also tests performed within the laboratory. Firstly, the *in vitro* laboratory tests results will be dis‐ cussed, then the *in vivo* data will be given. In addition to this, "squeaking", one of the most common concerns relating to ceramic-on-ceramic joints will be discussed.

#### **2.1.** *In vitro*

The majority of recent papers discussing the tribology (lubrication, friction and wear) of ce‐ ramic-on-ceramic hip joints detail tests done under 'severe' conditions such as malposition‐ ing, edge-loading or microseparation. There are, however, some earlier studies that describe how these joints operate under 'standard' conditions.

A well positioned ceramic-on-ceramic hip, tested under the loads and motions expected during the standard walking cycle, performs exceptionally well in terms of friction, lubrica‐ tion and wear [6-24]. Friction tests using different viscosities of carboxy-methyl cellulose (CMC) solution show that these joints operate close to full-fluid film lubrication with very low friction factors (0.002 at physiological viscosities) [6, 12]. These ceramic-on-ceramic joints have been shown to have very low surface roughness values that play a part in this low friction [6]. Tests were also performed using different viscosities of bovine serum [6, 12]. Bovine serum is often used in the laboratory as a replacement for the body's natural lubri‐ cating fluid, synovial fluid, as it contains proteins that act in a similar manner to those present in synovial fluid and although CMC fluids replicate the shear-thinning behaviour of synovial fluid, they do not contain any proteins. The introduction of these proteins into the lubricating fluid resulted in higher friction (0.03 at physiological viscosities) and mixed lu‐ brication. These results are shown as Stribeck plots in Figure 1. A Stribeck plot shows the measured friction factor plotted against Sommerfeld number (a dimensionless parameter dependent on the lubricant viscosity, the entraining velocity of the bearing surfaces, the joint radius and the load applied). A rising trend of friction factor with increasing Sommerfeld number is indicative of a full-fluid film lubrication regime, whereas a falling trend is nor‐ mally indicative of mixed lubrication. In full-fluid film lubrication, the surfaces are com‐ pletely separated by the lubricant and the friction generated is due solely to the shearing of the lubricant film. In mixed lubrication, the load is carried in part by the contact between the asperities of the bearing surfaces and also by the pressure generated within the lubricant. Although the bovine serum tests suggested that these ceramic-on-ceramic joints were oper‐ ating in mixed lubrication with some asperity contact, there was no surface damage evident on the joints after testing. It was speculated that the proteins that adhere to the ceramic sur‐ faces when using protein-based lubricants produce a sufficiently thick layer to penetrate the fluid film and result in protein-to-protein contact and shearing. It is likely that the subse‐ quent friction developed by the protein-to-protein contact is greater than that due to the shearing of the lubricant film alone. Therefore, although higher friction and mixed lubrica‐ tion is encountered when testing under more physiological lubricating conditions, there is still little asperity contact during the normal walking cycle [12].

(ISO 6474:1981 (second-generation ceramics), which was replaced with ISO 6474:1994 (thirdgeneration ceramics) and has now been replaced by ISO 6474-1:2010 (fourth-generation ce‐ ramics)) the performance of these all ceramic bearings has been greatly improved [4, 5]. The third and fourth-generation alumina ceramics are manufactured using hot isostatic pressing. This produces a material that is highly pure with a small grain size (≤ 2.5 μm and many manufacturers produce ceramics with even smaller grain sizes) that provides material strength and minimises the risk of fracture. The majority of the ceramic-on-ceramic joints

Ceramic-on-ceramic hip prosthesis performance will be reviewed in this chapter (*in vitro* and *in vivo*) along with a discussion of the concerns with ceramic-on-ceramic joints that hap‐ pen in a minority of cases such as joint "squeaking" and component fracture. The majority of articles reviewed discuss the performance of alumina-on-alumina joints. There are, how‐ ever, other ceramic materials available on the market today for use in orthopaedic surgery, for example BIOLOX® delta (an alumina matrix composite containing 72.5% alumina, 25.5% zirconia, and 2% mixed oxides). The reader will be made aware when joints made from this

