**2.2 Risk factors**

A number of preoperative risk factors for prosthetic joint infection have been identified and these include pre-existing patient co-morbidities such as obesity, diabetes mellitus, rheumatoid arthritis and a history of prior malignancy. Body mass index (BMI) greater than 40kg/m2 has been associated with a 9 fold increased risk of knee infection in our series (Dowsey & Choong 2009). Similar studies have shown that for every 1 kg/m2 increase in body mass index, there was an associated 8% increase in the risk of deep prosthetic joint infection. The association between obesity and deep prosthetic infections is particularly marked in the hip arthroplasty population with the risk of infection increasing from 0.9% in patients within the normal weight range, to 9.1% in morbidly obese(Choong, et al. 2007, Dowsey & Choong 2008). Diabetes mellitus also predisposes patients to deep prosthetic joint infection with 5.3% of diabetic patients developing a prosthetic joint infection in one study (Dowsey & Choong 2009, Yang, et al. 2001). Postulated mechanisms for the increased risk include impaired leucocyte function and impaired wound healing in diabetic patients. Rheumatoid arthritis has been associated with a higher risk of deep prosthetic joint infections. Patients with rheumatoid arthritis have been reported to have a greater than 2.5 fold increase in the risk of arthroplasty infection compared to patients with osteoarthritis (Bengtson & Knutson 1991, Poss, et al. 1984). Whether this is due to impaired immunity secondary to the underlying disease or whether it is a reflection of the increased use of immunosuppressive medications in this cohort remains unclear (Berbari, et al. 2006b). A diagnosis of malignancy not involving the index joint has been identified as a risk factor for the subsequent development of prosthetic joint infection(Berbari, et al. 1998).

Operative risk factors associated with deep prosthetic joint infection include higher American Society of Anaesthesiologist's (ASA) and National Nosocomial Infections Surveillance (NNIS) scores, bilateral surgery, knee arthroplasty, arthroplasty type and operating room conditions. Berbari et al showed that increasing NNIS score was associated with increasing risk of deep prosthetic joint infection; NNIS score 1 was associated with a 1.7 fold increase, increasing to 3.9 with a NNIS score of 2 (Berbari, et al. 1998). The ASA score, a component of the NNIS score, was also associated with an increased risk of arthroplasty infections (Pulido, et al. 2008). Pulido et al identified close to a six-fold increase in risk of prosthetic joint infection in patients undergoing simultaneous bilateral arthroplasty surgery. In the same study, knee arthroplasty, when compared to hip arthroplasty, was independently associated with a higher risk of developing deep prosthetic joint infection (Pulido, et al. 2008). The type of prosthesis used also appears to influence the risk of infection. A 20-fold increased risk of infection with metal hinged prosthetic knee joints compared to metal-to-plastic prostheses has been reported (Poss, et al. 1984).

A number of postoperative risk factors for prosthetic joint infection have also been identified. The most important of these appears to be postoperative wound complications including the presence of superficial infection and/or wound discharge (Bengtson & Knutson 1991, Surin, et al. 1983, Wymenga, et al. 1992). Superficial infection, occurs within 30 days of the operative procedure, only involves the superficial structures and additionally

the two years following prosthetic joint surgery. The incidence of knee arthroplasty infection within 2 years was 1.55% decreasing to 0.46% in the subsequent 8 years. Corresponding data in the hip arthroplasty population showed an incidence of 1.63% within 2 years and 0.59%

A number of preoperative risk factors for prosthetic joint infection have been identified and these include pre-existing patient co-morbidities such as obesity, diabetes mellitus, rheumatoid arthritis and a history of prior malignancy. Body mass index (BMI) greater than 40kg/m2 has been associated with a 9 fold increased risk of knee infection in our series (Dowsey & Choong 2009). Similar studies have shown that for every 1 kg/m2 increase in body mass index, there was an associated 8% increase in the risk of deep prosthetic joint infection. The association between obesity and deep prosthetic infections is particularly marked in the hip arthroplasty population with the risk of infection increasing from 0.9% in patients within the normal weight range, to 9.1% in morbidly obese(Choong, et al. 2007, Dowsey & Choong 2008). Diabetes mellitus also predisposes patients to deep prosthetic joint infection with 5.3% of diabetic patients developing a prosthetic joint infection in one study (Dowsey & Choong 2009, Yang, et al. 2001). Postulated mechanisms for the increased risk include impaired leucocyte function and impaired wound healing in diabetic patients. Rheumatoid arthritis has been associated with a higher risk of deep prosthetic joint infections. Patients with rheumatoid arthritis have been reported to have a greater than 2.5 fold increase in the risk of arthroplasty infection compared to patients with osteoarthritis (Bengtson & Knutson 1991, Poss, et al. 1984). Whether this is due to impaired immunity secondary to the underlying disease or whether it is a reflection of the increased use of immunosuppressive medications in this cohort remains unclear (Berbari, et al. 2006b). A diagnosis of malignancy not involving the index joint has been identified as a risk factor for

the subsequent development of prosthetic joint infection(Berbari, et al. 1998).

joints compared to metal-to-plastic prostheses has been reported (Poss, et al. 1984).

