Commentary The human microbiome—the ecosystem of bacteria living and interacting on (and in) human tissues—is enormously complex. Our understanding of its impact is evolving. Next-generation sequencing (NGS) techniques allow for the identification of microbial species and can be used to elaborately interrogate the composition of the bacterial community in specific anatomic sites. This has been extensively described in relation to the gut microbiome, the largest concentration of bacteria in the human body, where complex ecological interactions alter immunology1, pathogenesis2 and metabolic functions3. Far less is known about the microbiomes of anatomic sites previously considered sterile, such as the placenta, the central nervous system, and now the synovium. No one disputes that bacteria and bacterial DNA may transiently penetrate these protected spaces, but whether bacteria truly live there remains to be demonstrated. This preliminary study by Goswami et al. represents a relevant attempt to identify the microbiome of the human osteoarthritic hip and knee. The investigators obtained synovial fluid, tissue, and swabs from patients undergoing primary total hip or knee arthroplasty for osteoarthritis—patients with no overt signs of infection. Concordance in bacterial community composition among samples from the same patient was observed. Variations in diversity were seen depending on the hospital of enrollment and on whether patients had received a corticosteroid injection; these differences may be proxies for broader differences in patient-specific microbial composition. Surprisingly, comorbidities and demographic variables had no significant impact on the microbes found within a joint in this study. The detected microbes ranged from common to obscure—including some that are traditionally considered pathogenic and many others that are not. Some of the common sequences obtained included those of Staphylococcus, Streptococcus, and Pseudomonas. However, even though these are recognized orthopaedic pathogens, it would be surprising to confirm that they persist in clinically uninfected joints. Other bacteria found with frequency included some that are typically found in the environment but only occasionally cause opportunistic infections, such as Paenarthrobacter and Brevundimonas. The limitations of NGS techniques are important to consider. The presence of nucleic acids does not mean the presence of living microbes: remnants of dead bacteria, contamination introduced during sample collection and processing, and even miscategorized mammalian DNA in the vast NGS sequence libraries can all lead to positive findings even when no living bacteria are present. NGS findings are also difficult to confirm by alternate means (such as microscopy or culture), as the microbial burden is so low. A healthy degree of skepticism is warranted, given recent studies of low-biomass microbiomes in other tissues not known to harbor bacteria, such as the human fetus4,5. To their credit, the authors acknowledge these important limitations of their study. So, what can we learn from this study? First, the study serves as a reminder that the presence of microbial DNA in joints, in the absence of other findings of infection, is not diagnostic of infection. This is an important point particularly as it pertains to the molecular diagnosis of periprosthetic joint infection, which is gaining faster in popularity than in evidence. Optimizing molecular tests to diagnose infections is complicated by the implication that all joints may have subclinical quantities of bacterial DNA. Second, it is plausible that microbial nucleic acids are present in osteoarthritic joints. If the findings of this study are confirmed, further work should be done to establish whether these products are being produced intra-articularly by living bacteria, whether they are unique to diseased joints, and whether they represent a disease state that might benefit from treatment. The burden of proof should be high. Further confirmation of these findings with optimal controls, using alternate methodologies, and with replication across other populations (for instance, comparing osteoarthritic and healthy joints) will be needed to convince us that a stable microbiome exists within the joint capsule. Lastly, the techniques used in this study are less broadly available and much more complex than other existing diagnostic tools. The expertise to critically evaluate this type of methodology is beyond the scope of many clinicians. Further education and elaboration on clinically relevant findings are critical to realize the full clinical benefit of NGS in orthopaedics. In summary, this provocative paper offers evidence that microbial DNA is present in the osteoarthritic joint. The next steps that are needed will be to confirm these findings with appropriate controls at every juncture, and to evaluate whether these nucleic acids derive from living intra-articular bacterial communities or from transient circulation of dead microbes.
Read full abstract