Jan J. De Waele (a), Jerina Boelens (b) , and Isabel Leroux-Roels (b,c)
Purpose of review
Antimicrobial resistance (AMR) is increasing in ICUs around the world, but the prevalence is variable. We will review recent literature and try to answer the question whether this is a myth or a new reality, as well as discuss challenges and potential solutions.
AMR is diverse, and currently Gram-negative multidrug-resistant organisms (MDROs) are the main challenge in ICUs worldwide. Geographical variation in prevalence of MDROs is substantial, and local epidemiology should be considered to assess the current threat of AMR. ICU patients are at a high risk of infection with MDRO because often multiple risk factors are present. Solutions should focus on reducing the risk of cross-transmission in the ICU through strict infection prevention and control practices and reducing exposure to antimicrobials as the major contributor to the development of AMR.
AMR is a reality in most ICUs around the world, but the extent of the problem is clearly highly variable. Infection prevention and control as well as appropriate antimicrobial use are the cornerstones to turn the tide.
antimicrobial, antimicrobial resistance, antimicrobial stewardship, antimicrobial therapy, Gram-negative infection, multidrug resistance, sepsis
Sepsis is an important challenge in the ICU, with increasing incidences around the world in recent years . The mainstay of therapy in patient with sepsis and septic shock is antimicrobial therapy, apart from supporting failing organs and controlling the source of infection when relevant. Antimicrobial therapy should be administered in a timely manner, at the correct dose and should have the appropriate spectrum. As in most infections, the causative pathogen is unknown, empirical antimicrobial therapy will typically be broad-spectrum and may involve multiple antimicrobials (combination therapy) in many settings. Antimicrobial resistance (AMR) has been an issue as the discovery and subsequent development of antimicrobials, and resistance to an antimicrobial generally emerges less than 10 years after its introduction on the market . As such AMR is not a new nor a surprising phenomenon. However, the rate at which AMR is spreading around the world, the lack of new antimicrobials and the emergence of pathogens that are resistant to all available antimicrobials are worrying. AMR has been coined one of the greatest threats to healthcare [3& ,4] and may indeed impact many patients, and also indirectly compromise other interventions such as surgery or chemotherapy as infection is a common complication of many diseases or interventions. However, the true extent of AMR in the ICU is not easy to gauge. There may be bias in the literature, as often data on AMR come from tertiary care centers that may have a different patient population compared with other hospitals, or from specific countries with high incidences of AMR. Reporting of outbreaks may create the perception that infections caused by multidrug-resistant organisms (MDROs) are more prevalent than they actually are in reality. Finally, extrapolating epidemiological data on infections because of MDRO from certain countries or hospitals may not be the best way to estimate the true risk of AMR in an individual patient presenting with infection. Indeed, the epidemiology of AMR appears to be highly variable with important differences between countries, regions, hospitals, and even hospital wards. Although in one area a specific MDRO may be endemic, this same pathogen may be an uncommon finding in other areas. In this review, we will try to answer the question if AMR is a myth or a reality in ICUs around the world. We will take a critical approach to the most recent literature. Challenges and possible solutions will be discussed to put AMR in the ICU into proper perspective.
*Antimicrobial resistance in the ICU is a reality around the world, but there is important geographical variability; Gram-negative-resistant pathogens are the major challenge.
*Antibiotic stewardship and strict infection prevention and control are the cornerstones in preventing the spread of antimicrobial resistance in the ICU.
ANTIMICROBIAL RESISTANCE, MULTIDRUG RESISTANCE, … WHAT’S IN A NAME?