The performance of ceramic-on-ceramic hip joints has been evaluated using data obtained from joints operating well within the body, prostheses retrieved due to joint failure and also tests performed within the laboratory. Firstly, the *in vitro* laboratory tests results will be dis‐ cussed, then the *in vivo* data will be given. In addition to this, "squeaking", one of the most

The majority of recent papers discussing the tribology (lubrication, friction and wear) of ce‐ ramic-on-ceramic hip joints detail tests done under 'severe' conditions such as malposition‐ ing, edge-loading or microseparation. There are, however, some earlier studies that describe

A well positioned ceramic-on-ceramic hip, tested under the loads and motions expected during the standard walking cycle, performs exceptionally well in terms of friction, lubrica‐ tion and wear [6-24]. Friction tests using different viscosities of carboxy-methyl cellulose (CMC) solution show that these joints operate close to full-fluid film lubrication with very low friction factors (0.002 at physiological viscosities) [6, 12]. These ceramic-on-ceramic joints have been shown to have very low surface roughness values that play a part in this low friction [6]. Tests were also performed using different viscosities of bovine serum [6, 12]. Bovine serum is often used in the laboratory as a replacement for the body's natural lubri‐ cating fluid, synovial fluid, as it contains proteins that act in a similar manner to those present in synovial fluid and although CMC fluids replicate the shear-thinning behaviour of synovial fluid, they do not contain any proteins. The introduction of these proteins into the

discussed in this chapter were produced using third-generation ceramics.

540 Advances in Biomaterials Science and Biomedical Applications

**2. Ceramic-on-ceramic hip prosthesis performance**

how these joints operate under 'standard' conditions.

common concerns relating to ceramic-on-ceramic joints will be discussed.

material are discussed.

**2.1.** *In vitro*

**Figure 1.** Stribeck plot for ceramic-on-ceramic joints (Scholes *et al* (2006) [12]). CMC: CMC fluids; BS: bovine serum.

With this good lubrication and low surface roughness the wear volumes produced under 'standard' conditions in the laboratory have, inevitably, been shown to be very low (less than 0.4 mg/million cycles cf. approximately 35 mg/million cycles for conventional metal or ceramic-on-UHMWPE joints, see Table 1), and sometimes almost immeasurable. As these ceramic-on-ceramic joints are working close to full-fluid film lubrication there is little or no contacting of the asperities on the joint surfaces leading to this low wear and friction [6, 12]. Also, laboratory studies have shown that cup malpositioning and elevated swing phase load testing have not significantly affected the wear of these joints [21, 22, 25, 26], see Table 2. This combination should, therefore, lead to a very successful artificial joint.


**Table 1.** Wear rates found for well-positioned conventional metal or ceramic-on-UHMWPE and ceramic-on-ceramic joints under standard testing conditions

More severe loading conditions have, however, given slightly different results. Microsepara‐ tion was first introduced during the swing phase of the walking cycle on the Leeds hip wear simulator [30] to replicate the stripe wear sometimes seen on retrieved ceramic-on-ceramic joints. Microseparation was incorporated in the simulator studies at Leeds because a year earlier, it was suggested by Mallory *et al* (1999) [31] that this separation of the femoral head and acetabular cup can occur in conventional metal-on-UHMWPE joints. Nevelos *et al* (2000) [30] hypothesised that a similar mechanism may take place in ceramic-on-ceramic joints re‐ sulting in the visible stripe wear found on some retrievals. Microseparation was, therefore, set up in the simulator and involved separation of the femoral head and acetabular cup dur‐ ing the swing phase of walking leading to relocation (with rim contact and edge-loading) during heel strike before the head then relocated in the cup in the stance phase (see Figure 2). The surface damage caused by this rim contact resulted in what is known as stripe wear.

