**2. Epidemiology**

### **2.1 Incidence**

Identification of prosthetic joint infection currently relies on diagnostic criteria, which include: histopathologic evidence of acute inflammation of periprosthetic tissue, presence of a sinus tract, macroscopic purulence surrounding the prosthesis observed intraoperatively or two or more positive microbiological cultures with the same organism isolated from the prosthetic joint fluid or tissue(Berbari, et al. 1998). We report a rate of prosthetic joint infection between 1.0 – 2.0% in primary lower limb arthroplasty and this is congruent with current literature (Dowsey & Choong 2008, 2009, Swan, et al. 2011). In the United States the rate of infection in knee and hip arthroplasty was 0.92% and 0.88% respectively in a recent review of Medicare data. (Kurtz, et al. 2008). The majority of arthroplasty infections occur in

Infection in Primary Hip and Knee Arthroplasty 415

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

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

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,

reoperation on the infected hip (Surin, et al. 1983).

for the bacteraemia(Bengtson & Knutson 1991).

**2.3 Microbiology** 

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