AMR refers to the overall phenomenon of bacteria being resistant to antimicrobials. This is a natural phenomenon and microorganisms may be intrinsically resistant to certain antimicrobials. Although this may imply that certain antimicrobials cannot be used to treat pathogens, this is not problematic as these patterns are well known and this will not compromise empirical therapy. AMR may also be acquired by bacteria with multiple mechanisms responsible for this – these will be different for different antimicrobials and may include the production of a variety of b-lactamase enzymes, changes to the target proteins, loss of porins, or overexpression of transmembrane efflux pumps . Many of these resistance genes are localized on plasmids, mobile genetic elements that can easily transfer resistance within or between different bacterial species, in particular, Gram-negative bacteria. Multidrug resistance (MDR) may be the result of one or more of these phenomena; MDR is the situation where a pathogen has become resistant to multiple antimicrobials, and fewer options are available to effectively treat an infection. In 2012, Magiorakos et al.  proposed a conceptual framework in an attempt to harmonize the terminology used. MDR was defined as acquired nonsusceptibility to at least one agent in three or more antimicrobial categories, extensive drug resistance was defined as nonsusceptibility to at least one agent in all but two or fewer antimicrobial categories (i.e., bacterial isolates remain susceptible to only one or two categories) and pan-drug resistance was defined as nonsusceptibility to all agents in all antimicrobial categories. This classification implies that bacterial isolates should be tested against all or nearly all of the antimicrobial agents within the antimicrobial categories, which is rarely performed, yet these terms have been used in clinical practice around the world. Although this may not always be completely correct, these terms reflect increasing levels of AMR and are useful to grade the level of AMR. Surveillance networks such as European Antimicrobial Resistance Surveillance Network from the European Center for Disease Control use a standardized protocol and focuses on specific MDROs (https://www.ecdc.europa. eu/en/about-us/partnerships-and-networks/diseaseand-laboratory-networks/ears-net). At the bedside however, it is more common to name specific pathogens or resistance patterns when discussing AMR in patients. Clinicians would rather name methicillin-resistant Staphylococcus aureus (MRSA), Acinetobacter or carbapenem-resistance as challenges rather than AMR as a whole. This is not surprising as the pathogens or resistance patterns vary considerably across the world. The current use of AMR as a blanket term when referring to reduced susceptibility to antimicrobial drugs fails to address the important differences that exist between the pathogens in terms of how they may transfer resistance patterns to other pathogens as well as how transmission between patients occurs. Particularly, Gram-negative bacteria are often lumped together, but an increase in extended-spectrum b-lactamase (ESBL) producing – Enterobacterales has different implications and solutions than an increase in Acinetobacter colonization.
ANTIMICROBIAL RESISTANCE AS A SPECIFIC CHALLENGE FOR THE ICU
The prevalence of MDRO in the ICU is determined by both the influx of MDRO positive patients, as well as the number of patients who acquire MDROs during their stay in the ICU. The latter is mainly caused by patient-to-patient transmission via the hands, fomites, or environment [7& ]. Obviously, one cannot control the influx of patients, but this confirms the importance of admission screening for nosocomial pathogens, and the usefulness of surveillance cultures during the stay of the patient. This allows to monitor resistance levels and also evaluate the efficacy of infection control practices and early detection of outbreaks. In the ICU patient many if not all known risk factors for AMR development are present: often extensive exposure to antimicrobials, prolonged length of stay, use of invasive devices, a vulnerable host, and many more . The ICU therefore is an environment where there is a high risk of amplifying MDRO prevalence if infection control measures are not strictly applied. Although decolonization may seem an attractive option for gram-positive carriers – albeit that the evidence in the ICU is limited for that – a recent guideline does not recommend decolonizing multidrug-resistant Gram-negative (MDR-GN) carriers in any setting.
RECENT TRENDS IN GRAM-POSITIVE BACTERIAL INFECTIONS
The two major challenges in Gram-positive infections are MRSA and vancomycin-resistant enterococci (VRE). Although MRSA has been a challenge in many hospitals and ICUs in the past decades , currently the problem has been decreasing although it has not disappeared. This was also reflected in trends of bloodstream infections in the United States over 20 years where the incidence of MRSA was declining . The Abdominal Sepsis Study (ABSES) reported an overall prevalence of MRSA in intraabdominal infections (IAI) of 1% . Vancomycin resistance in Staph. aureus has been reported but remains at a low level throughout the world. Similarly, linezolid resistance may be encountered but is rarely a cause of concern in an individual patient. VRE is variably encountered, and outbreaks may occur [13,14]. In IAI, globally a prevalence of 2.8% was reported, again with considerable geographical variability . Patients in the ICU seem to be at a higher risk of acquiring VRE bacteremia  and also prolonged exposure to antimicrobials is an important risk factor for VRE bacteremia ; not surprisingly nosocomial VRE infections significantly increase hospital costs compared with susceptible infections.
