Chronic kidney disease (CKD) is an increasing global health issue, defined as decreased kidney function shown by a glomerular filtration rate of less than 60 ml/min/1.73 m², or markers of kidney damage, or both, for a duration of at least 3 months (Webster et al., 2017; Benjamin and Lappin, 2021). CKD is commonly progressive and when patients reach end-stage renal disease, they may require renal replacement therapies, such as hemodialysis, peritoneal dialysis (PD) or kidney transplantation (Murdeshwar and Anjum, 2021). CKD patients exhibit increased mortality and morbidity, especially associated with cardiovascular events (Webster et al., 2017; Benjamin and Lappin, 2021).

The decrease in kidney function leads to the accumulation of substances in the body that in normal conditions would be eliminated in the urine (Lisowska-Myjak, 2014). These substances, known as uremic toxins, build up inside the organism and cause alterations in various body sites, including the oral cavity, with higher levels of urea, ammonia and pH being commonly detected in the saliva of CKD patients (Gupta et al., 2015; Simões-Silva et al., 2017). These changes are thought to be major selective factors responsible for altering the oral microbiome, which has shown significant association with CKD status (Simões-Silva et al., 2017, 2018; Hu et al., 2018).

On the other hand, it has been hypothesized that the oral microbiome may contribute in some extent to the progression of CKD. Three mechanisms have been proposed that link oral infection, especially periodontitis, to secondary systemic effects (Li et al., 2000; Simões-Silva et al., 2018). These include the generalized spread of infection from the oral cavity as a result of transient bacteremia; generalized injury resulting from circulating toxins of oral microbial origin; and systemic inflammation as a result of immunological activation induced by oral microorganisms (Li et al., 2000; Simões-Silva et al., 2018).

Among other complications, the accumulation of uremic toxins in the body leads to a chronic inflammatory status which can cause the suppression of lymphocytes and cell-mediated immunity, making CKD patients more prone to infections than other populations (Gupta et al., 2015). In fact, infectious episodes persist as the main weakness of PD programs, with peritonitis and exit-site/tunnel infections remaining as the most common and relevant concerns for these patients (Mujais, 2006; Liakopoulos et al., 2017). While Gram-positive agents (mainly Staphylococcus spp. and Streptococcus spp.) are responsible for the majority of infections, Gram-negative bacteria (mainly Pseudomonas spp. and members of the family Enterobacteriaceae) are more likely to provoke more severe infections with poor outcomes (Li and Chow, 2011; Liakopoulos et al., 2017).

The Enterobacteriaceae family is of particular relevance in end-stage renal disease patients. Aside from being recognized agents of nosocomial infections, Enterobacteriaceae are responsible for 10 to 12% of all peritoneal dialysis-associated peritonitis, constituting a significant risk for patients on PD programs (Jain and Blake, 2006; Szeto et al., 2006). Regarding the oral microbiome, enterobacteria are frequent inhabitants of periodontal pockets (Gonçalves et al., 2007), potentially contributing to generalized inflammation and disease progression in these patients. Furthermore, Enterobacteriaceae are critical level pathogens regarding antibiotic resistance phenotypes. The mechanisms of antibiotic resistance in these bacteria may be intrinsic due to the presence of features encoded by chromosomal genes related with functions such as antibiotic inactivating enzymes and non-enzymatic paths (efflux pumps, permeability or target modifications) that are observed in all or most of the members of a given bacterial species (Ruppé et al., 2015). In contrast, acquired antibiotic resistance is due to gene mutation or, more frequently, to incorporation of genes received through mobile genetic elements, such as plasmids or phages that often harbor genes encoding ß-lactamases, aminoglycosides modifying enzymes, or non-enzymatic mechanisms (Cox and Wright, 2013).

