The role of molecular diagnostics in implant-associated bone and joint infection

Authors

  • P.-Y. Lévy,

    1.  Université d’Aix-Marseille, Unité des rickettsies, URMITE CNRS-IRD, Faculté de médecine
    2.  Pôle de Maladies Infectieuses, Marseille, France
    Search for more papers by this author
  • F. Fenollar

    1.  Université d’Aix-Marseille, Unité des rickettsies, URMITE CNRS-IRD, Faculté de médecine
    2.  Pôle de Maladies Infectieuses, Marseille, France
    Search for more papers by this author

Corresponding author: F. Fenollar, Université d’Aix-Marseille, Unité des rickettsies, URMITE CNRS-IRD, Faculté de médecine, 27 Boulevard Jean Moulin, 13385 Marseille Cedex 5, France
E-mail: florence.fenollar@univ-amu.fr

Abstract

Clin Microbiol Infect 2012; 18: 1168–1175

Abstract

Microbiological culture is the conventional method for establishing the diagnosis in implant-associated bone and joint infection, but it may lack both specificity and sensitivity. Molecular diagnosis has been an important step in the diagnosis of infectious diseases. We review the principles and the role of molecular diagnosis in improving the aetiological diagnosis of implant-associated bone and joint infection. Currently, molecular diagnosis mainly includes conventional broad-range PCR and specific PCR assays. These tools are efficient, but several pitfalls exist that necessitate rigour in all steps of the process. In implant-associated bone and joint infection, molecular assays have been shown to be useful in complementing culture techniques to identify microorganisms when patients have previously received antibiotics or in the presence of fastidious microorganisms. Broad-range PCR targeting the 16S rRNA sequence followed by sequencing must be performed in culture-negative specimens when infection is suspected on the basis of clinical signs and symptoms or inflammatory syndrome. This molecular tool has allowed not only increasing identification of anaerobic bacteria, such as Finegoldia magna, but also the discovery of the role of Tropheryma whipplei, an aetiological agent of implant-associated bone and joint infection in patients without Whipple’s disease. Real-time pathogen-specific PCR assays performed in a closed system are more sensitive and specific than broad-range PCR, but each assay is typically able to detect only a single microorganism. These assays should be performed to confirm the identification provided by broad-spectrum PCR, and also when broad-range PCR fails to detect a microorganism despite efficient DNA extraction.

Introduction

The diagnosis and treatment of implant-associated bone and joint infection is a major challenge [1]. These infections are mainly caused by microorganisms that grow in biofilms on the surface of the implant [1]. In the biofilm, microorganisms are difficult to diagnose and treat effectively. Indeed, they are recalcitrant to conventional antimicrobial agents and host immune responses [1]. The failure of aetiological diagnosis with traditional culture may be caused by prior antimicrobial exposure and fastidious microorganisms, but it is also possibly related to the reduced growth rate of biofilm microorganisms delaying or inhibiting growth in microbiological culture [1,2].

In culture-negative cases, treatment is performed on an empirical basis, with poor prognosis [1,3]. Molecular tools may help to establish an aetiological diagnosis, but molecular procedures must be rigorously performed because of the risks of interfering contamination [4]. The efficiency of PCR assays in detecting microorganisms depends on sample collection, PCR methodology, validation, and the interpretation of each PCR analysis [5]. Basic rules must be applied to obtain reliable results. In this article, we review the entire process of performing and interpreting molecular amplification assays, and their role in the diagnosis of implant-associated bone and joint infections.

Clinical Samples for PCR Assays

Collection

PCR assays should be performed on fluids collected by needle aspiration or on periprosthetic tissues obtained by surgical biopsy [5]. At least three specimens should be sampled for culture [3,6–8]. They should be shipped immediately to the laboratory after collection [5]. Prior to culture, a piece of each specimen should be stored at −20°C or −80°C for further analyses [5,9]. If fresh specimens are not available, PCR could be performed on paraffin-embedded biopsy specimens, but this alternative lacks both sensitivity and specificity [10]. Superficial wounds should be avoided, because of the frequent colonization by microbial flora from the surrounding skin [11]. Samples collected by swab should be avoided, as they may lack sensitivity [3]. The application to the specimens of an appropriate sonication technique with an ultrasound bath has recently shown encouraging results [12].