**Figure 2.** Microseparation as applied in the Leeds simulator by Nevelos *et al* (2000) [30]. (A) Swing phase: microsepa‐ ration. (B) Heel-strike: rim contact. (C) Stance phase: relocation.

Using this microseparation technique, Nevelos *et al* (2000) [30] found slightly higher wear rates with ceramic-on-ceramic joints than found under 'standard' conditions (see Table 2). This study, however, was relatively short-term (800,000 cycles). Longer-term tests per‐ formed by Stewart *et al* (2001) [32] on the same simulator gave similar results to those found by Nevelos *et al* (2000) [30]. Other workers have also found an increase in wear rate with microseparation [33]. More recent work performed by Al-Hajjar *et al* (2010) [25] studied BI‐ OLOX® delta couplings. These joints gave lower wear than that found for alumina-on-alumi‐ na joints. Microseparation of the BIOLOX® delta joint surfaces, again, resulted in higher wear. The wear rates found using microseparation are, however, still extremely low in com‐ parison with conventional joints (see Tables 1 and 2).

**Reference Material combination Components Wear rate (mg/million**

[27] Metal-on-UHMWPE UHMWPE cup 47.4 [27] Ceramic (zirconia)-on-UHMWPE UHMWPE cup 37.9 [28] Metal-on-UHMWPE UHMWPE cup ~33.7 [28] Ceramic (zirconia)-on-UHMWPE UHMWPE cup ~29.0 [29] Metal-on-UHMWPE UHMWPE cup ~40.9 [29] Ceramic (zirconia)-on-UHMWPE UHMWPE cup ~29.5 [13] Ceramic-on-ceramic Head and cup ~0.25 [16] Ceramic-on-ceramic Head and cup ~0.16 [17] Ceramic-on-ceramic Head and cup < 0.04 [18] Ceramic-on-ceramic Head and cup 0.09 [24] Ceramic-on-ceramic Head and cup ~0.20

**Table 1.** Wear rates found for well-positioned conventional metal or ceramic-on-UHMWPE and ceramic-on-ceramic

More severe loading conditions have, however, given slightly different results. Microsepara‐ tion was first introduced during the swing phase of the walking cycle on the Leeds hip wear simulator [30] to replicate the stripe wear sometimes seen on retrieved ceramic-on-ceramic joints. Microseparation was incorporated in the simulator studies at Leeds because a year earlier, it was suggested by Mallory *et al* (1999) [31] that this separation of the femoral head and acetabular cup can occur in conventional metal-on-UHMWPE joints. Nevelos *et al* (2000) [30] hypothesised that a similar mechanism may take place in ceramic-on-ceramic joints re‐ sulting in the visible stripe wear found on some retrievals. Microseparation was, therefore, set up in the simulator and involved separation of the femoral head and acetabular cup dur‐ ing the swing phase of walking leading to relocation (with rim contact and edge-loading) during heel strike before the head then relocated in the cup in the stance phase (see Figure 2). The surface damage caused by this rim contact resulted in what is known as stripe wear.

**Figure 2.** Microseparation as applied in the Leeds simulator by Nevelos *et al* (2000) [30]. (A) Swing phase: microsepa‐

ration. (B) Heel-strike: rim contact. (C) Stance phase: relocation.

joints under standard testing conditions

542 Advances in Biomaterials Science and Biomedical Applications

**cycles)**


It is still unknown if it is this microseparation that causes the stripe wear that is observed on some retrievals or whether this type of wear is due simply to edge-loading of the head on the cup through a different mechanism. Higher rates of wear including the appearance of stripe wear may also be due to a steep acetabular cup implantation angle or repeated dislo‐ cations [34]. Microseparation, or edge-loading, may occur during various different physical activities such as stair climbing, standing from squat position and deep flexion. It may, how‐ ever, also occur during walking. These simulator studies incorporating microseparation of the head and cup into the walking cycle are therefore a severe testing method. The resulting rim contact and stripe wear does, however, replicate the more severe conditions that these joints may encounter thus producing the same effect that is seen on some retrievals but not necessarily through the correct corresponding actions.