Antimicrobial resistant enterobacteria are quite prevalent in CKD patients and are the cause of serious drug-resistant infectious episodes (Wang et al., 2019). Probably because of frequent hospitalizations and frequent need for antibiotic therapy, patients with CKD tend to have more infections by drug resistant bacteria than people with normal kidney function (Fysaraki et al., 2013; Salloum et al., 2020). In 2014, outpatient hemodialysis facilities in the United States reported that, out of the isolates identified as agents of bloodstream infections and exit-site-related infections, 17.8% of Escherichia coli and 14.6% of Klebsiella spp. isolates were resistant to cephalosporins and 4.8% of Enterobacter spp. isolates were resistant to carbapenems (Nguyen et al., 2017). In another study conducted in a hemodialysis unit in a hospital in Algeria, 100% of K. pneumoniae isolates associated with catheter-related infections produced extended spectrum beta-lactamase and were resistant to at least gentamicin (Sahli et al., 2017). Regarding peritoneal dialysis, in PD patients who developed dialysis-associated peritonitis between 2002 and 2011, 54.3% of Enterobacteriaceae isolates were resistant to third-generation cephalosporins and 76.1% were resistant to fluoroquinolones (ciprofloxacin), and all of them produced extended spectrum beta-lactamase (Prasad et al., 2014). This data shows that a large percentage of enterobacteria responsible for dialysis-related infections in CKD patients also exhibit resistance to one or more antibiotic drugs.

Taking into consideration the above mentioned, this work aimed to assess the prevalence and antibiotic resistance profile of clinically relevant enterobacteria in the oral cavity of CKD patients undergoing PD.

Overall, our results show that the accumulation of uremic toxins in patients with advanced CKD and the consequent changes in the oral milieu (such as the increase in salivary urea, ammonia and pH) may indeed act as selecting factors on the oral microbiome, leading to an oral dysbiosis with proliferation of pathobionts in the oral cavity. In fact, 6 of the 8 species identified in Table 3 tested positive for urease production. Thus, it was observed that more urease-producing species were found in the oral cavity of CKD patients in comparison to the oral cavity of healthy controls, which is expected to be due to the previously reported adaptation of the microbiome to the altered conditions of the oral milieu, evidenced in Table 2 (Hu et al., 2018; Rysz et al., 2021).

The increased prevalence of Enterobacteriaceae spp. is significantly relevant in CKD patients. While enterobacteria are only occasionally detected in the oral cavity of healthy individuals, colonization is much more prevalent in immunocompromised patients, people who come in contact with hospital environments frequently, and people suffering from hyposalivation, oral lesions, and systemic disease, a scenario that fits CKD patients (Peleg and Hooper, 2010; Posse et al., 2017). Perhaps, all these reasons also contributed to the higher prevalence of Enterobacteriaceae in the oral microbiota of our CKD patients as well as to the fact that Enterobacteriaceae were responsible for approximately 8% of the peritoneal dialysis-associated peritonitis and 16% of exit-site infections in our group of patients, highlighting the clinical relevance of our results.

The increased prevalence of Raoultella ornithinolytica in CKD patients undergoing PD is a particularly significant finding. Formerly classified as Klebsiella ornithinolytica, R. ornithinolytica, a histamine-producing aquatic-commensal enterobacteria, is a virulent pathogen of community-acquired and emerging nosocomial infections, particularly associated with invasive procedures (Seng et al., 2016; Hajjar et al., 2018, 2020). R. ornithinolytica is able to occasionally survive in human saliva and a case report has pointed it as a cause of primary peritonitis in humans (Sibanda, 2014; Seng et al., 2016), thus aggravating the risk of its presence in the oral cavity of patients undergoing peritoneal dialysis. Regarding the oral microbiome, the high rate of colonization by Raoultella ornithinolytica is a very interesting finding in itself, since this species is still not described in the Human Oral Microbiome Database.¹

Although no association was found between previous Enterobacteriaceae infectious agents in the group of CKD patients and oral colonization by enterobacteria at the time of sampling, the difficulty in the identification of R. ornithinolytica and common misidentification as belonging to the Klebsiella genus should be considered, as a misidentification of the infectious agent is a possibility (Ponce-Alonso et al., 2016). Additionally, the hypothesis of an oral-peritoneal association regarding infection cannot be discarded without further studies, especially when all the Enterobacteriaceae-related infectious episodes reported in these patients occurred at least 1 year prior to the collection of the samples used in this study.

