Culture-negative periprosthetic joint infection: is there a diagnostic role for next-generation sequencing?

KEYWORDS:

• Culture-negative
• Periprosthetic joint infection
• Next-generation sequencing
• Metagenomic sequencing
• Infection
• Joint arthroplasty

1. Introduction

Periprosthetic joint infection (PJI) following joint arthroplasty is a devastating complication. Despite recent advances in prevention, infection remains a real challenge and an important mode of failure following total joint arthroplasty (TJA). PJI is the most common reason for revision of failed total knee arthroplasty (16.8% to 25.2%) and the third leading cause for revision total hip arthroplasty (14.8%) [1,2]. In addition to the healthcare cost implications, which are four times higher than for a typical primary TJA, there is considerable morbidity and mortality associated with PJI [3,4]. Multiple surgical interventions, prolonged hospitalization, and higher complication rates are typically encountered when managing patients with PJI [5–8]. Furthermore, a five-fold increase in mortality has been reported for patients treated for PJI in comparison to aseptic revision counterparts [3]. While prevention is the most effective strategy, reaching an early and accurate diagnosis of PJI remains crucial for the delivery of appropriate care and successful treatment [9]. In an effort to reach a diagnosis, a combination of serological, synovial, microbiological, histological and radiological investigations is currently performed [10–12].

However, our current understanding is based on culture as the standard for pathogen identification and diagnosis. Yet, culture yields negative results in 7–50% of PJI cases [13–16]. Management of culture-negative PJI patients poses an issue regarding the selection of antimicrobial therapy [13]. Subjecting PJI patients to multiple surgical procedures and prolonged antimicrobial treatment without knowing the infective microorganism(s) is unacceptable in the 21st century. The outcome of the management of these patients is suboptimal and studies suggest a 4.5 times greater rate of reoperation in the setting of culture-negative PJI [17].

Even with strategies to optimize yield, such as obtaining more samples, increasing culture incubation period, and implant sonication, the sensitivity of culture remains at 39–70% [13,18–20]. Although numerous factors, such as premature use of antibiotics and infection with fastidious organisms, have been implicated, the reason for culture-negative PJI mostly relates to the inability to isolate organisms that reside in a biofilm and are not planktonic. In addition, isolation of slow-growing organisms such as coagulase-negative Staphylococcus, the second most common organisms resulting in PJI, using traditional culture can be difficult also [21,22]. Furthermore, culture methods for identifying uncommon causes of PJI, such as mycobacteria and fungi (except Candida spp), can take a relatively long time and require specialized, and organism-specific, culture media. Culturing, gram staining, and biochemical tests are traditional assays that currently consume the manpower of clinical microbiology laboratories [23].

Consequently, culture-independent molecular technologies have garnered interest in pathogen identification. With the advent of massively parallel DNA sequencing technologies, known collectively as next-generation sequencing (NGS), it has become possible to rapidly and comprehensively sequence all of the microbial genetic material within a clinical sample (metagenomics) [24]. Clinical application of NGS has enabled detection of an otherwise elusive pathogen in two seminal case reports; Abiotrophia defectiva in culture-negative endocarditis and neuroleptospirosis in a child with culture-negative meningoencephalitis [25,26]. A recent statement report from the American Academy of Microbiology (AAM) and Society for Microbiology (ASM) [23,24] described NGS as having the potential to dramatically revolutionize the clinical microbiology laboratory by ‘replacing current time-consuming and labor-intensive techniques with a single, all-inclusive diagnostic test’. Theoretically, NGS offers a more sensitive, culture-independent method for detecting organisms and can even discern whether organisms are transcriptionally active (metatranscriptome), enabling antibiotic resistance projections [24,27].

Prior molecular methods for diagnosing PJI, such as PCR, have shown promise in over the last decade, but also certain limitations that have precluded widespread clinical translation for PJI [28,29]. Jahoda et al. reported that PCR could detect as low as 590 CFU of Staphylococcus aureus, thereby facilitating pathogen identification in PJI even in the setting of prior antibiotic use [30]. Achermann et al. showed multiplex PCR of sonication fluid had better sensitivity than sonicate fluid culture (78% vs. 62%) [31]. PCR has also shown benefit in detecting genes responsible for biofilm formation and antibiotic resistance [32,33]. However, the sensitivity of PCR has been questioned in recent years, with multiple studies reporting similar or lower sensitivity for detecting bacteria in PJI than traditional cultures [34–36]. In a recent meta-analysis of 20 studies examining PCR utility for PJI, the pooled sensitivity, specificity, and diagnostic odds ratio were lower than prior meta-analysis estimates, at 0.76, 0.94, and 0.94, respectively [37]. Furthermore, PCR permits primer-restricted identification of an organism in a sample, or the identification of an organism only if it makes up a significant proportion of the DNA within the sample [38–40].

Theoretically, NGS does not suffer from some of these limitations. Perceived advantages of NGS include the speed of processing – from just a few hours to sequence the whole metagenome – and its ‘open read’ nature via comparison of sequence reads against comprehensive curated library databases available that contain all known pathogens [23,41]. That said, several questions regarding the clinical application of NGS remain, and this review aims to present a balanced overview of the available literature.