If, and when, wear of these ceramic components occurs it is important to understand how the body is likely to react to these wear particles. It is well known that wear particle induced osteolysis is a major concern for conventional metal-on-UHMWPE joints [3] but is this the case for ceramic-on-ceramic prostheses? Promisingly, several workers have shown that the cellular response to ceramic particles is less severe than that due to polyethylene particles [35, 15]. An important point to note with the laboratory tests discussed above is that, even under extreme loading and motion conditions, these joints provide low wear with a minimal adverse tissue reaction to these wear particles. As stated by Fisher *et al* (2006) [15], the com‐ bination of low wear and low reactivity means that '...ceramic-on-ceramic bearings address the tribological lifetime demand of highly active patients.'.

#### **2.2.** *In vivo*

As shown above, the laboratory test results for ceramic-on-ceramic joints are very promising. Is this mirrored by the *in vivo* performance? There are many publications detailing the per‐ formance of ceramic-on-ceramic hip joints in the body. A selection of these papers is discussed below to give a general overview of joint performance using this material combination.

The short-term (mean follow-up of 50.4 months) performance of ceramic-on-ceramic joints was compared to that of metal-on-highly cross linked polyethylene (XLPE) joints in a study reported by Bascarevic *et al* (2010) [36]. Seventy-five metal-on-highly XLPE hips (72 patients) and 82 ceramic-on-ceramic hips (78 patients) were assessed. Both were found to work well with no revisions performed on the ceramic-on-ceramic joints compared with 2 revisions necessary for the metal-on-highly XLPE group. The authors commented that ceramic-on-ce‐ ramic components manufactured using third-generation ceramics were "especially suited for hip arthroplasty in young and active persons".

A comparative study was also performed by Amanatullah *et al* (2011) [37]. The short-term performance (60 months) of 125 ceramic-on-ceramic hip joints was assessed against 95 ce‐ ramic-on-UHMWPE prostheses. In this study both the material combinations preformed well and no statistically significant difference was found between the clinical outcome scores for the two types of prosthesis, however, one important point to note is that audible noise such as "squeaking" did occur in 3.1%. of the ceramic-on-ceramic hips. "Squeaking" is a phenomenon reported by others and this will be discussed later. It was noted that this is a short-term study (five years) and, as the metal-on-UHMWPE hips did not produce as low wear rates in the radiographic analysis, osteolysis may occur at a later stage.

Another short-term study comparing the results of 525 hips (421 ceramic-on-ceramic and 104 metal-on-UHMWPE) was reported by Johansson *et al* (2011) [38]. These joints had been implanted for an average of 59 months. The survival rates were 98% and 92% for the ceram‐ ic-on-ceramic and metal-on-UHMWPE hip joints respectively.

A short-term study (60 month follow-up) reported by Nikolaou *et al* (2012) [39] assessed and compared the performance of 36 metal-on-UHMWPE, 32 metal-on-highly XLPE and 34 third-generation ceramic-on-ceramic hip joints. The apparent wear of the bearing surfaces was assessed radiologically. The ceramic-on-ceramic joints showed the lowest wear (mean linear wear 0.035 mm) and the metal-on-highly XLPE gave nearly 3 times lower wear than the metal-on-UHMWPE joints (0.329 mm cf. 0.869 mm). At 60 months, no difference in clini‐ cal outcome was found for these three different material combinations. "Squeaking" was ob‐ served in 3 (8.8%) of the ceramic-on-ceramic joints; no revision procedures were necessary as a results of this squeaking though.

the head and cup into the walking cycle are therefore a severe testing method. The resulting rim contact and stripe wear does, however, replicate the more severe conditions that these joints may encounter thus producing the same effect that is seen on some retrievals but not