DNA extraction

Prior to molecular amplification, DNA extraction is an important step. Various protocols have been reported [5,9,13–16]. To optimize the quality of the DNA extracts, an overnight step in lysis buffer and proteinase K before extraction is strongly advised for bone and joint specimens, including fluids. Moreover, the use of automation, which allows the elimination of several manual steps and the risk of cross-contamination, should be preferred.

Interpretation of PCR Assays

PCR pitfalls and solutions

There is sometimes a lack of efficient DNA extraction, leading to false-negative results. The main critical points are the technique used and the cell wall of the bacteria [17]. The quality of DNA extracts must be evaluated by amplifying a fragment of a human gene, such as the β-globin or actin gene [5,9]. In cases of negativity, another specimen must be extracted at another time. The amount of background DNA from bacteria and the size and abundance of the amplified fragment may also play a role [17,18]. When possible, a short target (c. 100 bp) should be designed [18].

Molecular assays are also prone to contamination, which can occur from the time when the samples are obtained to their manipulation during the laboratory process [4,19]. Contamination may occur from one sample to another during the assay, or from laboratory surfaces, tubes, pipettes, the technician’s hair or clothes, or previous amplicons in the laboratory. Even if those risks can be lowered by following strict rules, single-use plasticware, tubes, water and reagents can still be a source of contamination [4,20,21]. Thus, it is imperative to carefully follow standard recommendations to prevent contamination (Fig. 1) [5,21,22].

Figure 1.

 The most frequent contamination pitfalls and the solutions for their prevention and detection.

Validation

The presence of positive and negative controls in each PCR assay is necessary to validate the assay. The use of positive controls (extracted bacterial DNA) allows confirmation that the PCR process was correct. DNA from a microorganism that does not commonly cause implant-associated bone and joint infection, such as Escherichia coli, is preferred to DNA from a common causative pathogen. The use of negative controls, processed from DNA extraction to PCR in parallel with the tested samples, is imperative to detect contamination. Water, a mixture of all reagents used in the PCR assay and DNA extracted from human tissue without infection can be used as negative controls [19].

Interpretation

Both negative and positive controls must be correct for interpretation of the PCR. If the positive control was not detected or if any amplification of the negative control occurred, the run is unreliable [23] and should be repeated. For broad-range PCR, each positive amplicon must be sequenced for accurate identification of the causative microorganism. An original sequence observed for the first time in a laboratory can usually be considered to be a true positive [10]. A similar sequence found in the same run from samples of two patients suggests potential contamination. Sequences from microorganisms that are commonly present in water or reagents (Pseudomonas spp. and Acinetobacter spp.) and those from microorganisms from skin (coagulase-negative staphylococci and Propionibacterium acnes) may result from contamination. The interpretation of the results is sometimes difficult, as coagulase-negative staphylococci are the most commonly observed pathogens in implant-associated infections. Thus, the microorganisms of the normal skin flora may be either contaminants or pathogens. If a result is doubtful or has a low predictive value, supportive data can help in the interpretation. PCR targeting a second gene can be performed with the same DNA extract, and another sample of the patient can be extracted and tested at another time.

Finally, the persistence of DNA from the causative agent has been reported in infective endocarditis from months to years after clinical cure. It is possible that, as with endocarditis, the link between the current episode and the amplified DNA needs to be evaluated for implant-associated bone and joint infections [24,25]. However, to the best of our knowledge, there are currently no published data on this potential diagnostic problem.

Usefulness of PCR Assays

The main indication for PCR assays is the lack of microorganism culture. No microorganisms are detected in approximately 5–15% of apparent infections [1]. However, suspicion of mixed infections, mainly in the case of open lesions, can be also an indication.

