**4.1 Clinical features**

The clinical diagnosis of prosthetic joints is challenging. Many typical symptoms of infection are often absent. Pain is the predominant symptom of prosthetic joint infections and is present in 90 to 100% of patients. The presence of fever is variable with 9 to 43% of patients in most case series having documented elevated temperatures (Canner, et al. 1984, Inman, et al. 1984, McDonald, et al. 1989, Miley, et al. 1982, Morrey, et al. 1989, Windsor, et al. 1990). In acute infections, erythema and swelling of the joint are often present, but are less common in more chronic infections (Del Pozo & Patel 2009, Miley, et al. 1982, Zimmerli, et al. 2004). A discharging sinus is associated with chronic, indolent presentations (Del Pozo & Patel 2009).

Zimmerli et al classifies arthroplasty infections as: Early (developing in the first three months after surgery), Delayed (occurring three to 24 months after surgery) and Late (greater than 24 months). This classification roughly correlates to important observed differences in the causative pathogens; with virulent organisms such as *Staphylococcus aureus* characteristically presenting earlier and more indolent pathogens such as coagulase negative Staphylococcus usually presenting later (Zimmerli, et al. 2004).

### **4.2 Laboratory studies**

Peripheral blood leucocytosis is a poor predictor of infected arthroplasty; less than 10% of patients with an infected prosthesis have an elevated white cell count in most series (Canner, et al. 1984, Inman, et al. 1984, Zimmerli, et al. 2004). Other biochemical tests, such as the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are more useful diagnostic tests for these infections. For patients with proven infection of knee or hip arthroplasty, the ESR had a sensitivity of 81-92% and a specificity of 90-96%, while the CRP had a sensitivity of 84-89% and a specificity of 83-96% (Bottner, et al. 2007, Spangehl, et al. 1999). There are however, limitations to the diagnostic utility of the ESR and CRP. These markers are normally elevated after primary uncomplicated arthroplasty; the ESR peaks in the first week and may remain elevated for up to a year, while the CRP peaks at

nidus for later relapse of infection (Costerton 1999, Stewart & Costerton 2001, Trampuz, et

The properties of the biofilm alter with time; with age many biofilms become increasingly resistant to antibiotics. Monzon et al demonstrated the efficacy of vancomycin against *Staphylococcus epidermidis* decreased as a biofilm aged. This phenomenon was not consistent with all antibiotics; the activity of rifampicin and tetracyclines was not altered (Monzon, et al. 2002). Using Ribosomal RNA Fluorescence In Situ Hybridization studies, Poulson et al assessed the growth rate of biofilms and demonstrated that the cellular turnover was significantly higher in younger biofilms compared to established biofilms (Poulsen, et al. 1993). This finding could account for the difference to antimicrobial susceptibility observed. Implant factors are also recognised to play a role in the pathogenesis of infection. Biochemical properties of prosthetic material influences bacterial adhesion and may impair host immune responses. For example, methyl methacrylate cement has been shown to

The clinical diagnosis of prosthetic joints is challenging. Many typical symptoms of infection are often absent. Pain is the predominant symptom of prosthetic joint infections and is present in 90 to 100% of patients. The presence of fever is variable with 9 to 43% of patients in most case series having documented elevated temperatures (Canner, et al. 1984, Inman, et al. 1984, McDonald, et al. 1989, Miley, et al. 1982, Morrey, et al. 1989, Windsor, et al. 1990). In acute infections, erythema and swelling of the joint are often present, but are less common in more chronic infections (Del Pozo & Patel 2009, Miley, et al. 1982, Zimmerli, et al. 2004). A discharging sinus is associated with chronic, indolent

Zimmerli et al classifies arthroplasty infections as: Early (developing in the first three months after surgery), Delayed (occurring three to 24 months after surgery) and Late (greater than 24 months). This classification roughly correlates to important observed differences in the causative pathogens; with virulent organisms such as *Staphylococcus aureus* characteristically presenting earlier and more indolent pathogens such as coagulase negative

