The prevalence of MDRO infections in cirrhosis varies across studies, and several factors may explain this heterogeneity. First, different definitions have been applied over time and across datasets, with many not meeting commonly adopted criteria (e.g., resistance to at least one agent in three different antibiotic classes). Second, systemic infection and colonization have often been used interchangeably, without providing useful information regarding the need for targeted antibiotic therapy. Third, MDRO prevalence may vary according to differences in surveillance protocols and/or healthcare facilities. Lastly, many infections in cirrhosis turn out to be culture-negative (e.g., pneumonia, spontaneous bacterial peritonitis), thus the actual prevalence could be underestimated.

According to available literature, MDR strains are responsible for 20–30% of culture-positive BI, with heterogeneity in prevalence and predominant strains across geographical areas, also highlighting environmental factors that could influence the spread of such infections. For instance, extended-spectrum beta-lactamase enterobacterales (ESBLE) are frequently isolated across Europe, North America, and Asia, whereas methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) predominate in North America. A prospective, observational, multicenter study that included forty-six centers worldwide reported that MDR accounted for 34% of culture-positive infections [12]. Data from a European cohort of more than 450 patients showed that 29% of culture-positive infections were caused by resistant rods [13]. Another multicenter study in Italy reported a 27% prevalence of MDR infections in 395 patients, mainly due to gram-negative rods [14]. These results were consistent with data from a large cohort of 876 patients listed for LT between 2006 and 2014, where the rate of MDR-related BI was 24.2% [15]. The prevalence of MDR pathogens in the ICU setting is variable according to studies [16, 17], but it may rise up to 30–50% of culture-positive infections.

Many risk factors for MDRO infection have been identified in ESLD patients, the majority of these being shared with other chronic conditions: repeated infection and/or hospitalizations, previous/recent exposure(s) to systemic antibiotics, long-term ICU or hospital stay, indwelling or central catheters, invasive procedures.

Among preventable factors, the role of prior antibiotic course(s) is worth mentioning. Antibiotic overuse is a major healthcare problem derived from many causes as environmental exposure, diffuse spread in agriculture and farming. However, it is interesting that in cirrhosis, since half of BI episodes are usually culture-negative, a targeted antibiotic therapy is not always feasible, therefore the odds of long-lasting empiric therapies are higher, with reduced use of de-escalating strategies.

Antibiotic prophylaxis has been advocated as a further risk factor for subsequent MDR infection, but this point remains debated. First, it should be mentioned that, according to Guidelines, patients with cirrhosis who require long-term antibiotic prophylaxis are a restricted subpopulation, namely those with low ascites fluid protein concentrations and those with prior episodes of SBP [18, 19]. Data on the association between norfloxacin prophylaxis and MDRO infection came from observational studies [20, 21], whereas more recent multicenter experiences did not confirm these findings [12, 22]. There is significant evidence that prophylaxis may enhance the risk of MDRO colonization, which is, in turn, another known risk factor for MDRO infection.

The association between MDRO colonization and further systemic infection from the same strain has been largely demonstrated in patients with and without cirrhosis. A recent study showed that ICU patients with cirrhosis had a higher MDRO colonization rate than patients without, confirming that such colonization increased by 7-folds the risk of being infected by the same strain. Notably, this risk was observed both for Gram -ve and Gram + ve strains, in two different Countries [17].

Currently, a single biomarker or a combination of many is not sufficiently accurate for diagnosing bacterial infection in ESLD. Therefore, diagnosis should be based on integrating clinical suspicion, biomarkers, radiological features, and microbiological samples. In cirrhotic patients, rapid diagnosis and initiation of appropriate empirical (or sometimes targeted) antibiotic therapy are crucial for patient survival. Diagnostic kits based on various technologies (such as MALDI-TOF MS and multiplex PCR) are increasingly available. These kits can identify the pathogen(s) and determine their antimicrobial susceptibility within 1 to 6 h, facilitating not only rapid empirical therapy institution but also a rapid de-escalation [23]. Although the high cost may pose a potential barrier to their use, their utilization should be encouraged to enable the initiation of effective therapy as quickly as possible, especially in cases of severe clinical infection, as in the case of high suspicion of MDRO infection [24].