Fastidious microorganisms

Molecular diagnosis has allowed spectacular data to be obtained for the identification of Kingella kingae among children with osteoarticular infections [13,15,26]. Currently, for patients with implant-associated bone and joint infections, PCR assays have mainly allowed the identification of Granulicatella spp., Staphylococcus aureus, even in cases of small-colony variant infection, and anaerobic bacteria [1,27]. Indeed, the implication of Finegoldia magna as a pathogen in prosthetic joint infection has been rediscovered by broad-spectrum PCR [28,29]. PCR has also allowed the establishment of the role of Tropheryma whipplei as an aetiological agent of prosthetic joint infection without other classic signs of Whipple’s disease [30–32]. More recently, Legionella micadadei has been reported to cause prosthetic joint infection, thanks to the systematic use of 16S rRNA PCR [33]. PCR could also help in the diagnosis of other fastidious microorganisms involved in implant-associated bone and joint infections, such as Mycobacterium tuberculosis or other Mycobacterium spp. [34–36].Thus, broad-spectrum PCR is useful in the diagnosis of difficult-to-culture organisms, especially when the pathogen is rare or unsuspected.

Previous antibiotic therapy

Antimicrobial therapy must be discontinued at least 2 weeks before the sampling of specimens [1]. In cases of revision surgery, perioperative prophylaxis must not be administered until after tissue specimens have been collected. However, in cases of a longer duration of antibiotic therapy or inadequate prophylaxis, broad-range PCR may help to establish an aetiological diagnosis [5,22,37].

Mixed infection

The potential inability of broad-range PCR to identify a mixture of bacterial species in a single specimen has been considered [38]. However, 16S rRNA PCR combined with cloning is able to detect mixed polymicrobial infections, allowing for identification of the main bacteria present in the mixture [5]. Moreover, with this approach, new potential microorganisms have been reported among patients with mixed infections, whereas no novel bacterial species were detected in monobacterial specimens [5]. These unusual microorganisms are not detected in culture, mainly because of overgrowth by rapidly growing species.

Conventional Broad-range PCR

Methodology

The 16S rRNA gene, the 23S rRNA gene, the rpoB gene and the 16S–23S intergenic spacer are present in all bacteria, and are thus particularly suitable for broad-range PCR [19]. Several protocols have been reported for 16S rRNA gene-based detection of bacteria [16,19,39–41]. However, only 16S rDNA amplification, systematically followed, when positive, by sequencing [5], allows accurate identification of the involved microorganisms, and must be used (Fig. 2). Moreover, nested or semi-nested PCRs that correspond to the reamplification of a PCR product must be avoided, owing to the high risk of contamination [32,42]. A polymicrobial infection must be suspected if a mixed sequence is observed on the DNA pherogram, and a cloning step with commercialized kits should be performed [10]. Sequencing of all of the clones is time-consuming, but it may help to identify new pathogens. Indeed, a sequence similarity of <97% of the 16S rRNA sequence is the criterion used to define a potentially new bacterial species [43–45]. The main primers that have been published for the diagnosis of implant-associated bone and joint infections are summarized in Table 1 [5,9].

Figure 2.

 Methodology of conventional broad-range PCR and specific PCR assays.

Table 1. Main primers that have been used and published for molecular amplification in implant-associated bone and joint infections (extract from Table 1 of Fenollar F, Levy PY, Raoult D. Usefulness of broad-range PCR for the diagnosis of osteoarticular infections. Curr Opin Rheumatol 2008; 20: 463–470)
ReferencesMicroorganismsTargeted sequencesForward primer: name and sequence (5′–3′)Reverse primer: name and sequence (5′–3′)
  1. Y = T or C; R = G or A; K = G or T.

