Liver transplantation (LT) is the only curative treatment for liver cirrhosis, but donor organs are scarce and therefore allocated at a late stage of the disease (Tacke et al., 2016; Adam et al., 2018). Patients on the waiting list for LT are prone to acute decompensation (AD), acute-on-chronic liver failure (ACLF), and hospitalization including application of broad-spectrum antimicrobials due to infections which are closely connected to AD and constitute the most frequent etiology of ACLF, meaning that cirrhotic patients frequently undergo antimicrobial therapy (Fernández et al., 2017; Trebicka et al., 2020). The prevalence of multidrug-resistant organisms (MDRO) in these patients is still rapidly increasing worldwide (Fernández et al., 2012, 2019; Salerno et al., 2017; Piano et al., 2019). Therapeutic options are further limited in case of infection with carbapenem-resistant Gram-negative bacteria (CRGN), often leading to clinical deterioration and death (Ferstl et al., 2017, 2018; Fernández et al., 2019). To prevent transmissions, implementation of infection control measures and an active surveillance of CRGN strains in transplantation centers are crucial (Geladari et al., 2017). Since patients evaluated for LT are regularly screened for CRGN at our center, we reasoned that this vulnerable population might be well-suited for the study of mechanisms leading to carbapenem resistance (CR).

CRGN carriage is detected by screening algorithms (e.g., from rectal swabs) using standard cultivation techniques, species identification, and antimicrobial susceptibility testing methods (Tomczyk et al., 2019). However, the exact genetic resistance determinants and transmission routes often remain unknown. CRGN can be distinguished in carbapenemase-producing (CP) and non-carbapenemase-producing (non-CP) bacteria. The first group harbors carbapenem-hydrolyzing β-lactamases. These isolates represent the most important hygiene threat (Nordmann et al., 2012; Sahuquillo-Arce, 2015) and account for approximately 10% of CRGN cases in patients with liver disease (Ferstl et al., 2017). The group of non-CP CRGN consists of a broad variety of bacterial isolates without expressing a defined carbapenemase. CR in this group is often the result of overexpression of β-lactamases, such as AmpC cephalosporinases or extended spectrum β-lactamases (ESBL) in conjunction with increased efflux or impermeability, for example, due to porin loss (Köhler et al., 1999; Rodríguez-Martínez et al., 2009; Sahuquillo-Arce, 2015; Goodman et al., 2016; Hamzaoui et al., 2018). Various possible explanations for CRGN occurrence exist: (i) CR might be an evolutionary result of mutations leading to loss of porins and/or overexpression of resistance genes in commensal bacteria; this process is triggered by the administration of broad-spectrum antimicrobials and subsequent antimicrobial selection pressure (Sommer et al., 2009; Meletis et al., 2014; Goodman et al., 2016; Iredell et al., 2016). (ii) CR can be acquired by so-called “plasmid hospitalism” which might be the result of a horizontal interspecies gene transfer (Göttig et al., 2015) and, (iii), maybe most important, CRGN can be transmitted directly from patient to patient via, for example, shortcomings in infection control measures. Here, we report on the WGS analysis of 18 CRGN isolates obtained from 24 detected CRGN isolates (from 23 patients) while screening of 351 patients evaluated for liver transplantation (Ferstl et al., 2021) over a period of 10years.

Colonization with CRGN is associated with repeated hospitalization, sepsis, and increased mortality in patients with advanced liver disease (Ferstl et al., 2017, 2021). Therefore, it is of utmost importance to avoid transmission of CRGN to patients that are awaiting or underwent LT.

The results of our retrospective study provide no evidence for a gross transmission of CRGN bacteria or CR genes/plasmids mediating resistance determinants from patient to patient for the herein analyzed patient cohort. These conclusions can be drawn out of the observations that (i) most bacterial isolates were identified as phylogenetically unrelated, (ii) carbapenemases (OXA-232, VIM-1, VIM-2, and OXA-72) were each found in not more than one isolate, respectively, and (iii) there was no evidence for plasmid-mediated spread of resistance determinants mediating CR. These observations contrast with studies from other centers where a certain sequence type and or carbapenemase type were dominantly witnessed (Clancy et al., 2013; Karampatakis et al., 2017; Macesic et al., 2018). Dissemination of certain sequence types carrying carbapenemases is known to vary drastically depending on region (Nordmann et al., 2011; Logan and Weinstein, 2017). Therefore, our data indicate that an individual assessment and surveillance of CRGN might be needed for each center.