It is important to note that all the species identified in this study are clinically relevant and constitute potential cause of infection, with many being associated with PD-related infections (Barraclough et al., 2009; Li and Chow, 2011). Although contamination concerns for patients undergoing peritoneal dialysis typically tend to focus on contamination from external origins, the oral microbiome may be a major source of contamination both through external and internal routes (Li and Chow, 2011; Smith and Nehring, 2021). Bacteria inhabiting the oral cavity can translocate into the bloodstream as a result from routine daily activities, such as tooth brushing or invasive dental procedures, becoming opportunistic infectious agents in distant body sites such as the peritoneum (Smith and Nehring, 2021). In fact, the risk associated with the presence of pathogenic bacteria in the oral-nasal cavities of CKD patients is already recognized, which is why these patients are frequently screened for nasal carriage of Staphylococcus aureus as a way to prevent infection and the efficacy of prophylactic intranasal antibiotics for the treatment of confirmed nasal carriage of S. aureus has been tested in several prospective studies (Grothe et al., 2014; Li et al., 2016). Furthermore, the presence of these bacteria in the oral cavity can stimulate the immune system and contribute to the inflammatory state in CKD, which then leads to cardiovascular events, the main cause of death in these patients (Li et al., 2000; Webster et al., 2017; Höfer et al., 2021).

The oral cavity has also recently been highlighted as a potential reservoir of antibiotic resistance genes (Roberts and Mullany, 2010; Jiang et al., 2018; de Sousa Moreira Almeida et al., 2020). Resistance against amoxicillin was observed in 83% of isolates, although in these cases it is possibly intrinsic (Cox and Wright, 2013). However, putative acquired genetic elements related with antibiotic resistance were observed in 10 out of 23 oral isolates, 3 from healthy individuals (sul1 in Escherichia coli and bla_SHV in Klebsiella pneumoniae) and 7 from CKD patients (sul1, intI1, bla_CTX-M and incF in Enterobacter asburiae, Raoultella ornithinolytica and Klebsiella oxytoca). Curiously, the 5 isolates harboring the gene intI1 and the one exhibiting the high mobile replicon type incF and bla_CTX-M genes all originated from CKD patients. These observations raise the interest to investigate the vulnerability of some population groups to be colonized by bacteria harboring acquired antibiotic resistance genes and mobile genetic elements (de Sousa Moreira Almeida et al., 2020). Overall, regarding the resistance phenotype, 100% of the Enterobacteriaceae isolates were resistant to at least one of the antibiotics tested and only 6 isolates, all originating from CKD-PD patients, exhibited multidrug resistance. Genetic elements related to resistance were also detected in approximately 43% of isolates. Antibiotic-resistant enterobacteria are known agents of dialysis-associated infections in CKD patients, frequently originating serious life-threatening complications (Wang et al., 2019). Therefore, it is important to keep screening isolates from CKD patients for antimicrobial resistances, in order to plan efficient therapeutic interventions in case of infection.

Our study presents several strengths and some limitations. The fact that the control and CKD groups differ in age and gender may constitute a limitation of our study, although a binary logistic regression was applied using the prevalence of Raoultella ornithinolytica (the only member of the Enterobacteriaceae family with a significantly different prevalence between groups) as the dependent variable and sex, age, and type of participants as independent variables, confirming that only the type of participants (controls or CKD-PD) made a unique statistically significant contribution. Moreover, though ideally the groups should display similar demographic characteristics, several studies have reported that the oral microbiome does not suffer significant alterations throughout adulthood, with no significant differences being reported by gender or age groups (adults and young adults) at species level. Alterations in the healthy oral microbiota are only significant before adulthood or after the age of 65–70 years-old (Crielaard et al., 2011; Belstrøm et al., 2014; Lira-Junior et al., 2018; Verma et al., 2021). As so, the healthy young adults constituting the control group in this study are nevertheless representative of healthy reference values. Regarding methodology, although MALDI-TOF MS is a universally used method that has been routinely used in clinical microbiology (He et al., 2010; Tsuchida et al., 2020) and is considered extremely adequate for the identification of enterobacteria (Richter et al., 2013), it presents some limitations, as is the example of the inability to differentiate Escherichia coli from Shigella species (Devanga Ragupathi et al., 2017).

