**3.6.2 Culture duration**

"Standard" cultivation periods (mostly ≤ 7 days in the literature) are generally questionable in infections where biofilms are involved, due to low cell counts of planktonic bacteria and impaired growth rates of sessile organisms in the biofilm. However, the issue was not addressed for a long time. Prolonged cultivation for 14 days was described sporadically (Ince et al., 2004) and even included as a standard recommendation into German practice guidelines, but not assessed under controlled conditions. Thus, our group systematically evaluated a 14-day culture period with periprosthetic tissue samples from hip and knee patients (Schäfer et al., 2008).

Using the algorithm described under 3.1.1 to distinguish infecting agents from contaminating strains, only 74% of the infections (caused by "early" agents) were found within the first week of cultivation (Schäfer et al., 2008). In the second week we not only identified a significant amount of additional infections, but also a completely different spectrum of causative ("late") species (Table 1).


Table 1. Spectrum of bacteria detected over a 14-day cultivation period.

Infections in Hip and Knee Arthroplasty:

**0**

**0**

**3.8 Stage at which samples are attained** 

**10**

**20**

**30**

**40**

**no. of samples**

**B**

the χ2 test.

**50**

**60**

**70**

**10**

**20**

**30**

**40**

**no. of samples**

**A**

**50**

**60**

**70**

Challenges to and Chances for the Microbiological Laboratory 447

**≥ 21 no. of culture-positive tissues**

**≥ 21 no. of culture-positive tissues**

Fig. 2. Correlation between positive histology and the number of culture-positive tissue biopsies. (A) early species. (B) late species. Statistical significance was demonstrated by

Although an additional risk and cost factor at first glance, preoperative evaluation of tissue samples in addition to joint aspiration can be helpful to identify the causative agent of arthroplasty infection accurately before the revision is carried out (Fink et al., 2008). This enables a one-stage replacement procedure, if clinically viable. Moreover, it allows to design

*P* **= .007** 

*P* **< .0001** 

 **histology + histology ø**

 **histology + histology ø**

Regarding the late species, the sensitivity would have been merely 27% if cultures had been monitored for only 7 days (Fig. 1).

Fig. 1. Detection rates of early and late species depending on the cultivation period. Whisker bars span the Hall-Wellner 95% confidence intervals.

### **3.7 Discrimination between infection and contamination**

There is no standardized procedure which would define infection over contamination accurately. Regarding tissue samples, usually a combined algorithm of neutrophil infiltration scores (2.3.2) and culture detection of identical organisms from multiple tissue samples is used. However, due to the missing consensus criteria the approaches vary considerably between studies (Atkins et al., 1998; Bori et al., 2007; Fink et al., 2008; Ko et al., 2005; Mirra et al., 1976; Pandey et al., 1999).

A problem we encountered was that the algorithms we have adopted to define infections (Atkins et al., 1998; Pandey et al., 1999; Virolainen et al., 2002) were evaluated in the context of "standard" microbiological cultivation periods. Thus, with prolonged culture duration (3.6.2) a larger amount of contaminants might have impaired the significance of this algorithm. However, our findings allowed us to refute the concern that prolonged culture of tissue biopsies could lead to over-proportional contamination rates (Schäfer et al., 2008). It became clear that among both the "early" and the "late" agents (Table 1) a highly significant correlation existed between positive histology and the number of culture-positive tissues (Fig. 2).

Regarding the late species, the sensitivity would have been merely 27% if cultures had been

Fig. 1. Detection rates of early and late species depending on the cultivation period. Whisker

**0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 days of incubation**

 **early species late species**

There is no standardized procedure which would define infection over contamination accurately. Regarding tissue samples, usually a combined algorithm of neutrophil infiltration scores (2.3.2) and culture detection of identical organisms from multiple tissue samples is used. However, due to the missing consensus criteria the approaches vary considerably between studies (Atkins et al., 1998; Bori et al., 2007; Fink et al., 2008; Ko et al.,

A problem we encountered was that the algorithms we have adopted to define infections (Atkins et al., 1998; Pandey et al., 1999; Virolainen et al., 2002) were evaluated in the context of "standard" microbiological cultivation periods. Thus, with prolonged culture duration (3.6.2) a larger amount of contaminants might have impaired the significance of this algorithm. However, our findings allowed us to refute the concern that prolonged culture of tissue biopsies could lead to over-proportional contamination rates (Schäfer et al., 2008). It became clear that among both the "early" and the "late" agents (Table 1) a highly significant correlation existed between positive histology and the number of

bars span the Hall-Wellner 95% confidence intervals.