Peripheral blood leucocytosis is a poor predictor of infected arthroplasty; less than 10% of patients with an infected prosthesis have an elevated white cell count in most series (Canner, et al. 1984, Inman, et al. 1984, Zimmerli, et al. 2004). Other biochemical tests, such as the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are more useful diagnostic tests for these infections. For patients with proven infection of knee or hip arthroplasty, the ESR had a sensitivity of 81-92% and a specificity of 90-96%, while the CRP had a sensitivity of 84-89% and a specificity of 83-96% (Bottner, et al. 2007, Spangehl, et al. 1999). There are however, limitations to the diagnostic utility of the ESR and CRP. These markers are normally elevated after primary uncomplicated arthroplasty; the ESR peaks in the first week and may remain elevated for up to a year, while the CRP peaks at

inhibit complement and lymphocyte activity (Panush & Petty 1978, Petty 1978).

al. 2003, Zimmerli 2006).

**4. Diagnosis** 

**4.1 Clinical features** 

**4.2 Laboratory studies** 

presentations (Del Pozo & Patel 2009).

Staphylococcus usually presenting later (Zimmerli, et al. 2004).

day 2 and may remain elevated for 3 weeks (Aalto, et al. 1984, Larsson, et al. 1992, Shih, et al. 1987).

The search for other biochemical markers of infection has included interleukin 6 (Il-6), tumour necrosis factor α (TNF-α) and procalcitonin C. Il-6 and TNF-α are cytokines released by monocytes and macrophages in the setting of infection (Bottner, et al. 2007). Procalcitonin is a precursor of calcitonin, and has been shown to be a specific marker of bacterial sepsis (Fernandez Lopez, et al. 2003). In a review by Bottner et al of 78 patients undergoing revision arthroplasties Il-6, TNF-α and procalcitonin were all significantly elevated in patients with confirmed septic loosening. The sensitivity and specificity respectively of Il-6 was 95% and 87%, TNF-α 43% and 94% and procalcitonin 33% and 98%. (Bottner, et al. 2007). Il-6 is elevated in the post-operative period for primary arthroplasty however, in a study by Shah et al, Il-6 was shown to return to normal levels within 2 days of the operation. Therefore there is potential diagnostic utility of Il-6 over CRP and ESR in the early post-operative period if infection is suspected, particularly in the first 21 days (Shah, et al. 2009).

Synovial fluid characteristics can be used to assist in diagnosis of prosthetic joint infections. In a study by Trampuz and colleagues, the leucocyte count was significantly higher in patients with prosthetic joint infection with a median of 18.9 x 103/µL (range, 0.3 to 178 x 103/µL) compared to a median leucocyte count of 0.3 x 103/µL (range, 0.1 to 16 x 103/µL) in patients with aseptic loosening. Using receiver operating characteristic (ROC) curves the authors found a synovial total white cell count 1.7 x 103/µL and a leucocyte differential of greater than 65% neutrophils had a sensitivity and specificity of 94%, 88% and 97%, 98% respectively (Trampuz, et al. 2004).

### **4.3 Radiological studies**

Plain radiographs lack sensitivity and specificity in diagnosing septic arthroplasty. Findings such as lucency around the prosthesis can be noted in both septic and aseptic loosening situations (Figure 1 A-D). In early infection plain radiographs are frequently normal (Miller 2005).

Technetium-Methylene Diphosphonate (MDP) bone scintigraphy is a sensitive test for prosthetic joint infection (Figure 1 E-G), but lacks specificity, as it does not differentiate between aseptic and septic loosening(Ghanem, et al. 2009). The bone scan can also remain positive for a year following primary arthroplasty. Bone scan does have a high negative predictive value therefore bone scans potentially can be used to exclude infection in the setting of a painful prosthetic joint (Smith, et al. 2001). Similar findings have been documented with newer modalities such as 18F-Fluoro-deoxyglucose positron emission tomography (FDG-PET) (Delank, et al. 2006, Zoccali, et al. 2009). A recent meta-analysis of FDG-PET reported a sensitivity of 82.1% and specificity of 86.6% for the presence of prosthetic joint infection, and hence this may be a useful test if available(Kwee, et al. 2008).

