Antimicrobial resistance (AMR) in patients with sepsis is a growing problem worldwide. ^, However, the more subjective the sepsis clinical syndrome is, the less specific the actual sepsis diagnosis. Sepsis illustrates this dilemma, that is, about one-third of the cases defined by the sepsis clinical syndrome do not have an infectious etiology. To make matters more complex, out of the patients who actually have an infectious etiology for sepsis, only 60% to 70% have a positive bacterial culture from any site, which makes the selection of antibiotics more prone to errors due to the lack of susceptibility patterns. In addition, the vast majority of IV antibiotics are administrated in intensive care units (ICUs) around the world. The combination of these 3 factors is the inevitable over utilization of inappropriate antibiotics, which may lead to individual harm due to the direct drug side effects in patients already critically ill, as well as to major shifts in the ICU microbial population, which may lead to more critically ill patients being exposed to multi-drug resistant infections.

This article will describe the most current world scenario of AMR in patients with sepsis, from epidemiology to diagnostics to therapeutics, in order to provide a comprehensive understanding of today’s challenges and the potential solutions for the future.

Sepsis and AMR represent significant, growing, and intertwined threats to global health. Sepsis remains a major challenge in all settings, but one that disproportionately impacts lower income countries, regarding both incidence and mortality. Even in more affluent settings it remains a public health concern that is not likely to diminish any time soon due to the aging of the population and growing incidence of comorbid conditions.

The growing awareness about the challenge of sepsis led the World Health Organization to publish their first global report on sepsis epidemiology. According to data from the Global Burden of Disease study, the number of sepsis cases, worldwide, in 2017 was estimated to be close to 49 million with approximately 11 million (∼22.5%) sepsis-related fatalities reported. Of these, more than 80% of cases and deaths occurred in countries with low to middle socio-demographic indexes, highlighting the unequal burden that low-income and middle-income countries (LMICs) face.

The usual clinical requirement to initiate appropriate (ie, active in vitro against the causative organism) antimicrobials as soon as possible in septic shock and, to a lesser extent, in sepsis, often leads to the use of empirical broad spectrum antimicrobials. This strategy is particularly relevant in settings of widespread AMR and increases the risk of further development of resistant strains, thus potentially creating a vicious cycle.

In the recently published WHO global AMR surveillance report, ^, the percentage of infections caused by AMR bacteria is over 17% (meaning 1 out of every 6 infections regardless of setting). There is significant geographic heterogeneity but, even in the lower incidence areas of Europe and the western pacific, the percentage is around 10%.

The issue of resistance is particularly worrying in gram-negative bacteria where resistance to quinolones and 3rd generation cephalosporins has become endemic in several regions and resistance to carbapenems is showing an alarming growth.

Of course, the impact of AMR is greatest in the critically ill as illustrated in the prospective Eurobact II cohort where, among 2042 ICU patients with ICU-acquired bloodstream infection (BSI), those with DTR gram-negatives had significantly increased 28 day mortality when compared with those infected by susceptible strains (19.2 vs 10.1%; OR 1.71; P <.001). This scenario is clearly worse in LMICs, as demonstrated in a multicenter study in Brazil, where 50.5% of BSI among ICU patients were caused by an antimicrobial resistant bacteria from the WHO priority list, and reported carbapenem resistance rates in 45.5% and 94.4% of Enterobacterales and Acinetobacter baumannii , respectively.

Another large study from the United States, also focusing on BSI analyzed the impact of “discordant” therapy (meaning that none of the empirically used antibiotics was active in vitro against the causative MO) in 26,036 patients. Of these, 83% ( n = 21,608; 42.2% with sepsis and 14.2% with septic shock) received antibiotics on the same day that blood cultures were collected, which was used as a surrogate marker for severe infection. Infection by an antimicrobial-resistant strain was associated with a markedly increased risk of receiving discordant therapy (aOR = 9.09 [95% CI 7.68–10.76]; P <.001). Discordant empirical therapy independently increased mortality by 37% (aOR = 1·37 [95%CI 1.24–1.51]; P≤ .001). Thus, sepsis represents an ongoing public health threat and the rising incidence of AMR increases the mortality risk by decreasing the proportion of patients receiving appropriate empirical antibiotics.