When one looks to the study strengths, firstly, it highlights how the uremic state in CKD can cause a dysbiosis of the oral microbiome and lead to a proliferation of enterobacteria, which are clinically relevant pathogens for patients on PD programs. Additionally, the identification of Raoultella ornithinolytica as a colonizer of the oral cavity is of extreme importance. The isolation of R. ornithinolytica from saliva samples has been reported before in the literature (Heggendorn et al., 2013; Derafshi et al., 2017), but never in such a high prevalence as in this study and never as an oral colonizer of CKD patients. It is important to keep in mind that this is an emerging species both in terms of infectious episodes in humans and in terms of resistance to antibiotic agents (Seng et al., 2016; Hajjar et al., 2020), two factors that make R. ornithinolytica a clinically relevant cause of concern, not only to CKD patients, but to the general population. The profiling of Enterobacteriaceae isolates in terms of antibiotic resistance, both regarding resistance phenotypes and the presence of genetic elements, is another strong point of this paper. As mentioned before, antibiotic-resistant enterobacteria are responsible for a significant percentage of infections in CKD patients (Prasad et al., 2014; Nguyen et al., 2017; Sahli et al., 2017; Wang et al., 2019), which makes their screening crucial for the clinical follow-up of said patients, given it can provide information for efficient antibiotic prescription in case of infection. Of note that infections are a major problem in peritoneal dialysis programs, leading to patients’ deaths or exclusion from peritoneal dialysis programs (Li et al., 2016). Therefore, the search of predisposing factors that lead to peritonitis and catheter-related infections in CKD-PD patients is of outmost importance. Finally, our study also contributes to the profiling of antibiotic resistance genes in the oral microbiome. In recent years, the oral cavity has been highlighted as a reservoir for resistance genes (de Sousa Moreira Almeida et al., 2020) and the profiling of these genes provides extremely relevant information for dentistry and the treatment of oral infections.

Our results suggest that the accumulation of uremic toxins in CKD patients induces alterations in the oral milieu, which, in turn, exert selective pressure on the oral microbiome, leading to its dysbiosis with significantly increased prevalence of urease-producing enterobacteria, specifically of Raoultella ornithinolytica, an opportunistic enterobacteria. Antibiotic resistance phenotyping revealed mostly intrinsic resistances, although resistances to sulfamethoxazole and streptomycin were also detected. Multidrug resistance was exclusively found in CKD-PD patients isolates. These results, together with the detection of acquired antibiotic resistance genes, alert to the possibility of the oral microbiota as a potential reservoir of resistance genes.

Enterobacteriaceae are known opportunistic pathogens of CKD patients, and this CKD-induced oral dysbiosis can potentially lead to an increase of resistant infectious and inflammatory episodes, contributing, in turn, to the progression of the disease and to its mortality rate. For all these reasons, studying the oral microbiome and its associated resistome and comprehend its role in infection episodes and in the progression of systemic diseases is of extreme relevance to ensure the health and well-being of CKD patients undergoing peritoneal dialysis.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee for Health of Centro Hospitalar Universitário de São João: (approval number 200/18) and Ethics Committee for Health of Faculdade de Medicina Dentária da Universidade do Porto: (approval number 12/2019). The patients/participants provided their written informed consent to participate in this study.

RM, AR, and BS-M contributed to conception and design of the study. CC, AM-R, CC, NS, LP, CF and AG contributed to data acquisition. CC, AM-R, CC, CF, RM, AR, CM and BS-M contributed to data analysis and interpretation. CC, CF, AA, RM, AR, CM and BS-M performed the statistical analysis. CC wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

This work is a result of the project POCI-01-0145-FEDER-029777, co-financed by Competitiveness and Internationalisation Operational Programme (POCI), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF) and through national funds by the FCT – Fundação para a Ciência e a Tecnologia. CC fellowship was supported by FCT/MCTES scholarship with the reference 2020.08540.BD. This work and CF were financially supported by FEDER through project “Assessing the risks associated with environmental antibiotic resistant bacteria: propagation and transmission to humans” (PTDC/CTA-AMB/28196/2017) – Programa Operacional Competitividade e Internacionalização, and by National Funds from FCT – Fundação para a Ciência e a Tecnologia and was hosted by CBQF through FCT project UIDB/50016/2020.

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.