Computer tomography (CT) and magnetic resonance imaging are not considered useful imaging modalities due to artefact from the metal prosthesis interfering with interpretation of imaging findings. However newer CT scanners can minimise this effect and may be useful in detecting abnormalities of the soft tissues in periprosthetic infections (Figure 2 A-E) but do not diagnose periprosthetic bone abnormalities well (Cyteval, et al. 2002)

Infection in Primary Hip and Knee Arthroplasty 421

Fig. 2. (A) Localised infective sinus at the centre of incision used for total knee joint

component. (C) Magnified image showing obvious periprosthetic lysis (arrows). (D) Computer tomogram showing lysis under tibial component extending through medial cortex as cloaca (arrow). (E) Computer tomogram demonstrating soft tissue abscess

formation (arrows) in continuity with intramedullary suppuration.

replacement. (B) plain radiograph showing periprosthetic sclerosis and lysis under the tibial

Fig. 1. (A) Painful (left) cementless hip prosthesis in situ. (B) Note extrinsic scalloping of anterior cortex of femoral diaphysis (box). (C) Magnified image of anterior femoral cortex with extrinsic scalloping (arrows) caused by soft tissue abscess (D). (E) Nuclear bone scan (TcMDP) demonstrating mild uptake over left proximal femur. Indium white cell scan at (F) 4 hours and (G) showing marked retention of nuclear tracer at 20 hours.

Fig. 1. (A) Painful (left) cementless hip prosthesis in situ. (B) Note extrinsic scalloping of anterior cortex of femoral diaphysis (box). (C) Magnified image of anterior femoral cortex with extrinsic scalloping (arrows) caused by soft tissue abscess (D). (E) Nuclear bone scan (TcMDP) demonstrating mild uptake over left proximal femur. Indium white cell scan at (F)

4 hours and (G) showing marked retention of nuclear tracer at 20 hours.

Fig. 2. (A) Localised infective sinus at the centre of incision used for total knee joint replacement. (B) plain radiograph showing periprosthetic sclerosis and lysis under the tibial component. (C) Magnified image showing obvious periprosthetic lysis (arrows). (D) Computer tomogram showing lysis under tibial component extending through medial cortex as cloaca (arrow). (E) Computer tomogram demonstrating soft tissue abscess formation (arrows) in continuity with intramedullary suppuration.

Infection in Primary Hip and Knee Arthroplasty 423

hybridization (FISH) and immunofluorescent microscopy (IFM). Both PCR and FISH target specific regions of bacterial genetic material, commonly bacterial ribosomal RNA (rRNA). The advantage of using rRNA is that it is highly conserved in bacterial species compared to most protein encoding genes. Both methods can use broad range oligonucleotide primers or more targeted primers including genus and species specific

A number of studies investigating the role of bacterial 16s rRNA PCR have been performed. Sensitivities of this technique ranged from 63-100% in detecting bacteria involved in prosthetic joint infection (De Man, et al. 2009, Hoeffel, et al. 1999, Mariani, et al. 1996, Moojen, et al. 2007). In a study by Mariani et al 50 patients with symptoms following total knee arthroplasty underwent synovial fluid and intraoperative tissue sampling for culture and PCR; cultures were positive in fifteen specimens compared to 32 specimens when PCR was applied (Mariani, et al. 1996). Likewise Tunney et al used PCR in a study of 120 patients undergoing prosthetic hip joint revision. The explanted prosthesis underwent ultrasonification and this fluid was cultured and underwent 16s DNA PCR. Standard microbiologic cultures were positive in 22% of patients, compared to 72% of patients with positive results from PCR (Tunney, et al. 1999). The limitation of these studies was a paucity of correlation with clinical or histological features of infection. In a review of 34 patients with confirmed prosthetic joint infection Vandercam et al found that PCR was positive in 31 of 34 patients (91.2%), compared to positive microbiological culture in 22 of 34 patients (64.7%). Of import, eight of the nine patients with positive PCR but negative culture results had received antibiotic therapy in the prior ten days (Vandercam, et al. 2008). Despite these promising results, the weakness of 16s ribosomal RNA PCR techniques is the low specificity and high false positive rate. In a study by Clarke et al 29% of the patients without septic arthritis (on the basis of clinical, radiological, biochemical, intraoperative findings, culture and histology) had positive PCR results, this was particularly pronounced in the cohort undergoing revision arthroplasty for aseptic loosening where 46% of patients had positive PCR (Clarke, et al. 2004). The high false positive rate may be due to a number of factors including contamination of specimen or the reagents and detection of necrotic bacterial DNA (Bauer, et al. 2006). Importantly though, many patients labelled as having aseptic loosening may in fact have had low grade chronic infection contributing to prosthesis loosening. Given that there is no gold standard to define prosthetic joint infection, the