A number of postoperative risk factors for prosthetic joint infection have also been identified. The most important of these appears to be postoperative wound complications including the presence of superficial infection and/or wound discharge (Bengtson & Knutson 1991, Surin, et al. 1983, Wymenga, et al. 1992). Superficial infection, occurs within 30 days of the operative procedure, only involves the superficial structures and additionally

Operative risk factors associated with deep prosthetic joint infection include higher American Society of Anaesthesiologist's (ASA) and National Nosocomial Infections Surveillance (NNIS) scores, bilateral surgery, knee arthroplasty, arthroplasty type and operating room conditions. Berbari et al showed that increasing NNIS score was associated with increasing risk of deep prosthetic joint infection; NNIS score 1 was associated with a 1.7 fold increase, increasing to 3.9 with a NNIS score of 2 (Berbari, et al. 1998). The ASA score, a component of the NNIS score, was also associated with an increased risk of arthroplasty infections (Pulido, et al. 2008). Pulido et al identified close to a six-fold increase in risk of prosthetic joint infection in patients undergoing simultaneous bilateral arthroplasty surgery. In the same study, knee arthroplasty, when compared to hip arthroplasty, was independently associated with a higher risk of developing deep prosthetic joint infection (Pulido, et al. 2008). The type of prosthesis used also appears to influence the risk of infection. A 20-fold increased risk of infection with metal hinged prosthetic knee

between two to ten years (Kurtz, et al. 2010, Ong, et al. 2009)

**2.2 Risk factors** 

includes one of the following features: purulent discharge, isolation of micro-organisms through aseptic sampling techniques or clinical features of infection (Horan, et al. 1992). Applying this definition, patients with a postoperative surgical site infection had around a 36 fold increase in the risk of the subsequent development of a deep prosthetic wound in one study(Berbari, et al. 1998). Similarly, in another study examining patients with deep prosthetic infections acquired in the perioperative period, 25 of the 26 patients described preceding wound complications, which included; the persistent drainage of fluid from the wound, development of a haematoma under the wound, a superficial infection or a stitch abscesses (Poss, et al. 1984). Of note we have demonstrated that the use of closed suction drainage in total knee arthroplasty is protective of prosthetic knee infection and this may be due the role of a drainage tube in minimizing haematoma formation (Dowsey & Choong 2009). Early post-operative persistent discharge of fluid from the wound has been associated with a 3.2 times higher risk of deep prosthetic joint infection. Often in these cases the same pathogenic organisms isolated from the discharging fluid is later recovered at time of reoperation on the infected hip (Surin, et al. 1983).

Postoperative medical complications including atrial fibrillation and myocardial infarction have also been implicated as risk factors for deep prosthetic joint infection, with a 6-fold and 20-fold respective increase reported (Pulido, et al. 2008). One postulated mechanism to account for this association is that standard management of these medical conditions includes anticoagulation. In the postoperative period this may increase the risk of bleeding and haematoma formation near the wound, which in itself may increase the risk of infection. Secondly, these medical complications may necessitate longer inpatient hospital stay, which may be associated with nosocomial acquisition of infection. Allogenic blood transfusion was also identified as conferring a twofold increased risk of prosthetic infection, again the risk may be via an association with bleeding and haematoma formation near the wound, or possibly as a marker of complications and prolonged hospitalisation (Pulido, et al. 2008).

Nosocomial infections, particularly urinary tract infections have also been identified as risk factors for deep prosthetic joint infections. Surin et al demonstrated that patients with remote infections in the postoperative period were three times more likely to develop deep infections. Over three quarters of these infections were urinary tract infections. Interestingly however, there was no correlation between causative agents of the nosocomial infection and the micro-organism ultimately isolated from the infected prosthesis (Surin, et al. 1983). These results have been confirmed by other studies (Pulido, et al. 2008, Wilson, et al. 1990). Bengston et al highlighted the significance of skin infections in haematogenous seeding of the prosthesis. One third of patients with haematogenous seeding in this cohort had concurrent or preceding skin infections that were identified as the probable primary focus for the bacteraemia(Bengtson & Knutson 1991).

