Clinical Metagenomics of Infective Endocarditis

Overview

Infective endocarditis (IE) is an infection of cardiac valves. IE mainly involves bacteria, more rarely fungi. IE is an uncommon diseases with an estimated incidence of 1-12 cases per 100,000 inhabitants per year. The diagnostic of IE relies on the culture of biological samples (blood cultures and per-operative samples) in the bacteriology laboratory in order to identify the pathogen and its susceptibility to antimicrobials. Nonetheless in about 10% of the cases, the blood cultures remain negative, due to antibiotics taken before harvesting, to non-culturable bacteria or to aseptic phenomena. Clinical metagenomics is defined as the application of high-throughput sequencing (NGS) followed by a specific bioinformatics analysis to obtain clinical information, i.e. pathogen identification and the prediction of their susceptibility to antimicrobials. The metagenome of a sample (i.e. all the genomes of the organisms present) virtually contains all the information necessary for bacteriological diagnosis: what is the pathogenic bacteria , and to which antibiotics it is susceptible. Hence, using clinical metagenomics in the context of IE appears seducing in order to overcome the limitations of conventional methods based on culture. Here, we propose to assess the performance of clinical metagenomics in the diagnostic of IE.

Background and Rationale Infective endocarditis (IE) is an infection of cardiac valves. IE mainly involves bacteria, more rarely fungi. IE is an uncommon diseases with an estimated incidence of 1-12 cases per 100,000 inhabitants per year. The diagnostic of IE relies on the collection of biological samples (blood cultures and per-operative samples) and their culture in the bacteriology laboratory, referred to as conventional methods. This process, practically unchanged since the time of Pasteur at the end of the 19th century, has the disadvantage of allowing the only detection of the bacteria that can be grown under the usual conditions of a laboratory. If most pathogenic bacteria can do so, some bacteria that may be involved in IE (eg Coxiella sp., Bartonella sp., Tropheryma whipplei) require very specific conditions to multiply. In addition, the prior intake of antibiotics by the patient before the sample collection may negatively influence the culture of the samples. The diagnosis of IE is evoked by considering a range of clinical arguments (fever, microbiological risk behavior (positive blood culture(s), serology (for Coxiella burnettii), echography (presence of an intra-cardiac mass, abscess, leakage and / or disinsertion of a valve), all of which are minor and major criteria for determining the certainty of IE. The treatment of IE is based on prolonged antibiotics (4 to 6 weeks for the majority of cases) and valvular surgery (according to precise indications). In about 10% of the cases, however, the blood cultures remain negative. These are known as negative culture IE, which may be due to antibiotics taken before sampling, to non-culturable bacteria or to aseptic phenomena. The broad-range PCR which consists of amplifying by PCR then sequencing a fraction of the 16S RNA encoding gene can then be used. It allows the identification of the pathogenic bacterium, but does not give information on any resistance acquired to antibiotics. Concept of clinical metagenomics Clinical metagenomics is defined as the application of high-throughput sequencing (NGS) followed by a specific bioinformatics analysis to obtain clinical information, i.e. pathogen identification and the prediction of their susceptibility to antimicrobials. The metagenome of a sample (i.e. all the genomes of the organisms present) virtually contains all the information necessary for bacteriological diagnosis: what is / are the pathogenic bacteria (s), and to which antibiotics it/they is/are susceptible. The concept of clinical metagenomics has developed in parallel with the new DNA sequencing technologies introduced in the mid-2000s and much more efficient in terms of throughput than the sequencing method described by Sanger. While this concept is attractive, there are still many obstacles to its implementation. First, clinical specimens from bacterial infections usually contain a high concentration of leukocytes, the genome of which is about 1000 times larger than that of bacteria. Thus, the first limiting step for clinical metagenomics is the need to obtain sufficient bacterial DNA to allow the preparation of quality libraries for sequencing, but also to reduce the concentration of human DNA whose sequencing is unnecessary in this context . Methods are available but their evaluation for clinical metagenomics purposes is necessary. Secondly, the complexity of the bioinformatics data to be managed by a microbiologist requires that data be exploited as automated as possible together with a user-friendly interface, which is not the case today even if online platforms are being developed. The taxonomic assignment of sequences, their assembly, the identification of genes and chromosomal mutations associated with antibiotic resistance, and the establishment of a link between the resistance determinant and the host bacterium are additional obstacles to the implementation of clinical metagenomics. Finally, if the time taken to obtain results that can be exploited by the microbiologist tends to decrease, it is now comparable to that of culture, at a much higher cost. However, sustained competition between sequencer manufacturers should maintain their decline as has been the case over the last decade. Relevance of the use of clinical metagenomics in IE The use of clinical metagenomics in the context of IE therefore seems relevant for several reasons: * Most IE are monomicrobial, which support the good performance of clinical metagenomics according to our preliminary results. * The culture of intraoperative samples performed in an IE context is sometimes negative, due to antibiotic pre-treatment and / or non-cultivation under the routine conditions of the pathogen (eg Bartonella spp or Coxiella spp.), leaving room for improvement. * The treatment of IE is a long-term treatment that requires accurate diagnosis in accordance with pathogens' susceptibility to antibiotics. If the pathogen is not found in culture, broad spectrum antibiotics should be administered to the patient with a double risk of treatment failure and toxicity. However, clinical metagenomics could provide information on antibiotic sensitivity even in the case of a negative culture. * In most cases, the microbiological diagnosis of IE is not an emergency diagnosis, which is compatible with the use of clinical metagenomics for which the time of implementation is at best 48-72h. Hence, we propose to assess the performance of clinical metagenomics in the diagnostic of IE.