If, and when, wear of these ceramic components occurs it is important to understand how the body is likely to react to these wear particles. It is well known that wear particle induced osteolysis is a major concern for conventional metal-on-UHMWPE joints [3] but is this the case for ceramic-on-ceramic prostheses? Promisingly, several workers have shown that the cellular response to ceramic particles is less severe than that due to polyethylene particles [35, 15]. An important point to note with the laboratory tests discussed above is that, even under extreme loading and motion conditions, these joints provide low wear with a minimal adverse tissue reaction to these wear particles. As stated by Fisher *et al* (2006) [15], the com‐ bination of low wear and low reactivity means that '...ceramic-on-ceramic bearings address

As shown above, the laboratory test results for ceramic-on-ceramic joints are very promising. Is this mirrored by the *in vivo* performance? There are many publications detailing the per‐ formance of ceramic-on-ceramic hip joints in the body. A selection of these papers is discussed

The short-term (mean follow-up of 50.4 months) performance of ceramic-on-ceramic joints was compared to that of metal-on-highly cross linked polyethylene (XLPE) joints in a study reported by Bascarevic *et al* (2010) [36]. Seventy-five metal-on-highly XLPE hips (72 patients) and 82 ceramic-on-ceramic hips (78 patients) were assessed. Both were found to work well with no revisions performed on the ceramic-on-ceramic joints compared with 2 revisions necessary for the metal-on-highly XLPE group. The authors commented that ceramic-on-ce‐ ramic components manufactured using third-generation ceramics were "especially suited

A comparative study was also performed by Amanatullah *et al* (2011) [37]. The short-term performance (60 months) of 125 ceramic-on-ceramic hip joints was assessed against 95 ce‐ ramic-on-UHMWPE prostheses. In this study both the material combinations preformed well and no statistically significant difference was found between the clinical outcome scores for the two types of prosthesis, however, one important point to note is that audible noise such as "squeaking" did occur in 3.1%. of the ceramic-on-ceramic hips. "Squeaking" is a phenomenon reported by others and this will be discussed later. It was noted that this is a short-term study (five years) and, as the metal-on-UHMWPE hips did not produce as low

Another short-term study comparing the results of 525 hips (421 ceramic-on-ceramic and 104 metal-on-UHMWPE) was reported by Johansson *et al* (2011) [38]. These joints had been implanted for an average of 59 months. The survival rates were 98% and 92% for the ceram‐

wear rates in the radiographic analysis, osteolysis may occur at a later stage.

ic-on-ceramic and metal-on-UHMWPE hip joints respectively.

below to give a general overview of joint performance using this material combination.

necessarily through the correct corresponding actions.

544 Advances in Biomaterials Science and Biomedical Applications

the tribological lifetime demand of highly active patients.'.

for hip arthroplasty in young and active persons".

**2.2.** *In vivo*

Early to mid-term results were also reported by Stafford *et al* (2012) [40]. At a mean followup of 59 months six of these 250 ceramic-on-ceramic hips were revised. Two were revised for recurrent dislocation secondary to impingement, two for deep infection, one for recur‐ rent dislocation and one due to fracture of the femoral head. Although no patients experi‐ enced "squeaking", six described a grinding or crunching noise that was experienced mainly during deep flexion. One of these "squeaking joints" was a BIOLOX® delta ceramic prosthesis. Again, no revision surgery was performed on these noisy joints.

Mesko *et al* (2011) [41] assessed the outcome of 930 ceramic-on-ceramic hips over 10 years that were implanted by nine different surgeons. The survivorship at 10 years (96.8%) was referred to as 'excellent' with 0.9% fracture, 2.3% dislocation and 2.5% reported incidents of noise such as clicking, squeaking, popping, or creaking.

Another 10 year follow-up was reported by Yeung *et al* (2012) [42]. The clinical information was available for 244 hips (227 patients) and the radiographic information was available for 184 hips (172 patients). The success of these joints led to an overall survival rate of 98% (again, with revision for any reason as the end point).