#### **2.3 Microbiology**

Staphylococcus species account for approximately half of all prosthetic joint infections; this includes *Staphylococcus aureus* and coagulase negative Staphylococcus species, both methicillin sensitive and resistant. Gram-negative bacilli infections and polymicrobial infections are the two next most common groups of pathogens described. Other grampositive bacteria such as Streptococcus and Enterococcus species occur less commonly (Bengtson & Knutson 1991, Berbari, et al. 1998, Fitzgerald, et al. 1977, Moran, et al. 2007, Pandey, et al. 2000, Pulido, et al. 2008, Steckelberg & Osmon 2000). Importantly, in all series,

Infection in Primary Hip and Knee Arthroplasty 417

may be much more delayed, particularly with low virulence organisms such as coagulase negative staphylococcus species (Steckelberg & Osmon 2000). Furthermore, up to 50% of suspected prosthetic joint infections of haematogenous origin present within the first two years (Deacon, et al. 1996). However it is important to note that distinguishing between whether an episode of bacteraemia led to haematogenous seeding of a prosthetic joint or whether the primary source of the bacteraemia was a subclinical prosthetic joint infection

The pathogenesis of prosthetic joint infections is intimately connected to the property of biofilm formation by microorganisms. The presence of this biofilm can have a critical effect on the likely success of treatment for a number of reasons. Bacteria can exist in two unique forms; the free living or planktonic forms characterised by rapid cellular division, and the stationary or sessile forms characterised by slower cellular division (Costerton 1999,

The sessile bacteria secrete an extracellular matrix or slime. Together the microorganisms and this matrix comprise what is known as 'the biofilm'. The abiotic matrix performs a number of functions including provision of anchorage onto structures to support the sessile colonies(Donlan & Costerton 2002). It also facilitates communication between bacteria within the biofilm. This communication termed 'quorum sensing', is analogous to the paracrine signalling in multicellular organisms and enables the bacteria to regulate their gene synthesis(Gristina & Costerton 2009). Importantly, the matrix can provide bacteria with protection from antimicrobial chemicals and from host defense mechanisms. This impairment of host defense mechanisms has been demonstrated in a number of in vitro models. For example, the extracellular slime produced by *Staphylococcus epidermidis* can

The concentration of antibiotic required to inhibit the growth of bacteria in biofilms is higher than that required to kill free-living bacteria. The mean inhibitory concentration (MIC) of many antibiotics is higher with the sessile forms than corresponding planktonic forms. Studies have demonstrated up to a 1000 fold increase in the MIC to particular antibiotics for bacteria moving from the planktonic to the sessile phenotype (Amorena, et al. 1999, Jones, et al. 2001, Rose & Poppens 2009, Schwank, et al. 1998, Souli & Giamarellou 1998, Stewart & Costerton 2001). This poses a major challenge for clinicians interpreting the reported antibiotic susceptibility results of bacteria, as our standard laboratory antibiotic susceptibility testing uses only the planktonic forms of bacteria. Newer technologies including the Calgary Biofilm Device can enable antibiotic susceptibility testing of the sessile phenotype of bacteria, but at present these are limited to a research setting and are not

There are a number of postulated mechanisms for the apparent resistance of biofilm residing bacteria to the effects of antibiotics. Firstly, the antibiotic may be deactivated at the surface of the biofilm. Secondly, the altered nutritional and biochemical environment within the biofilm may alter the activity of the antibiotics. Thirdly, antibiotics, particular cell wall active antibiotics such as betalactam antibiotics, rely on rapid growth and reproduction of the microorganism for their effect. These antibiotics are effective against the planktonic phenotype but have limited efficacy against the sessile phenotype as cellular turnover is greatly reduced. Finally the sessile forms act as 'spore-like' structures, which may act as a

inhibit the phagocytic activity of neutrophils(Shiau & Wu 1998).

can be problematic.

**3.2 The role of biofilms** 

Costerton, et al. 1995).

widely available(Ceri, et al. 1999).


a small number of cases meet the definition for prosthetic joint infection, and yet remain culture negative on standard microbiologic techniques.

1. (Moran, et al. 2007), 2. (Sharma, et al. 2008), 3. (Pandey, et al. 2000), 4. (Steckelberg & Osmon 2000), 5. (Pulido, et al. 2008), 6. (Berbari, et al. 1998), 7. (Bengtson & Knutson 1991), 8. (Fitzgerald, et al. 1977), 9. (McDonald, et al. 1989)

Table 1. Microbiological isolates in reported literature (percent)