Fenollar et al. [5]Eubacteria16S rRNA536F, CAGCAGCCGCGGTAATACrp2, ACGGCTACCTTGTTACGACTT
Fihman et al. [9]Eubacteria16S rRNA536f, CAGCAGCCGCGGTAATAC1050r, CACGAGCTGACGACA
Eubacteria16S rRNA16S-241bp-F, GGAGGAAGGTGGGGATGACG16S-241bp-R, ATGGTGTGACGGGCGGTGTG
Eubacteria16S rRNA16S-cons-F, YGGCGRACGGGTGAGTAA16S-GP-R, CCGATCACCCTCTCAGGTCG for GP
and
16S-GN-R, AGTTAGCCGGTGCTTCTTCT for GN
Fenollar et al. [5] Staphylococcus aureus RpoB Saur.Rpob.F, GTTTGAATTGCATGGTAGCGTSaurRpob.R, GAAGCAATGATATCTGCTGGTG
Streptococcus group B RpoB SagaRpob.F, CAATTGCAGAGCATATCGATGGSagaRpob.R, TCCAACAATAGTAACAACACGG
Streptococcus pneumoniae RpoB SpncuRpob.F, GGTAGAAGCTGGTACGATTATGACSpneuRpob.R, GATCAGTTGGAGCAACAACCT
Enterococcus spp. RpoB StrpF, AARYTIGGMCCTGAAGAAATStrpR, TGIARTTTRTCATCAACCATGTG
Granulicatella adiacens RpoB GadiaRpob.F, TGTAACTCTAACACTTGTCCGAGadiaRpob.R, GGACGTCACGGTAATAAAGGG
Staphylococcus epidermidis RpoB SepiRpob.F, GTGATACGTCCATGTAATCCASepiRpob.R, TTTGACAGCTGATGAAGAGGA
Enterobacteriaceae RpoB CM7, AACCAGTTCCGCGTTGGCCTGGCM31b, CCTGAACAACACGCTCGGA
Escherichia coli RpoB EcolRpob.F 5′-TTCACCAACGATCTGGATCACEcolRpob.R 5′-GAAGAACAGGTTCTCGAACAG
Pseudomonas aeruginosa RpoB PaerRpob.F, TGTACACCAACGACATCGACPaerRpob.R, AAGAACAGGTTGCCGAACAG
Clostridium perfringens EabccCloPerF, GACACATTAGTTGGGAAGTGCloPerR, GTTGTTAACCAATGAGCAGC
Propionibacterium acnes 16S–23S rRNAPacnF, CTAAGGAGTTTTTGTGAGTGGPacnR, CTTTGCACAACACCACGTC
Prevotella spp.CfxA2-like lactamasePrevoF, GGATAAACTTGACCCAAAGACPrevoR, GATGTATAGTTAGAGTAAGCC

Limits of conventional broad-range PCR

Although most studies agree about the good specificity of broad-range PCR in the diagnosis of implant-associated bone and joint infections, discrepancies have been reported about its sensitivity, with several studies showing poor sensitivity (<50%) [5,9,40,46–48]. Several biases could have influenced these data, such as the lack of pretreatment prior to DNA extraction, the lack of control of the extracted DNA quality, or the small number of studied patients [49]. Among the three studies that have involved more than 50 patients, the sensitivity of 16S rRNA PCR for osteoarticular infections (both prosthetic and not) was determined to be 92.5% vs. 89% for culture [5] in the first study. The second study reported a PCR sensitivity of 73.3% (53.8% for prosthetic joint infections and 88.2% for infections without prostheses), whereas for cultures, the sensitivity was 96.7% [9]. In the third study on prosthetic joint infections, culture and PCR had similar sensitivities (72.6% and 70.4%) and specificities (98.3% and 97.8%) [22].

Specific PCR assays

For pathogen-specific PCR, the assays should be performed with quantitative real-time PCR in a closed system, in which not only the amplification but also the identification of amplified products with DNA probes with specific annealing within the target-amplified region are coupled in a single vessel, reducing the risk of contamination (Fig. 2) [10]. If fluorescence-labelled oligonucleotide hybridization probes are not used, the identification must be confirmed by sequencing when an amplified product is detected. Overall, this technique has other advantages over conventional PCR, including speed, simplicity, reproducibility, and quantitative capacity. Pathogen-specific PCR has previously been shown to be more sensitive than 16S rDNA PCR in osteoarticular infections for S. aureus and M. tuberculosis [5,50]. Moreover, all of the currently available sequenced bacterial genomes (2848) allowed for the best DNA targets to be chosen for specific assays, with greater sensitivity than that of broad-range PCR [20]. Thus, pathogen-specific PCR is useful not only when a more sensitive diagnosis is needed, but also to confirm the results obtained with broad-range PCR [5].