RECENT TRENDS IN GRAM-NEGATIVE BACTERIAL INFECTIONS
The most important pathogens involved are ESBL producing Enterobacterales, MDR Pseudomonas aeruginosa, carbapenemase-producing Enterobacterales (CPE), carbapenem-resistant Enterobacterales (CRE), and Acinetobacter spp. Although Gram-positive bacterial infections may be no longer perceived as major threats in most ICUs, MDR in Gram-negative bacterial infections are the major challenge around the world. Clinicians perceive MDR-GN as an important threat. In a recent survey ESBL, carbapenem-resistant Acinetobacter baumanii and carbapenem-resistant Klebsiella pneumoniae were named as the main threats but there was little if any consensus on what the preferred strategies were to combat these pathogens . Potential solutions cited by respondents included cohorting, formulary restrictions, increasing the nurse/patient ratio among others. Considerable differences were found according to the background of the respondents (intensivists versus infectious diseases specialists). Importantly, the majority of the participants were based in countries where AMR is very prevalent, for example, Greece, Turkey, South Africa, among others – which will undoubtedly have affected the results. Rates of reported ESBL colonization are increasing in ICUs worldwide, because of the successful spread of these microorganisms in both the community and hospital [18&&]. It is important to differentiate ESBL E. coli from non-E. coli (mainly ESBL K. pneumoniae) in this respect, as the transmissibility of these MDRO is significantly different, with a 3.7 times higher risk of transmission for non-E. coli ESBL [19& ]. Universal screening for carriage of ESBL through surveillance cultures and a strategy of contact precautions in carriers appears of limited value to prevent cross-transmission when compliance with standard precautions was high, and no outbreaks were involved. This also does not contribute to a better use of carbapenems . One of the striking observations is the huge geographical variation in MDR-GN pathogens, both in overall prevalence as well as in the specific MDRO types that are most prevalent. In the ABSES study, focusing on IAIs, resistance in Gram-negative bacteria varied from 9% in Western Europe to 46% in Africa and the Middle East . Carbapenem-resistant Gram-negative bacteria were almost not encountered in Western Europe (0.5%), compared with 15% in Eastern and Southern Europe. Clearly, geographical differences are huge, and therefore the local ecology is key in determining the preferred strategies to combat AMR. Outbreaks are an important source of MDR-GN infections throughout the world, particularly with P. aeruginosa or Acinetobacter: a recent review focusing on MDR P. aeruginosa and Acinetobacter outbreaks found Acinetobacter outbreaks to be found more often in ICUs compared with MDR P. aeruginosa outbreaks that occurred both in ICUs and non-ICUs [21& ]. Not surprisingly, transmission via contact (either person-to-person or via environmental surfaces) was more often cited as the cause of the outbreak for Acinetobacter, whereas in P. aeruginosa, transmission via contaminated devices or water was the culprit more frequently. Infection prevention and control (IPC) measures including surveillance, isolation, closure of wards, was applied more often in Acinetobacter outbreaks. Although the overall perception is that MDRGN infections are on the rise, a number of studies did however report a reduction in MDRO colonization and/or infection, particularly of the more problematic pathogens such as MDR P. aeruginosa or Acinetobacter. In a longitudinal study covering patients admitted to six New York ICUs between 2006 and 2014, Furuya et al.  reported a 6–7% decrease in MDRO, which could not be linked to the introduction of universal contact precautions; the authors attributed these effects to other hospitalwide improvements in infection control during the study period. Italian investigators found that a multifaceted intervention, aimed to reduce selection and transmission of MDRO led to dramatic reductions . For all Gram-negative microorganisms including carbapenemase producing K. pneumoniae, Acinetobacter, and MDR P. aeruginosa, resistance dropped from 91 to 13%. The role of surfaces in cross-contamination within the ICU should not be underestimated in the transmission of MDROs. A recent study focusing on the spatiotemporal dynamics of multidrug-resistant bacteria on ICU surfaces found that – apart from the huge differences in positive cultures from surfaces between US and Pakistani ICUs – MDR isolates with high resistance gene burdens are commonly found. Many resistance genes are shared by multiple species; moreover, coassociation of Acinetobacter baumannii and Enterococcus faecium was found on multiple surfaces with these species establishing synergistic biofilms in vitro . New disinfectants have been found to be more effective in eliminating biofilms, hence controlling this route of transmission . Handwashing sinks have also been identified as potential reservoirs of MDRO, in particular CPE/CRE and P. aeruginosa [26,27]. Removing the sinks and water-free patient care led to a reduction of acquisition of Gram-negative organisms in a Dutch ICU . Although an increase in MDROs in communityacquired infections such as pneumonia has been reported worldwide , the prevalence of these pathogens is overall low, and it appears that when it develops, it is limited to patients with either recent antibiotic exposure and/or chronic respiratory diseases and/or healthcare exposure and/or previous colonization.
WHAT TO EXPECT IN THE NEXT DECADE(S)?
