**5.1 Preoperative diagnosis**

The most common indication for THA is idiopathic osteoarthritis. Within this diagnosis group, those with an atrophic pattern of bone response to osteoarthritis are at increased risk of acetabular prosthesis loosening31. Proximal femoral bone geometry may also affect prosthesis survival, with large non-tapering femoral canal shape (stove-pipe) being associated with an increased risk of aseptic loosening31.

Although the final pathway to the development of aseptic loosening is process of mechanical failure of the construct driven by inflammatory-mediated bone loss, multiple factors mediate an individual's susceptibility to this process. These may broadly be divided into patient, surgical, and prosthesis-related factors (Figure 4). Although not an exhaustive list, some of these proposed factors that have been identified and validated will be

Fig. 4. Summary of risk factors that influence the development of aseptic loosening

The most common indication for THA is idiopathic osteoarthritis. Within this diagnosis group, those with an atrophic pattern of bone response to osteoarthritis are at increased risk of acetabular prosthesis loosening31. Proximal femoral bone geometry may also affect prosthesis survival, with large non-tapering femoral canal shape (stove-pipe) being

**4. Risk factors for osteolysis** 

discussed.

**5. Patient risk factors 5.1 Preoperative diagnosis** 

associated with an increased risk of aseptic loosening31.

Higher rates of prosthesis loosening also occur in patients who have undergone arthroplasty for post-traumatic arthritis and osteonecrosis when compared with primary osteoarthritis. However, it is thought that this finding may relate to higher activity levels and increased bearing surface wear, rather than being a function of the pre-operative diagnosis32,33.

A number of preoperative diagnoses carry a possible increased risk of prosthesis failure through associated medication. Patients taking systemic steroids have been found to have a higher risk of reoperation34. Non-steroidal anti-inflammatory drugs (NSAIDs) have been implicated in impaired bone healing, and patients taking NSAIDs have higher reoperation rates, although NSAID use may be acting as a marker of a painful prosthesis rather than contributing directly to prosthesis failure34.

Poorer prosthesis survival might be expected in patients with inflammatory arthropathy due to its inflammatory pathogenesis and the historic frequent use of corticosteroids in its treatment (that are associated with loss of bone mass through osteoblast suppression). However, Furnes *et al*, in a large arthroplasty registry-based study, found no difference in THA survival between patients with rheumatoid arthritis versus those with osteoarthritis35. Rud-Sorensen *et al* found that the risk of stem revision due to aseptic loosening was lower in rheumatoid patients versus primary osteoarthritis, whilst acetabular prosthesis survival was similar36.

Furnes *et al* and Bordini *et al* have reported higher acetabular revision rates due to aseptic loosening in patients with a primary diagnosis of developmental dyplasia of the hip compared to primary osteoarthritis37,38. Rates of acetabular prosthesis failure are higher in younger patients and those with greater graft coverage of the cup39. The role of these factors is unclear, but may relate to activity levels, or mechanical factors influencing prosthesis support.

## **5.2 Body mass index and obesity**

The Health Survey for England 2009 showed that over the last 16 years there has been marked increase in the proportion of the population that are obese. This proportion increased from 13% of men in 1993 to 22% in 2009 and from 16% of women in 1993 to 24% in 200940. The mean BMI of a patient undergoing THA in England and Wales has increased over the last 5 years from 27.4 to 28.4. Likewise the percentage of patients classed as either obese or morbidly obese has risen from 29% in 2004 to 37%

Historically, obesity has been deemed a relative contraindication for THA41, as the joint reaction force experienced at the hip is directly proportional to body weight, and thus obesity was considered a risk factor for prosthesis failure. Obesity is associated with a higher incidence of perioperative complications including cardiovascular and respiratory events42, venous thrombosis43, wound infection44, and dislocation45. However, despite the increase in joint load in these patients, no consistent increase in bearing wear or osteolysis has been shown across study populations46,47 and thus obesity is not a clear risk factor for osteolysis.