In our cohort, non-CP CRGN isolates were the predominant resistance determinants accounting for 14 out of 18 isolates. For non-CP CRGN, CR is often mediated by ESBL and/or AmpC genes in conjunction with porin deficiencies (Köhler et al., 1999; Rodríguez-Martínez et al., 2009; Hamzaoui et al., 2018). Porin-mediated resistance cannot be easily transferred between pathogens and might even come with a fitness cost for the respective isolate (Nordmann et al., 2012; Wong et al., 2019); however, the underlying mutations can arise over time and there is evidence that non-CP-mediated resistances are more likely to arise under antimicrobial pressure (Goodman et al., 2016). Obviously, this is especially troublesome as antimicrobials are indispensable in the treatment of patients with decompensated cirrhosis and ACLF.

Our results strongly suggest that three of the analyzed CRGN derive from a carbapenem-susceptible ancestor. Therefore, strict infection control measures for patients at risk even at those time points before the detection of CR bacteria are required. Such measures might avoid the spread of certain bacterial ancestor strains with a high genetic plasticity for evolving CR resistance which is of special importance as spread of such carbapenem-susceptible pathogens will mostly not be noticed. In case of the three isolate of our study, careful analysis of patient data records did not uncover any potential transmission event. For identification of such events, it is necessary to trace all contact points of patients to the healthcare system. For instance, transmission of CRGN via an outpatient clinic was reported (Klein et al., 2021) and it is highly likely that a part of LT patients is under a shared medical supervision. Patients might have crossed at healthcare facilities, such as general practitioners, primary or secondary hospitals, or even nursing and rehabilitation facilities. As relevant outpatient histories could not be retrieved due to the retrospective design and demise of almost all patients, the only possibility here was to reconstruct any patient-to-patient encounters inside of our hospital but defined crossing points were not detected.

Although the in-house hygiene plan with strict screening algorithms for MDRO allows quick and reliable CRGN detection in cirrhotic inpatients, acquisition of carbapenem-susceptible common ancestors might occur unregistered as screening for antimicrobial-susceptible bacteria is not done routinely. Tracking down such strains, however, requires a comprehensive study design including all participating healthcare providers and is also of high microbiological complexity as there is no easy screening algorithm for carbapenem-susceptible pathogens carrying an intrinsic susceptibility for development of carbapenem resistance.

In summary, rather than transmission of CRGN or CR determinants within our cohort, our data support a model according to which CRGN evolve under antimicrobial pressure. Transmission of carbapenem-susceptible ancestor pathogens may be an early step within this process (see Figure 6). This implicates that (i) the transmission of potential ancestors should be avoided by strict infection control measures even before occurrence of CR bacteria and (ii) the wise use of antimicrobials should be pursued especially in patients with carbapenem-susceptible multidrug-resistant Gram-negative bacteria. Antimicrobial stewardship efficiently reduces antimicrobial burden in, for example, oncologic patients (Pillinger et al., 2020), yet its value in patients with decompensated cirrhosis and ACLF still needs to be demonstrated. While hospital hygiene stays an integral part in the care of cirrhosis and LT, further strategies to avoid acquisition of CR clearly need to be implemented. To this end, prospective studies including close-mesh microbiological screening and a surveillance of pathogens using WGS are urgently warranted.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ncbi.nlm.nih.gov/sra/SRP327397/.

The studies involving human participants were reviewed and approved by Ethics Committee of the University Hospital Frankfurt am Main, Germany. The patients/participants provided their written informed consent to participate in this study.

VK, M-WW, TS, and PF designed the study. DV, MH, SG, and TW were involved in microbiological screening procedures, sample collection, and sample preparation. TS performed bioinformatic analysis. PF, WB, SZ, OW, and M-WW were responsible for clinical patient care and patient charts. VK, TS, and PF wrote the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the State of Hesse, Germany (“Hessisches Universitäres Kompetenzzentrum für Krankenhaushygiene” and the LOEWE-Center “ACLF-I, project P5”). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.