2. Pertinent studies

Next-generation sequencing (NGS) has already shown promise in the context of PJI for organism detection. Tarabichi et al. first demonstrated the utility of 16S-amplicon targeted NGS with the detection of Streptococcus canis in a culture-negative PJI patient [42]. In a prospective study of 78 patients, NGS was a useful adjunct for organism detection in 81.8% of culture-negative PJI where intra-operative tissue samples were analyzed [16]. Furthermore, in a series of 86 synovial fluid samples, high concordance with microbiological culture was seen with NGS of synovial fluid alone [43]. However, these studies were partly limited by sample size and future work is needed to determine the clinical significance of aseptic revisions that tested positive for microbial DNA (25%; n = 9/36 aseptic revision TJA) on NGS [16,43].

Using metagenomic shotgun sequencing, Thoendel et al. detected a wide range of PJI pathogens and suggested this method may aid in identifying the infecting organism in CN-PJI [44]. Using metagenomic sequencing, they were able to identify known pathogens in 94.8% of culture-positive PJIs, and new potential pathogens were detected in 43.9% (43/98) of culture-negative PJIs. Street et al. sequenced sonicate fluid from prosthetic joint and other orthopedic device infections, with 88% species-level sensitivity versus sonicate fluid culture [45]. While this form of NGS offers the potential for functional characterization of the organisms sequenced, the authors cited high levels of host DNA contamination despite best efforts in their laboratory address this.

In addition, emerging NGS data from three separate institutions suggest that PJI is polymicrobial at the microbial DNA level in a significant proportion of sequenced PJI cases [16,44,45], which contrasts with an established understanding of PJI based on culture modalities. Data presented at the 2019 American Academy of Orthopedic Surgeons meeting further suggests that the majority of PJI patients who later failed during longitudinal follow-up by a new organism (detected on culture) had the infective organism isolated by NGS during the initial treatment resection arthroplasty (88.7% of failures) [46]. This signal appears to support the clinical relevance of the NGS microbial DNA signal detected in the setting of PJI.

3. Perspective

While seems likely that NGS may occupy an increased role for the diagnosis of infection in the future, the translation of sequencing data into a clinically useful instrument in the setting of PJI remains critically limited. There is a lack of understanding of the relevance of this DNA signal. Preliminary data suggests that analytical microbial DNA can be detected, but we still face the difficulty of identifying a potential pathogen or combination of pathogens, but not knowing whether it is active and clinically relevant or not. Furthermore, with such high sensitivity, we need to understand the clinical importance of identifying unusual pathogens that may not necessarily be involved in the disease process.

In addition, there is concern regarding possible contamination and what may be natively present in the asymptomatic joint as part of a microbiome. We need to carefully consider the role that multiple pathogens may play and the way that they interact. NGS may implicate polymicrobial processes much more often than current techniques. There is also the issue of added cost and limited access as NGS is currently available in very few institutions. Highly specialized equipment, trained technicians, and bioinformatics expertise are also a prerequisite for effective sequencing runs [24]. Infectious disease specialists also have apprehensions about the broad-spectrum treatment of the polymicrobial metagenomic signal often detected within the joint, given antimicrobial stewardship implications. Without exploring clinical the relevance of this signal in PJI, it will remain difficult to understand the significance of the additional organisms identified on genomic sequencing.

4. Conclusion

The nature of the periprosthetic joint infection is complex. Infective organisms exist in the form of biofilm, take refuge inside osteoblasts and bony canaliculi. Thus, complete reliance on microbiological culture to diagnose these infections is suboptimal in certain cases of PJI. Next-generation sequencing is showing promise for pathogen identification in the setting of PJI. However, further multicenter work and randomized trials examining patient treatment outcomes will be necessary to translate this experimental technology into the clinical arena, validate the diagnostic benefits of NGS analysis, overall healthcare cost, and antimicrobial stewardship implications for management of PJI.

That said, it appears inevitable from the growing non-orthopedic literature that molecular sequencing technologies are here to stay, and we must embrace the paradigm shift in our understanding of periprosthetic infection that it may bring. In the words of the great American innovator R. Buckminster Fuller: ‘If you want to teach a new way of thinking, don’t bother trying to teach people. Instead, give them a tool, the use of which will lead to new ways of thinking’.

Declaration of interest

K Goswami has no conflicts of interest to disclose. J Parvizi reports grants, royalties, stocks, and/or consultancy with the following entities: Alphaeon, Ceramtec, Ceribell, ConvaTec, Corentec, Cross Current Business, Datatrace, Elsevier, Ethicon, Heron, Hip Innovation Technology, Intellijoint, Invisible Sentinel, JayPee, Joint Purification Systems, MDValuate, MicrogenDx, Parvizi Surgical Innovations, Physician Recommended Nutriceuticals, PRN-Veterinary, SLACK, Stryker, Tenor, Tissue Gene, Wolters Kluwer, and Zimmer, outside of the submitted work. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewers Disclosure

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was not funded.