In a multicentre study performed by Capello *et al* (2008) [43], 452 patients (475 hips) were assessed. The majority of the hips implanted were ceramic-on-ceramic (380) whilst the re‐ mainder (95) were conventional metal-on-UHMWPE and used as a comparison. The aver‐ age patient age was 53 years and there was an average 8 year follow-up period. The authors found the clinical results to be excellent. The ten-year Kaplan-Meier survivorship (with revi‐ sion of either component for any reason given as the end-point) was stated as 95.9% for the ceramic-on-ceramic hip joints and 91.3% for metal-on-UHMWPE. The ceramic joints, there‐ fore, performed better than the conventional joints and because one of the major causes of failure for conventional metal-on-UHMWPE joints is late aseptic loosening due to osteolysis, a longer term study is likely to give a greater difference in survivorship ratings. Of the ce‐ ramic-on-ceramic failures requiring revision, only 2/380 (0.5%) were due to ceramic frac‐ tures. Therefore, overall it was concluded that the third-generation ceramic-on-ceramic hip joints performed well clinically and radiographically in this young patient group. Although these joints performed relatively well "squeaking" was reported in 3/380 (0.8%) cases for the ceramic-on-ceramic joints.

In the study reported by Lee *et al* (2010) [44], the ten year survival rate of 88 ceramic-on-ce‐ ramic hip joints, with revision of either the head or the cup for any reason being as the end point, was 99.0%. These results were taken after a minimum of 10 years postoperatively and are, indeed, very promising.

A long-term study on earlier generation ceramic joints was performed by Hernigou *et al* (2009) [45]. This was a comparative study which investigated the wear and osteolysis of 28 bilateral arthroplasties (one ceramic-on-ceramic and the contralateral ceramic-on-UHMWPE). All of these joints (ceramic-on-ceramic and ceramic-on-UHMWPE) had lasted 20 years without the need for revision surgery. Computer tomography (CT scan) was used to assess the number and volume of osteolytic lesions present. Fewer osteolytic lesions were found on the side with the ceramic-on-ceramic joint; there were 13 cases of pelvic osteolysis found and 15 joints showed evidence of femoral osteolysis (cf. 24 joints with pelvic osteoly‐ sis and 23 with femoral osteolysis for the ceramic-on-UHMWPE joints). Also, in each pa‐ tient, the diameter, surface and volume of osteolysis was substantially lower on the side with the ceramic-on-ceramic hip implant than the ceramic-on-UHMWPE joint. It must also be remembered that these joints were manufactured using first-generation ceramics and the ceramics produced to the standard expected for third and fourth-generation ceramics are ex‐ pected to perform even better.

Another longer-term study (mean 20.8 years), again with earlier generation ceramics, per‐ formed by Petsatodis *et al* (2010) [46] also showed good results. There were 100 patients in this study, all with at least one ceramic-on-ceramic hip joint (109 hips) and, again, a young patient population (average age 46 years). The cumulative rate of survival was quoted as be‐ ing 84.4% at 20.8 years.

Although these results are promising, in the majority of cases, the follow-up period particu‐ larly for the third generation ceramics, was only short-term. It will, therefore be very inter‐ esting to evaluate the performance of these third and fourth-generation ceramic-on-ceramic joints, along with the metal and ceramic-on-highly XLPE, on a longer-term basis. These re‐ sults are eagerly anticipated.

The papers discussed so far have shown exceptional performance of ceramic-on-ceramic joints and suggests that these joints may perform better than metal or ceramic-on UHMWPE. This, however, is not reflected in the data described in the National Joint Regis‐ try (NJR) of England and Wales (2011) [1]. Promisingly though, the NJR states that there is 'little substantive difference' in the risk of revision for ceramic-on-ceramic, ceramic-on-poly‐ ethylene or metal-on-polyethylene joints. It is, however, not stated whether this polyethy‐ lene is UHMWPE or XLPE and the differences in performance between these two materials is not listed.