### **5.3 Bearing-surface wear and activity level**

Patient activity level associates with osteolysis. It is thought this association operates primarily though the production of wear of the bearing surface. Flugsrud showed that patients who undertake intermediate to intense activity are four times more likely than the less-active to develop acetabular prosthesis loosening48. A recent study with five to ten year

Risk Factors for Aseptic Loosening Following Total Hip Arthroplasty 283

follow up has shown that 24% of patients who have engaged in high levels of activity

Traditionally the rate of polyethylene wear has been reported as a function of time. The results from ex-vivo hip simulator experiments have shown that the number of hip cycles is proportional to the rate of wear of prosthesis surface50. In vivo, there is a great range of wear rates between individual as a consequence of differing activity levels51. Several validated assessment tools have been developed to measure activity levels in arthroplasty populations52, and Schmalzried *et al* showed that wear in patients is a function of activity53. There are no clear guidelines outlining what levels of activity can be undertaken following THA although the proportion of patients participating in athletic activity following THA ranges between 52 – 83%54-56. Whilst low-impact activities such as walking, swimming and cycling have always been recommended following THA, some patients participate in more high-impact and competitive sports. The increasing participation in athletic activity and higher post-operative expectations can partly be explained by the increasing numbers of younger patients undergoing THA. 42% of men and 31% of women who underwent THA in England and Wales in 2009 were under the age of 65 years2. A large number of patients over the age of 65 are also participating in

Several investigators have shown a relationship between high levels of polyethylene wear and osteolysis/aseptic loosening, and the concept of a wear-rate 'threshold' (commonly defined as 0.1mm/year) below which osteolysis occurs very rarely, has been suggested. Wilkinson et al quantitated the association between wear and osteolysis and found no evidence to support this concept. In a case-control study of 230 hips after cemented Charnley THA with a metal on polyethylene bearing they showed that the risk of osteolysis increased with each quintile increase in wear, from very low levels of wear, below the suggested threshold, through to high levels57. They subsequently showed that the risk of osteolysis showed a similar pattern of consistently increasing risk ratio with each wear rate quintile in a separate cohort study of patients with 319 hybrid THAs using a metal on

Within a given ethnic population the sequence of DNA between individuals is 99.5% identical. However, variability within the code does occur and gives rises to the phenotypic variability within the population. These variants occur at approximately every 1000 nucleotide base pairs of the code. This variation, where it occurs in >1% of the population is termed a polymorphism. The most common type of variant is a single letter change in the DNA sequence, termed a single nucleotide polymorphism (SNP). There are thought to be around 10 million common SNPs in the human genome. The individual specific risk of common diseases is thought to be influenced by the sum of many genetic variations, each potentially causing small changes in biological function and consequently subtle changes in

Patients vary in their osteolytic response to particulate wear debris. Some show little bone resorption in the presence of marked prosthesis wear whereas others undergo marked osteolysis following a small amount of prosthesis wear (Figure 6)57. Macrophage responsiveness to in-vitro particulate debris stimulation also varies between individual60,

developed femoral osteolysis, and had higher revision rates49.

high levels of activity49.

**5.4 Genetic factors** 

phenotype59.

conventional polyethylene bearing (Figure 5).

Fig. 5. Continuous dose-response relationship between prosthesis bearing wear and risk of osteolysis. Top panel shows data from a study of 115 cases and 115 controls after Charnley THA with consistent increase in proportion of subjects with osteolysis with increasing wear rate (by quintile)57. Bottom panel shows survivorship analysis in a cohort study of 319 hybrid THAs followed for a minimum of 10 years, and showing a similar dose-response relationship between osteolysis and polyethylene wear58.

follow up has shown that 24% of patients who have engaged in high levels of activity developed femoral osteolysis, and had higher revision rates49.