FISH is a technique that uses labelled oligonucleotide probes that hybridise to specific genetic regions on bacteria and are subsequently visualised using fluorescent microscopy or flow cytometry (Amann & Fuchs 2008, Moter & Gobel 2000). Probes to detect bacterial rRNA or other genetic targets are available and these include species specific probes, therefore allowing identification and simultaneous observation of the different bacteria. FISH also allows an appreciation of the architectural arrangement of the organisms within the biofilm which can assist in differentiating true infections from contamination (McDowell & Patrick 2005, Moter & Gobel 2000). In orthopaedic infections it has been demonstrated that *Staphylococcus aureus* and *Staphylococcus epidermidis* could be visualised and differentiated in an experimental biofilm. Additionally, in a clinical case of septic loosening of a hip prosthesis, *Staphylococcus epidermidis* was visualised using FISH techniques in periprosthetic tissue samples(Krimmer, et al. 1999). FISH has otherwise not yet been widely

primers (Amann & Fuchs 2008).

specificity of PCR remains difficult to judge.

applied to prosthetic joint infections in a clinical setting.

### **4.4 Histopathology diagnosis**

Intraoperative frozen section histopathologic studies of periprosthetic tissue can be used as an adjunctive test for the diagnosis of prosthetic joint infections. An early paper showed a correlation between the polymorphonuclear leucocyte (PMN) count in tissue on histopathologic examination and the diagnosis of infection (Mirra, et al. 1976). Subsequent studies using frozen section histopathology for revision arthroplasty (using a PMN count of five to ten cells per high power field to diagnose infection) had a sensitivity of 50-93% and sensitivity of 77-100% (Bori, et al. 2006, Frances Borrego, et al. 2007, Ko, et al. 2005, Nunez, et al. 2007). It should be noted that inflammatory conditions such as rheumatoid arthritis may also cause a high PMN count, hence lowering specificity (Mirra, et al. 1976).

### **4.5 Microbiology diagnosis**

The identification of the causative pathogen in a prosthetic device infection is of paramount importance. It allows for the institution of appropriate management strategies for infection including selection of the most appropriate antibiotic to target the pathogen, while minimising unnecessary antibiotic overuse, thus decreasing the incidence of drug toxicity and generally permitting simpler drug regimens to improve patient adherence.

It has earlier been noted that culture negative prosthetic joint infections continue to occur. Recent studies have focused on methods to increase the sensitivity of microbiological diagnostic techniques to address this problem. In a prospective study, which aimed to establish microbiological criteria for the diagnosis of prosthetic joint infection in revision arthroplasty, Atkins et al found that the isolation of indistinguishable microorganisms from three or more periprosthetic tissue samples has a sensitivity of 65% and a specificity of 99.6% for prosthetic joint infection. Utilising mathematical modelling the authors recommended that five to six intraoperative specimens of periprosthetic tissue be obtained to optimise the likelihood of a microbiologic diagnosis in prosthetic joint infection. They also noted that routine gram staining of periprosthetic tissue at revision arthroplasty had a very low sensitivity (12%) and the authors recommended that gram stain should be abandoned in revision arthroplasty cases, instead relying on culture (Atkins, et al. 1998).