Although, as discussed above, many of these ceramic-on-ceramic hip joints perform excep‐ tionally well, early dislocation is seen as a possible concern due to limited modular neck length and other factors. In the majority of the published literature referred to in this text [47, 36, 48, 43, 49, 50, 46, 51-57, 41, 40], the occurrence of dislocation is 0% - 2.3%. The num‐ ber of dislocations was higher (6%) in a study reported by Chevillotte *et al* (2011) [58], how‐ ever, no reoperation was required for any of these cases. Colwell *et al* (2007) [53] summarised other published literature detailing the number of dislocations in ceramic-onceramic hip joints in comparison with metal-on-polyethylene and found no difference be‐ tween the two material combinations. In addition to this Amanatulla *et al* (2011) [37] and Bascarevic *et al* (2010) [36] showed no difference in the number of dislocations for ceramicon-ceramic and ceramic-on-UHMWPE or metal-on-highly XLPE hip prostheses. Often there is no need for revision surgery after dislocation, unless it is recurrent dislocation.

point, was 99.0%. These results were taken after a minimum of 10 years postoperatively and

A long-term study on earlier generation ceramic joints was performed by Hernigou *et al* (2009) [45]. This was a comparative study which investigated the wear and osteolysis of 28 bilateral arthroplasties (one ceramic-on-ceramic and the contralateral ceramic-on-UHMWPE). All of these joints (ceramic-on-ceramic and ceramic-on-UHMWPE) had lasted 20 years without the need for revision surgery. Computer tomography (CT scan) was used to assess the number and volume of osteolytic lesions present. Fewer osteolytic lesions were found on the side with the ceramic-on-ceramic joint; there were 13 cases of pelvic osteolysis found and 15 joints showed evidence of femoral osteolysis (cf. 24 joints with pelvic osteoly‐ sis and 23 with femoral osteolysis for the ceramic-on-UHMWPE joints). Also, in each pa‐ tient, the diameter, surface and volume of osteolysis was substantially lower on the side with the ceramic-on-ceramic hip implant than the ceramic-on-UHMWPE joint. It must also be remembered that these joints were manufactured using first-generation ceramics and the ceramics produced to the standard expected for third and fourth-generation ceramics are ex‐

Another longer-term study (mean 20.8 years), again with earlier generation ceramics, per‐ formed by Petsatodis *et al* (2010) [46] also showed good results. There were 100 patients in this study, all with at least one ceramic-on-ceramic hip joint (109 hips) and, again, a young patient population (average age 46 years). The cumulative rate of survival was quoted as be‐

Although these results are promising, in the majority of cases, the follow-up period particu‐ larly for the third generation ceramics, was only short-term. It will, therefore be very inter‐ esting to evaluate the performance of these third and fourth-generation ceramic-on-ceramic joints, along with the metal and ceramic-on-highly XLPE, on a longer-term basis. These re‐

The papers discussed so far have shown exceptional performance of ceramic-on-ceramic joints and suggests that these joints may perform better than metal or ceramic-on UHMWPE. This, however, is not reflected in the data described in the National Joint Regis‐ try (NJR) of England and Wales (2011) [1]. Promisingly though, the NJR states that there is 'little substantive difference' in the risk of revision for ceramic-on-ceramic, ceramic-on-poly‐ ethylene or metal-on-polyethylene joints. It is, however, not stated whether this polyethy‐ lene is UHMWPE or XLPE and the differences in performance between these two materials

Although, as discussed above, many of these ceramic-on-ceramic hip joints perform excep‐ tionally well, early dislocation is seen as a possible concern due to limited modular neck length and other factors. In the majority of the published literature referred to in this text [47, 36, 48, 43, 49, 50, 46, 51-57, 41, 40], the occurrence of dislocation is 0% - 2.3%. The num‐ ber of dislocations was higher (6%) in a study reported by Chevillotte *et al* (2011) [58], how‐ ever, no reoperation was required for any of these cases. Colwell *et al* (2007) [53] summarised other published literature detailing the number of dislocations in ceramic-on-

are, indeed, very promising.

546 Advances in Biomaterials Science and Biomedical Applications

pected to perform even better.

ing 84.4% at 20.8 years.

sults are eagerly anticipated.

is not listed.