Strategy for the use of PCR Assays

PCR assays offer several advantages in the diagnosis of implant-associated bone and joint infection, but its use should be restricted, as proposed in Fig. 3. Indeed, PCR assays should be limited to culture-negative cases when infection is suspected on the basis of clinical signs and symptoms or inflammatory syndrome is present, as highlighted by blood test results and purulent samples [1]. Thus, broad-range 16S rDNA PCR must be performed when the culture was not positive. In suspected polymicrobial infections (open lesion, two or more bacterial species isolated with culture, and mixed sequence on the DNA pherogram), a 16S rDNA PCR assay followed by cloning procedures may also be performed. Specific PCR assays should be performed to confirm the results of broad-range PCR, and also when broad-range PCR fails to detect a microorganism despite efficient DNA extraction. The main epidemiological and clinical criteria for guiding the choice of specific PCR assays to be performed in a given patient are summarized in Table 2 [1,51–54]. S. aureus-specific PCR assays could also be performed in parallel with broad-range PCR, because this bacterium is one of the most common microorganisms associated with prosthetic joint infections [3].

Figure 3.

 Suggested strategy for the management of specimens in the microbiological laboratory and the use of molecular diagnostics in implant-associated bone and joint infection.

Table 2. Main relevant epidemiological and clinical information for guiding the choice of specific PCR assays to be performed in a given patient with implant-associated bone and joint infection
Epidemiological and clinical criteriaSpecific PCR assays
  1. aThe usual manifestations of delayed infections (3–24 months after surgery) are subtle, with persistent joint pain and/or implant loosening.

  2. bThe usual manifestations of early infections (<3 months after surgery) are acute, with joint pain and fever as well as effusion, redness and warmth at the surgical site.

A stay in an endemic area and ingestion of unpasteurized dairy products [54] Brucella melitensis
Children aged <5 years [13] Kingella kingae
Rheumatoid arthritis, chronic steroid use, treatment with monoclonal antibody against tumour necrosis factor-α, history of Mycobacterium tuberculosis infection involving another system, and pulmonary diseases [52] M. tuberculosis
Delayed infectiona, especially in patients with shoulder prostheses [1,53] Propionibacterium acnes
Rheumatoid arthritis, early and acute haematogenous infectionb [1,51] Staphylococcus aureus
Delayed infectionb [1] Staphylococcus epidermidis

Antibiotic Therapy Prescription Based on PCR Results

Among the drawbacks of molecular diagnosis is the lack of antibiotic susceptibility results (except for specific resistance genes, such as the genes conferring resistance to methicillin, quinolones, and rifampicin) [12,49,55,56]. However, when rigorously used, molecular tools provide robust diagnoses and provide indications for the prescription of the best empirical antibiotic treatment for the identified microorganisms, or specific treatment in cases when M. tuberculosis, Brucella spp. or T. whipplei infections are diagnosed [1,30]. If physicians are faced with a doubtful result and are hesitant to initiate treatment, they must contact their laboratory to discuss the result and to check the need to confirm its accuracy.

Conclusion and Perspectives

Conventional broad-range PCR has been an important advance in the diagnosis of infectious diseases. For implant-associated bone and joint infections, broad-range PCR followed by sequencing should not be performed in routine cases, but to complement culture, mainly for culture-negative cases, when infection is suspected. In addition to broad-range PCR, pathogen-specific real-time PCR is also useful. The technical procedures of PCR assays must be strictly performed, and the interpretation of the results must be rigorous. Even if the antimicrobial susceptibility is not available, the PCR result orients physicians towards the best empirical strategy. Molecular assays represent an additional tool with which to improve the diagnosis of implant-associated bone and joint infections on the basis of interdisciplinary teamwork. Indeed, physicians and microbiologists must interact, mainly for low predictive value results.

In the future, with the availability of a large number of complete bacterial genomes, specific primer sets and probes targeting a larger spectrum of bacteria could be developed, allowing for accurate and fast aetiological diagnosis. Multiplex PCR assays may also improve diagnosis [12]. The potential role of fluorescence in situ hybridization, a molecular technique that uses fluorescently labelled probes to detect RNA or DNA, and that is already used in infective endocarditis, remains to be evaluated for implant-associated bone and joint infections [17]. Currently, molecular diagnosis cannot be performed routinely in every laboratory. However, the rapid evolution and improvement of the technology could allow procedure standardization with automated workstations to perform sample processing, analysis, and product detection, to open these techniques to wider application in many laboratories. Molecular diagnosis is a more expensive diagnostic method than bacterial culture. However, the cost-effectiveness of this strategy has not yet been evaluated.

Transparency Declaration

The authors declare no conflicts of interest.

Ancillary