Prolonged cultures may also help to improve the diagnostic yield. An increase in positive culture results of 24.6% when culture incubation of periprosthetic tissue samples was increased from 3 to fourteen days has been reported and in particular the isolation of fastidious organisms, such as Propionibacterium species was increased (Schafer, et al. 2008).

A number of techniques have been developed in an attempt to disrupt the biofilm and increase the yield of microbiological cultures. One such technique is ultrasonification whereby the explanted prosthesis is placed in a sterile polyethylene bag then in a sterile anaerobic jar, Ringer's solution is added and sonification is performed. The sonicate fluid is cultured aerobically and anaerobically. One study comparing sonification to standard tissue culture involving 331 patients of whom 79 had prosthetic joint infections; sonification yielded an additional 14 microbiological diagnosis with a reported sensitivity of 78.5% and specificity of 98.8%. The authors noted that sonification was particularly useful in cases where patients had received antibiotics perioperatively (Trampuz, et al. 2007).

#### **4.6 Molecular techniques**

Newer molecular techniques have been applied to prosthetic joint infections to increase the diagnostic yield including polymerase chain reaction (PCR), fluorescent in situ

Intraoperative frozen section histopathologic studies of periprosthetic tissue can be used as an adjunctive test for the diagnosis of prosthetic joint infections. An early paper showed a correlation between the polymorphonuclear leucocyte (PMN) count in tissue on histopathologic examination and the diagnosis of infection (Mirra, et al. 1976). Subsequent studies using frozen section histopathology for revision arthroplasty (using a PMN count of five to ten cells per high power field to diagnose infection) had a sensitivity of 50-93% and sensitivity of 77-100% (Bori, et al. 2006, Frances Borrego, et al. 2007, Ko, et al. 2005, Nunez, et al. 2007). It should be noted that inflammatory conditions such as rheumatoid arthritis may

The identification of the causative pathogen in a prosthetic device infection is of paramount importance. It allows for the institution of appropriate management strategies for infection including selection of the most appropriate antibiotic to target the pathogen, while minimising unnecessary antibiotic overuse, thus decreasing the incidence of drug toxicity

It has earlier been noted that culture negative prosthetic joint infections continue to occur. Recent studies have focused on methods to increase the sensitivity of microbiological diagnostic techniques to address this problem. In a prospective study, which aimed to establish microbiological criteria for the diagnosis of prosthetic joint infection in revision arthroplasty, Atkins et al found that the isolation of indistinguishable microorganisms from three or more periprosthetic tissue samples has a sensitivity of 65% and a specificity of 99.6% for prosthetic joint infection. Utilising mathematical modelling the authors recommended that five to six intraoperative specimens of periprosthetic tissue be obtained to optimise the likelihood of a microbiologic diagnosis in prosthetic joint infection. They also noted that routine gram staining of periprosthetic tissue at revision arthroplasty had a very low sensitivity (12%) and the authors recommended that gram stain should be abandoned in

Prolonged cultures may also help to improve the diagnostic yield. An increase in positive culture results of 24.6% when culture incubation of periprosthetic tissue samples was increased from 3 to fourteen days has been reported and in particular the isolation of fastidious organisms, such as Propionibacterium species was increased (Schafer, et al. 2008). A number of techniques have been developed in an attempt to disrupt the biofilm and increase the yield of microbiological cultures. One such technique is ultrasonification whereby the explanted prosthesis is placed in a sterile polyethylene bag then in a sterile anaerobic jar, Ringer's solution is added and sonification is performed. The sonicate fluid is cultured aerobically and anaerobically. One study comparing sonification to standard tissue culture involving 331 patients of whom 79 had prosthetic joint infections; sonification yielded an additional 14 microbiological diagnosis with a reported sensitivity of 78.5% and specificity of 98.8%. The authors noted that sonification was particularly useful in cases

Newer molecular techniques have been applied to prosthetic joint infections to increase the diagnostic yield including polymerase chain reaction (PCR), fluorescent in situ

also cause a high PMN count, hence lowering specificity (Mirra, et al. 1976).

and generally permitting simpler drug regimens to improve patient adherence.

revision arthroplasty cases, instead relying on culture (Atkins, et al. 1998).

where patients had received antibiotics perioperatively (Trampuz, et al. 2007).