A multidisciplinary management is of paramount importance for preventing the spread of resistant rods and ensuring a timely, effective treatment [4, 25]. Clinical history and the use of tools to predict the risk of MDRO infection should be employed when selecting antibiotic therapy. In recent years, infection severity has guided the choice of empirical antibiotic therapy, typically consisting of two antibiotics from different classes [6, 8, 23].

When an MDRO infection is highly suspected, and when the clinical presentation is severe (e.g., septic shock) the use of appropriate large-spectrum therapies, involving new molecules such as ceftazidime/avibactam, cefiderocol, daptomycin is justified. It should be followed by a rapid de-escalation, if possible, once microbiological tests turned out positive, to reduce resistance rates. An appropriate administration of such antibiotics—such as continuous infusion in the case of beta-lactams [26]. – appears of utmost importance, too, to increase their effectiveness, especially in sickest patients with ascites who have a high volume of distribution.

The management of MDRO infections, especially if severe, also relies on non-antibiotic therapies, which are similar to what is commonly applied in sepsis. These strategies include adequate caloric support, appropriate resuscitation strategies, management of vascular catheters responsible for the infection, and source control [27].

The outcome of patients with ESLD, especially with ACLF, appears to be significantly reduced when a superimposed infection occurs. For instance, a study by Fernandez and Colleagues demonstrated that bacterial infection significantly increased mortality in patients with ACLF (90-d survival 49% vs 72%), being these independently associated with mortality, both in patients with ACLF-1 and ACLF-2 [28]. The prognosis seems to be even worse when the infection is caused by MDRO strains. Indeed, the aforementioned multicenter study from Europe confirmed that these infections were associated with a lower resolution rate (71.4% vs. 87.6%), a higher prevalence of severe sepsis/shock (31.9% vs. 12.2%,), ACLF (67.5% vs. 45.6%) and, ultimately, a significantly higher 28-day mortality (35.1% vs. 18.1%) [13]. Taken together, these data pose many questions regarding sickest patients with ACLF awaiting LT and experiencing MDRO infection. On one hand, the bad short-term prognosis in terms of transplant-free survival, counterbalanced by the good post-operative survival in patients with ACLF [29], could represent a factor to proceed with the transplant. Conversely, a severe, uncontrolled MDRO infection with multiorgan failures should be viewed as a condition where the transplantation may be potentially inappropriate.

Therefore, the decision to proceed with transplantation in patients with MDRO-controlled infection should be made on a case-by-case basis, identifying the correct timing and balancing the benefits of transplantation with the risks associated with the type of infection (pandrug, extended drug resistance), source control, and available therapeutic options for the post-operative phase [23, 30, 31].

Preventing and reducing modifiable risk factors for MDRO infection in ESLD not only represents a research field, but it will be also a major goal for the future. A continuous update of local epidemiology (indeed, many studies have documented significant changes in the prevalent MDR strain within the same ward) and the stratification of patients at higher risk, based on the available tools, may be helpful and cost-effective preventive strategies.

A further option is to foster antibiotic stewardship, which appears essential for reducing antibiotic pressure, and consequently, the risk of colonization and infection from MDR organisms [4].

Since rectal colonization confers a high risk of subsequent invasive infection, fecal microbiota transplantation could be promising preventive strategy in pre-transplant candidates. Bajaj et al. summarized the current evidence regarding this option in many settings of cirrhosis, with favorable results [32]. There are, however, major points to be considered. First, indications are heterogeneous (alcoholic hepatitis, Cl. Difficile infection, hepatic encephalopathy) and, at present, do not specifically include resolution of colonization. Moreover, the role of microbiota transplantation as a game changer in liver disease progression, through an improvement of inflammation, warrants investigations. Second, it is currently unknown what is the preferable method (capsules, enemas, colonoscopy). Third, there are safety issues regarding donor screening and the risk of intestinal infection.

Finally, infection prevention requires a series of "non-pharmacological" strategies that are markedly beneficial for patients, such as the proper management of devices (e.g., venous and arterial vascular catheters, ventilators), as well as hand hygiene and the use of safety devices. We believe that training new operators and retraining experienced ones, along with evaluating the trends of MDRO infections and continuously updating protocols, can be valid options.