Traditionally the rate of polyethylene wear has been reported as a function of time. The results from ex-vivo hip simulator experiments have shown that the number of hip cycles is proportional to the rate of wear of prosthesis surface50. In vivo, there is a great range of wear rates between individual as a consequence of differing activity levels51. Several validated assessment tools have been developed to measure activity levels in arthroplasty populations52, and Schmalzried *et al* showed that wear in patients is a function of activity53. There are no clear guidelines outlining what levels of activity can be undertaken following THA although the proportion of patients participating in athletic activity following THA ranges between 52 – 83%54-56. Whilst low-impact activities such as walking, swimming and cycling have always been recommended following THA, some patients participate in more high-impact and competitive sports. The increasing participation in athletic activity and higher post-operative expectations can partly be explained by the increasing numbers of younger patients undergoing THA. 42% of men and 31% of women who underwent THA in England and Wales in 2009 were under the age of 65 years2. A large number of patients over the age of 65 are also participating in high levels of activity49.

Several investigators have shown a relationship between high levels of polyethylene wear and osteolysis/aseptic loosening, and the concept of a wear-rate 'threshold' (commonly defined as 0.1mm/year) below which osteolysis occurs very rarely, has been suggested. Wilkinson et al quantitated the association between wear and osteolysis and found no evidence to support this concept. In a case-control study of 230 hips after cemented Charnley THA with a metal on polyethylene bearing they showed that the risk of osteolysis increased with each quintile increase in wear, from very low levels of wear, below the suggested threshold, through to high levels57. They subsequently showed that the risk of osteolysis showed a similar pattern of consistently increasing risk ratio with each wear rate quintile in a separate cohort study of patients with 319 hybrid THAs using a metal on conventional polyethylene bearing (Figure 5).

### **5.4 Genetic factors**

282 Recent Advances in Arthroplasty

2 test, P<0.001

osteolysis no osteolysis

12345

wear quintile

Fig. 5. Continuous dose-response relationship between prosthesis bearing wear and risk of osteolysis. Top panel shows data from a study of 115 cases and 115 controls after Charnley THA with consistent increase in proportion of subjects with osteolysis with increasing wear rate (by quintile)57. Bottom panel shows survivorship analysis in a cohort study of 319 hybrid THAs followed for a minimum of 10 years, and showing a similar dose-response

relationship between osteolysis and polyethylene wear58.

0

10

20

30

number of subjects

40

50

Within a given ethnic population the sequence of DNA between individuals is 99.5% identical. However, variability within the code does occur and gives rises to the phenotypic variability within the population. These variants occur at approximately every 1000 nucleotide base pairs of the code. This variation, where it occurs in >1% of the population is termed a polymorphism. The most common type of variant is a single letter change in the DNA sequence, termed a single nucleotide polymorphism (SNP). There are thought to be around 10 million common SNPs in the human genome. The individual specific risk of common diseases is thought to be influenced by the sum of many genetic variations, each potentially causing small changes in biological function and consequently subtle changes in phenotype59.

Patients vary in their osteolytic response to particulate wear debris. Some show little bone resorption in the presence of marked prosthesis wear whereas others undergo marked osteolysis following a small amount of prosthesis wear (Figure 6)57. Macrophage responsiveness to in-vitro particulate debris stimulation also varies between individual60,

Risk Factors for Aseptic Loosening Following Total Hip Arthroplasty 285

the development of heterotopic ossification following THA. Malik *et al* have also shown associations between aseptic loosening and other candidate loci within the genes encoding matrix metalloproteinase 1 and the vitamin D receptor67, mannose-binding lectin70, and the

Recent studies using beadchip assays have shown that many genes are differentially expressed in wear debris-induced cells and tissues72-74, and have highlighted our limited understanding of the spectrum of biological mediators involved in the pathogenesis of osteolysis. The identification of further risk loci is required to further understanding of the pathogenesis of aseptic loosening. This would potentially allow for the development of screening tools, and provide investigational targets for prophylaxis or treatment with the aim of reducing the need for revision surgery, and its associated morbidity and mortality.