**4.4 Histopathology diagnosis** 

**4.5 Microbiology diagnosis** 

**4.6 Molecular techniques** 

hybridization (FISH) and immunofluorescent microscopy (IFM). Both PCR and FISH target specific regions of bacterial genetic material, commonly bacterial ribosomal RNA (rRNA). The advantage of using rRNA is that it is highly conserved in bacterial species compared to most protein encoding genes. Both methods can use broad range oligonucleotide primers or more targeted primers including genus and species specific primers (Amann & Fuchs 2008).

A number of studies investigating the role of bacterial 16s rRNA PCR have been performed. Sensitivities of this technique ranged from 63-100% in detecting bacteria involved in prosthetic joint infection (De Man, et al. 2009, Hoeffel, et al. 1999, Mariani, et al. 1996, Moojen, et al. 2007). In a study by Mariani et al 50 patients with symptoms following total knee arthroplasty underwent synovial fluid and intraoperative tissue sampling for culture and PCR; cultures were positive in fifteen specimens compared to 32 specimens when PCR was applied (Mariani, et al. 1996). Likewise Tunney et al used PCR in a study of 120 patients undergoing prosthetic hip joint revision. The explanted prosthesis underwent ultrasonification and this fluid was cultured and underwent 16s DNA PCR. Standard microbiologic cultures were positive in 22% of patients, compared to 72% of patients with positive results from PCR (Tunney, et al. 1999). The limitation of these studies was a paucity of correlation with clinical or histological features of infection. In a review of 34 patients with confirmed prosthetic joint infection Vandercam et al found that PCR was positive in 31 of 34 patients (91.2%), compared to positive microbiological culture in 22 of 34 patients (64.7%). Of import, eight of the nine patients with positive PCR but negative culture results had received antibiotic therapy in the prior ten days (Vandercam, et al. 2008). Despite these promising results, the weakness of 16s ribosomal RNA PCR techniques is the low specificity and high false positive rate. In a study by Clarke et al 29% of the patients without septic arthritis (on the basis of clinical, radiological, biochemical, intraoperative findings, culture and histology) had positive PCR results, this was particularly pronounced in the cohort undergoing revision arthroplasty for aseptic loosening where 46% of patients had positive PCR (Clarke, et al. 2004). The high false positive rate may be due to a number of factors including contamination of specimen or the reagents and detection of necrotic bacterial DNA (Bauer, et al. 2006). Importantly though, many patients labelled as having aseptic loosening may in fact have had low grade chronic infection contributing to prosthesis loosening. Given that there is no gold standard to define prosthetic joint infection, the specificity of PCR remains difficult to judge.

FISH is a technique that uses labelled oligonucleotide probes that hybridise to specific genetic regions on bacteria and are subsequently visualised using fluorescent microscopy or flow cytometry (Amann & Fuchs 2008, Moter & Gobel 2000). Probes to detect bacterial rRNA or other genetic targets are available and these include species specific probes, therefore allowing identification and simultaneous observation of the different bacteria. FISH also allows an appreciation of the architectural arrangement of the organisms within the biofilm which can assist in differentiating true infections from contamination (McDowell & Patrick 2005, Moter & Gobel 2000). In orthopaedic infections it has been demonstrated that *Staphylococcus aureus* and *Staphylococcus epidermidis* could be visualised and differentiated in an experimental biofilm. Additionally, in a clinical case of septic loosening of a hip prosthesis, *Staphylococcus epidermidis* was visualised using FISH techniques in periprosthetic tissue samples(Krimmer, et al. 1999). FISH has otherwise not yet been widely applied to prosthetic joint infections in a clinical setting.