It is difficult and even impossible to predict the future evolution of AMR, but if recent trends continue, then a persistent increase may be expected. Obviously, this is something that requires immediate action. In several countries, national action plans have been developed to reduce the number of infections by MDRO in the next years. Insights in the mechanisms that are associated with the spread of MDR pathogens are increasing and may assist in devising strategies to better control the spread of AMR. This is not limited to the control of outbreaks. Better risk stratification may be a tool to refine strategies aimed at MDROs, but with current tools, this will be difficult to obtain. Patient risk factors for acquiring MDR-GN (e.g., mechanical ventilation, antimicrobial exposure, immunosuppression among many others) have been studied extensively but the accuracy of scores predicting either colonization or infection is variable, although the latter may perform slightly better [30& ]. Often the external validity of the scores is limited, and their use at the bedside is questionable. Compliance with hand hygiene remains the cornerstone to prevent transmission of MDROs in ICUs. In settings of high compliance, screening for MDRO colonization and contact precautions for MDRO has little additional effect . However, in most settings low compliance remains a problem, also in a high-risk environment as the ICU. A recent systematic review showed that the hand hygiene compliance in ICU’s worldwide is suboptimal and varies according to geographic region (65% in highincome countries, 9% in low-income countries), type of ICU (67% in neonatal, 41% in pediatric, and 58% in adult) and healthcare worker (HCW) profile (43% for nursing staff, 33% for physicians, and 54% for other staff) [32&&]. Equally important are correct environmental cleaning and disinfection of medical devices and avoiding creating reservoirs in the ICU, for example, high touch surfaces, sinks where biofilm formation is looming.
HOW TO MOVE ON FROM HERE
Tackling AMR requires an integrated approach, both at the societal and at the hospital level. Antimicrobial stewardship programs (ASP) have been introduced in many hospitals and are a key factor to reduce antimicrobial exposure, the primary driver of AMR. Developing better antimicrobial strategies is urgently needed, and pharmacokinetic/pharmacodynamic-based optimized dosing may offer opportunities to enhance the efficacy of antibiotic treatments and prolonging the shelf life of antimicrobials. This applies to both new and already available antibiotics. At the same time, care should be taken to use the newly developed antimicrobials such as ceftazidime/avibactam and ceftolozane/ tazobactam in an optimal way, reserving them for infections with pathogens that have few if any other treatment options left. In this context, acknowledging the importance of IPC in controlling MDR is crucial, and IPC should be closely intertwined with ASPs in every hospital. But also other factors that – often indirectly – contribute to the development or spread of MDR should be considered, for example, HCW staffing and workload, infrastructure, and architecture. Furthermore, adequate knowledge among HCW about the transmission of MDRO is critical. In a large French study, awareness of transmission was low, and knowledge of associated control measures was low . Although this was less of an issue in HCW in the ICU, this shows that education still remains important. Lack of knowledge was also cited by respondents in a Greek survey on this topic . Individual difference between MDR pathogens should be acknowledged as this distinction is essential for better strategies to combat AMR. IPC strategies to limit spread may differ for different pathogens, and a more pathogen-oriented approach may be desirable. In all, the future of AMR maybe not so grim if we at least acknowledge the challenges we face now and continue to improve the management of these patients. Several elements may contribute to this such as new insights in routes of transmission leading to new strategies to further reduce this for example avoiding the use of water in the ICU. Also the development of new antimicrobial and more importantly nonantimicrobial strategies to combat infections, and the potential use of fecal microbiota transplantation as a strategy to abolish MDRO colonization. This will require a continued awareness among all HCW of the potential threats, as well as accurate and real-time data of AMR within the hospital but also at the regional and national level. Now largescale data are often only released years later – as the variability is large these are often of limited use and unit-specific monitoring is crucial.
It would be inappropriate to say that AMR is a myth in the ICU – although the pathogens that are a concern may evolve over time, this phenomenon is a reality in ICUs across the world. However, the reality varies from unit to unit, and although in some units MDROs have become endemic, they remain a rare finding in others. Although outbreaks may occur in any setting, these can often be controlled by identifying infected and colonized patients, adequate antibiotic therapy of infected patients and enforcing IPC measures. On the other hand, we argue that it is a myth that AMR is a problem we can do nothing about. IPC has a critical role in preventing spread and controlling outbreaks. More importantly, we need to stop relying on new antimicrobials only, and implement a holistic view on AMR in the ICU and beyond.