2005; Mirra et al., 1976; Pandey et al., 1999).

culture-positive tissues (Fig. 2).

**3.7 Discrimination between infection and contamination** 

monitored for only 7 days (Fig. 1).

**0%**

**10%**

**20%**

**30%**

**40% 50%**

**60%**

**detection rate**

**70%**

**80%**

**90%**

**100%**

Fig. 2. Correlation between positive histology and the number of culture-positive tissue biopsies. (A) early species. (B) late species. Statistical significance was demonstrated by the χ2 test.

### **3.8 Stage at which samples are attained**

Although an additional risk and cost factor at first glance, preoperative evaluation of tissue samples in addition to joint aspiration can be helpful to identify the causative agent of arthroplasty infection accurately before the revision is carried out (Fink et al., 2008). This enables a one-stage replacement procedure, if clinically viable. Moreover, it allows to design

Infections in Hip and Knee Arthroplasty:

our own approach.

a routine basis.

laboratory.

(Fink et al., 2008, 2009).

**4.1.2 Hip infections** 

**4.1.3 Knee infections** 

samples are obtained for histological analysis.

Challenges to and Chances for the Microbiological Laboratory 449

during revision surgery. We have experienced that the inflammation process can be assessed with high reliability if 5 samples each are obtained for culture and for histologic analysis. The definition that i) growth of indistinguishable bacteria in ≥ 2 specimens or ii) microbial growth in one specimen combined with a histology score of 3+ (≥ 5 neutrophils per high-power field in 10 fields) (Atkins et al., 1998; Pandey et al., 1999; Virolainen et al., 2002) has proven feasible with respect to the clinical outcomes (Fink et al., 2008, 2009). Until now we use native tissue biopsies for culture. Tissue mincing is simple to perform and not too prone to contamination if carried out under a laminar air flow workbench. In our opinion, the prolonged incubation period of 14 days we carry out before cultures are cleared is a decisive measure. The persuasiveness of this approach has been shown in detail in 3.6.2. Although it would be interesting to compare the allegedly most sensitive sonication culture method directly with prolonged tissue cultivation, we doubt that the cumbersome and potentially contamination-prone sonication concept would prove significantly superior to

As we are convinced that prolonged cultivation over 14 days is the key to detecting infecting organisms with optimal sensitivity, we also currently refrain from using PCR techniques on

If overnight storage of unprocessed samples is necessary, which occurs in less than 5% of cases in our setting, tissue specimens are kept at 4°C and processed at the laboratory the next morning with highly reproducible results. Synovial fluid, when inoculated into pediatric blood culture vials immediately post-drawing, is stable for at least 24 hours at room temperature. Subsequent supplementation with the appropriate enhancing medium, which is necessary to cultivate blood-free sample fluids, can then be done at the

Taken together, our diagnostic measures have contributed significantly to the high eradication rates we observe with the treatment of both hip and knee arthroplasty infections

By default, we carry out two-stage revisions of infected hips in our clinic (Fink et al., 2009). Localized antibiotics applied via cement spacer and systemic antibiotics are customized for

The general sampling algorithm is depicted schematically in Fig. 3. In addition, we obtain pre-operative tissue biopsies for culture and histology if the joint aspiration culture is

We perform pre-operative tissue biopsies rather than joint aspiration if it is clear that revision operation is necessary due to an unstable implant (Fig. 4). Five biopsies for culture are obtained in a blind fashion without instillating fluid in the intra-articular space (to avoid possible losses in sensitivity due to sample dilution). Afterwards standard arthroscopy is performed to rule out possible joint damage, and during this process 5 additional tissue

We define infection using the same combined culture and histology algorithm as in biopsy samples taken during revision surgery. Our experience is that pre-operative biopsies are

administration at the time of revision surgery (Fink et al., 2011).

negative but the risk assessment suggests indicates septic implant failure.

more sensitive than culture from aspirated synovial fluid (Fink et al., 2008).

an individual regimen of systemic and localized antibiotic treatment for two-stage approaches using cement spacers supplemented with antibiotics (Fink et al., 2011). However, the utility of pre-operative biopsies is controversial between studies, mainly due to differences regarding the number of biopsies obtained and the definitions of infection (Fink et al., 2008, 2009; Meermans & Haddad, 2010).