Infection in Primary Hip and Knee Arthroplasty 425

infection due to *Staphylococcus aureus*, absence of implant loosening and strict adherence to treatment (Trebse, et al. 2005). However, antibiotic suppression alone is generally reserved for patients with significant comorbidities in whom surgery is contraindicated, who are without evidence of systemic infection and where tolerable oral antibiotics are available. Given the low likelihood of cure, many clinicians view this as long term, often lifelong suppressive therapy, embarked upon without curative intent (Steckelberg & Osmon 2000,

Interpretation of the current literature describing the outcomes of patients having one- and two-stage exchange procedures is challenging owing to the heterogeneity of the patient populations, the causative organisms and the surgical techniques (including use of antibiotic impregnated cement), the differences in the duration of patient follow up and probable publication bias. The greatest concern with one-stage exchange procedures is the implantation of the prosthesis into an infected field with subsequent reinfection of the revised arthroplasty. In one-stage exchange, reported success rates range from 38-100%; but there is significant variability in the definition of success which includes freedom from infection, freedom from pain or simply the presence of a functional joint (Callaghan, et al. 1999, Jamsen, et al. 2009, Steckelberg & Osman 2000). In examination of the outcomes of onestage exchange revision hip arthroplasty, 80% (range 57-92%) of patients have been reported to remain infection free after one-stage exchange without the use of antibiotic cement (Steckelberg & Osman 2000). When antibiotic impregnated cement was used, 88% (range 76- 100%) of patients have been reported to remain infection free at follow up (Callaghan, et al. 1999, Jackson & Schmalzried 2000, Langlais 2003, Steckelberg & Osman 2000). Results for one-stage exchange in knee arthroplasty revision are in general worse than for hips, with only 65% (range 57-100%) of patients remaining free of recurrence of infection at follow up (Steckelberg & Osmon 2000). On the basis of these results, one-stage exchange of an infected prosthesis is rarely advised for prosthetic knee infections(Trampuz & Zimmerli 2008, Zimmerli, et al. 2004). There are, however some advantages with one-stage exchange; patients undergo a single operation and generally require a shorter period of hospitalisation

Consensus recommendations for one-stage exchange suggest that it should only be considered where there is minimal soft tissue damage and where less virulent organisms are involved (Hirakawa, et al. 1998, Jackson & Schmalzried 2000, Miley, et al. 1982, Trampuz & Zimmerli 2008, Zimmerli, et al. 2004). The presence of sinus tract is considered a relative contraindication for one-stage exchange. Ideally the causative agent should be known prior to resection arthroplasty and treatment commenced

In two-stage exchange procedures, reimplantation is delayed for a variable length of time from 2 weeks to several months. Spacers impregnated with antibiotic are commonly inserted to maintain limb length and improve patient mobility during that interval (Leunig, et al. 1998). Antibiotics with activity against the isolated pathogen are administered for at least 6 weeks. Tissue samples are often routinely taken from the periprosthetic tissue at the time of reimplantation for microbiological culture to assess the efficacy of the interim treatment (Insall, et al. 1983, Wilson, et al. 1990, Windsor, et al. 1990). In infections with

Zimmerli, et al. 2004).

in total.

preoperatively(Zimmerli, et al. 2004).

**5.3 Exchange arthroplasty** 

Immunofluorescence microscopy (IFM) is another novel nonculture technique for diagnosing prosthetic joint infections. In immunofluorescence microscopy, samples are mixed with monoclonal antibodies (MAb) to specific antigens on bacterial cell walls. Samples are then incubated with a second antibody conjugated with a fluorescent dye. The bacteria are then visualised using fluorescence microscopy(Tunney, et al. 1999). As with FISH, IFM can be used to assess the biofilm structure and can detect multiple pathogens (McDowell & Patrick 2005, Tunney, et al. 1999). In the study by Tunney et al IFM was performed on the sonicate fluid from explanted prostheses using Mab for both Propionibacterium and Staphylococcus species. (Tunney, et al. 1999).