This section focuses on prevention strategies aimed at mitigating AMR in patients with sepsis, with particular emphasis on interventions that reduce the incidence and transmission of infections at ICUs. The following strategies are framed within a preventive hierarchy, in which infection prevention and control (IPC) measures constitute the cornerstone strategy, antimicrobial stewardship provides complementary protection at the patient level and potentially to reduce overall AMR rates, and effective communication within ICU teams enables sustained implementation.

Early recognition of septic patients and the prompt initiation of appropriate antimicrobial therapy are associated with reduced mortality in septic shock patients. ^, In this context, the clinical microbiology laboratory plays a critical role in identifying infectious agents and determining their antimicrobial susceptibility profiles. Diagnostic stewardship is a systematic approach to optimizing test selection, execution, and result reporting to ensure that appropriate diagnostics are applied to the appropriate patient at the appropriate time, thereby generating accurate and clinically actionable information that guides therapeutic decision-making, promotes appropriate antimicrobial use, mitigates AMR, improve resource utilization, and ultimately enhance patient outcomes. ^,

Traditional culture-based methods are largely employed for pathogen identification and antimicrobial susceptibility testing, AST ^,,. Blood cultures processed through continuous monitored automated systems (eg, Bactec FX, VersaTREK, BacT/ALERT Virtuo) usually requires time to allow the growth of such pathogens and can be affected by numerous factors including skin antisepsis, blood volume collected, incubation delay of blood culture bottles, and previous administration of antibiotics. ^,,,,, Traditional diagnostic tests may require up to 96 hours to yield results, which limits their usefulness for timely antimicrobial decision-making. During this interval, patients typically remain on empirical therapy that may be ineffective or unnecessarily broad-spectrum, contributing to AMR, drug-related toxicity, and increased health care costs. ^,,,,,

The incorporation of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has revolutionized diagnostic speed by substantially reducing the time required for pathogen identification. ^, However, this methodology still depends on blood culture positivity that typically requires approximately 12 to 24 hours. Despite MALDI-TOF MS limitation such as reduced performance in polymicrobial infections, difficulty distinguishing closely related species, and inability to routinely detect resistance genes, it provides more accurate pathogen identification than conventional staining methods performed after blood culture positivity, thereby supporting more appropriate antimicrobial prescribing. A meta-analysis of 21 studies including 14,515 patients demonstrated that MALDI-TOF MS significantly reduced the time to pathogen identification by 22.86 hours and the time to effective antibiotic therapy by 5.07 hours compared with conventional phenotypic methods. Earlier initiation of effective therapy was associated with a 23% reduction in in-hospital mortality among patients with BSIs, with greater benefit observed when MALDI-TOF MS was implemented alongside ASPs (35% mortality reduction). In addition, implementation of MALDI-TOF MS was associated with a 0.73-day reduction in hospital length of stay and an average cost saving of US$ 4140.18 per hospitalization, despite the high upfront cost of the technology. However, all analyzed studies were focused on patients with BSIs, which limits the ability to estimate its true impact on outcomes among patients with sepsis or septic shock. A single study evaluating septic patients found that a modified workflow using MALDI-TOF MS and rapid susceptibility testing for positive blood cultures reduced time to pathogen identification from 28.3 to approximately 10 hours and time to optimal antimicrobial therapy from 36.1 to 17.9 to 21.6 hours, accelerating targeted treatment and potentially improving outcomes.