Component fracture in ceramic-on-ceramic hips is another cause for concern for many sur‐ geons and patients. The fracture rates of ceramic joints have been dramatically reduced since the introduction of third and fourth-generation ceramics with the new material proc‐ essing methods. Ceramic-on-ceramic joints have strict regulations that must be abided by with regard to material properties such as burst strength; as discussed in work reported by Salih *et al* (2009) [59]. However, fracture of these components does still occur, leading to catastrophic joint failure and revision surgery. In the majority of cases though, this fracture rate is extremely low (0% - 0.5%) [60, 61, 43, 49, 45, 50, 46, 51, 48, 55, 57, 40]. Some reports do, however, give a higher rate of post-operative fracture (1-2.3%) [62, 44, 37, 42, 47, 39]. Frac‐ ture is often associated with trauma, however a multicentre review reported by Park *et al* (2006) [63] discussed the performance of 357 third-generation ceramic-on-ceramic THRs and fracture occurred under normal activities in six of these hips (1.7%). This design used a poly‐ ethylene-ceramic composite liner within a titanium alloy shell and the high rate of fracture led to the authors discontinuing its use. It must be recognised though that, in general, the rate of fracture for ceramic-on-ceramic hip joints is very low. In fact, an article describing the fracture of ceramic joints (Hannouche *et al* (2003) [64]) stated that during a 25 year period (1977 – 2001) 11 (less than 0.004%) of the 3300 ceramic-on-ceramic joints reported on failed due to fracture. This low rate of fracture is from ceramics produced using earlier versions of the ISO standard and, so, this is expected to reduce even further and many studies using third-generation ceramics have reported no failures due to fracture [51, 36, 65, 66], Fracture must, however, still be recognised as a risk (albeit very low) with ceramic-on-ceramic joints.

For what other reasons does failure occur? Savarino *et al* (2009) [67] analysed the clinical, ra‐ diographic, laboratory and microbiological data from 30 retrieved ceramic-on-ceramic hip components. They concluded that failure was due to malpositioning of the joint during sur‐ gery leading to mechanical instability, or trauma or infection. Loosening in this selection of joints was not due to wear debris induced osteolysis. They indicated that the wear debris produced by these joints and the osteolysis present were the effect of the loosening, rather than the cause. It is stated that correct positioning of the implant is crucial. This has also been stated by other workers [54].

A case study was reported by Nam *et al* (2007) [68] discussing a failed ceramic-on-ceramic hip joint where failure was stated as being caused by alumina debris-induced osteolysis. A sixty-three year old woman underwent bilateral THR in 1998. Eight years later the patient returned to the clinic for routine follow-up and was experiencing no discomfort or worrying symptoms. However, the radiographs taken showed expansive osteolytic lesions. After revi‐ sion surgery a histologic analysis of the retrieved tissues showed alumina particles within the cytoplasm of macrophages and in intercellular tissue suggesting wear particle induced osteolysis. Alumina wear particle-induced osteolysis is, however, a very rare phenomenon.

Chang *et al* (2009) [69] reported on the clinical and radiographic outcomes when using thirdgeneration ceramic-on-ceramic joints in revision THR of 42 failed metal-on-UHMWPE hips. This was an interesting study as most published literature discuss the choice of bearing ma‐ terial for primary surgery and, as stated by the authors, '...few studies have focussed on the choice of bearing surface in revision...'. This was a young patient group (mean age: 48.8 years, range: 32 – 59 years) and the mean length of time between primary and revision sur‐ gery was 9.5 years (range: 3.3 – 16.1 years). The mean duration of follow-up after this revi‐ sion surgery was 64 months (range: 38 – 96 months). At the time of publication of this article, no hips needed additional revision surgery and no hips showed radiolucent lines, acetabular cup migration or osteolysis. This study gives very favourable results for the use of ceramic-on-ceramic hip joints in revision surgery, especially for the younger patient as the likelihood of the need for further revision is greater.