Prosthesis design factors, aside from those that modulate wear, contribute to risk of osteolysis. Modularity allows intra-operative adjustment of bearing surfaces, prosthesis length and offset. However, it also creates additional interfaces within the construct at which generation of debris through wear may occur. Such interfaces include the trunion between the femoral head and stem at which corrosive wear may occur, and backside wear between an acetabular liner and its shell at which abrasive wear may occur, and potentially several other prosthetic component junctions in highly modular systems. Hydroxyapatite coating of the prosthesis may prevent osteolysis following injection of intra-articular particles by sealing the implant-bone interface from their ingression though the promotion of osseointegration at this interface75,76, but may also be a source of third-body wear. Selection of bearing diameter is also a factor. The use of larger head sizes reduces the risk of dislocation, but increase volumetric wear77. The need for a thinner liner to accommodate the

The metal on polyethylene bearing couple remains the gold standard for THA. However, the manufacturing and sterilization process of polyethylene has changed over time with the aim of improving its wear rate characteristics. The earliest prostheses were made with non-cross-linked ultra-high molecular weight polyethylene (UHWPE) that was irradiated to render it sterile for patient use. The process of sterilization with ionizing radiation leads to cross-linking within the polymer. Cross-linking improves wear resistance of the material, but also causes the formation of free radicals. Free radical species cause the oxidation of UHMWPE over time. Polyethylene oxidation degrades UHMWPE, and

Several production techniques have been developed to reduce the generation of free radicals, including annealing and melting. Melting reduces free radical concentration more than annealing but adversely affects the yield stress and fatigue resistance of the polymer. Annealing below melting point has a less adverse effect on the mechanical properties, but is less effective than melting at free radical removal. Sterilization in an oxygen-free environment also produces more cross-linking and reduces free radical production78. Irradiation in an inert gas and vacuum packing is also now routinely carried out to reduce pre-implantation oxidation, however this does not prevent oxidation occurring in vivo. Faris

larger head may also cause increased contact stresses and an increase in wear.

RANK/OPG pathway71.

**6. Prosthesis risk factors** 

**6.1 Prosthesis design** 

**6.2 Polyethylene wear** 

decreases its wear resistance.

and monocytes (PBMCs) taken from patients with a susceptibility to osteolysis exhibit quantitatively greater inducible cytokine responses to particulate debris in-vitro versus patients without this susceptibility61 . It is suggested that this inter-patient variability may have a genetic basis.

Fig. 6. Patients exhibit variable osteolytic responses to wear debris. a) radiograph showing marked polyethylene wear, but no osteolytic response, b) radiograph showing mild wear but pronounced femoral and acetabular osteolysis with prosthesis loosening.

Variation within the genes encoding inflammatory cytokines have been associated with osteolysis. Wilkinson *et al* showed an association between variability within the DNA encoding the tumor necrosis factor (TNF) promoter region (dbSNP rs361525) and risk of osteolysis following THA62. Subjects with osteolysis were approximately twice as likely to carry the variant DNA code as those subjects with no osteolysis. This association has been replicated in an independent population by Ambruzova *et al63*. Gordon *et al* have reported genetic variation within the genes encoding Interleukin-1 receptor antagonist (IL-1RN) and IL-6 is also associated with osteolysis64. Similar associations have also been identified in other populations65-67.

Variation within genes that regulate bone turnover also associate with osteolysis. Gordon *et al* showed that carriage of the dbSNP rs288326 variant in the FRZB gene encoding secreted frizzled-related protein-3 (Frp3), a regulatory glycoprotein within the osteogenic Wnt signaling pathway that modulates mesenchymal stem cell differentiation of osteoblasts68, associated with susceptibility to osteolysis following THA69. Its carriage also associated with the development of heterotopic ossification following THA. Malik *et al* have also shown associations between aseptic loosening and other candidate loci within the genes encoding matrix metalloproteinase 1 and the vitamin D receptor67, mannose-binding lectin70, and the RANK/OPG pathway71.

Recent studies using beadchip assays have shown that many genes are differentially expressed in wear debris-induced cells and tissues72-74, and have highlighted our limited understanding of the spectrum of biological mediators involved in the pathogenesis of osteolysis. The identification of further risk loci is required to further understanding of the pathogenesis of aseptic loosening. This would potentially allow for the development of screening tools, and provide investigational targets for prophylaxis or treatment with the aim of reducing the need for revision surgery, and its associated morbidity and mortality.