Similarly, to accelerate pathogen identification in septic patients, molecular assays, most based on multiplex real-time polymerase chain reaction (PCR), have been developed. Assays like BioFire Blood Culture Identification (BCID and BCID 2; bioMérieux, Inc., Hazelwood, MO), Unyvero BCU Application (Curetis GmbH, Germany), and Verigene (Nanosphere, Northbrook, IL), can provide results between 1-6h, and when combined with knowledge of local epidemiology, may support antimicrobial optimization and de-escalation. ^,,,,, The operational advantages of multiplex PCR panels are as follows: (i) the ability to rapidly identify pathogens and resistance genes, including in polymicrobial infections; (ii) operational simplicity with reduced dependence on specialized expertise; and (iii) availability for continuous (24 hour) processing features that are particularly valuable during off-hours, when trained personnel may not be available. ^, However, key limitations include the requirement for positive blood cultures, incomplete coverage of less frequent bacterial/fungal species or resistance genes, and potential misidentification of closely related organisms. ^,, Rapid diagnostic testing for BSIs was systematically evaluated by an expert panel convened by American Society for Microbiology in collaboration with Centers for Disease Control and Prevention and Battelle consultants using a GRADE-based approach across 51 studies. Most available evidence on the clinical impact of rapid diagnostic testing derives from retrospective observational studies and is therefore of limited strength, with insufficient data to compare specific technologies (eg, molecular assays vs MALDI-TOF MS). Despite these limitations, expert consensus strongly supports implementing rapid diagnostics in combination with structured, active communication of actionable results, as this integrated approach consistently reduces time to targeted therapy and hospital length of stay, whereas rapid tests used in isolation are not recommended. Evidence for reductions in ICU length of stay remains limited. Rapid diagnostics with active communication may also contribute to lower non-specific, in-hospital, and infection-related mortality; however, supporting evidence is low quality and recommendations are weak due to heterogeneity and uncertainty across studies.

Detection of a resistance gene does not necessarily predict phenotypic resistance; therefore, AST remains essential to confirm susceptibility and guide therapeutic optimization. AST results are critical for adjusting antimicrobial regimen, particularly in septic patients, in whom determination of the minimum inhibitory concentration (MIC) of the causative pathogen enables dose optimization to maximize the probability of achieving pharmacokinetic/pharmacodynamic (PK/PD) targets. To reduce AST turnaround time (TAT), both EUCAST and CLSI recommend the rapid antimicrobial susceptibility testing approach, which uses 150 μL of blood from a positive blood culture bottle to inoculate antimicrobial plates ^, Although this technique can shorten AST TAT, it increases laboratory workload because plates must be read at multiple time points. Another limitation is that interpretive breakpoints are standardized only for selected bacterial species. ^, To further reduce AST TAT, rapid antimicrobial susceptibility platforms have been developed. These systems rely on different technological principles as shown in Table 1 . ^,,,,, To date, the clinical impact of novel rapid AST methodologies has been evaluated for the Accelerate PhenoTest BC Kit (AXDX). In the largest multicenter study, including 854 patients with BSIs, implementation of this platform significantly reduced the median time to optimal antimicrobial therapy from 40.9 to 23.7 hours (17.2 hour reduction; P <.0001), the time to first antimicrobial modification from 24.2 to 13.9 hours ( P <.0001), and the time to antimicrobial de-escalation from 36.0 to 27.2 hours ( P =.0004). Among patients with Gram-positive bacteremia, the system shortened the time to optimal therapy from 36.3 to 20.4 hours ( P =.01) and reduced the duration of unnecessary MRSA coverage from 58.4 to 29.7 hours ( P =.04). No differences in the mortality (8.7% vs 6.0%), length of stay (7.0 vs 6.5 days), and adverse drug events were noticed after employing AXDX AST interpretations and PK/PD targets associated with clinical response are derived from breakpoints established using standardized reference methods that determine MICs with defined inocula. ^,,,, Machine-learning approaches that infer antimicrobial susceptibility from genomic data show promise, particularly for pathogen–drug combinations in which resistance is predominantly driven by chromosomal mutations in well-characterized genes (eg, Salmonella spp. and fluoroquinolones) However, predictive performance will be challenging for agents such as carbapenems and β-lactam/β-lactamase inhibitor combinations, in which mechanisms of resistance are complex and MIC values can be influenced by inoculum effect. ^,

Roche Molecular Diagnostics introduced LightCycler SeptiFast as the first commercially available nucleic acid amplification test (NAAT) for the rapid identification of pathogens directly from whole blood, followed by the development of additional NAAT-based platforms. Table 1 summarizes these NAAT technologies designed for the direct detection of microorganisms and selected resistance genes without the need for prior blood culture positivity. ^, Despite their rapid turnaround time, these assays have demonstrated heterogeneous diagnostic performance, including reports of suboptimal sensitivity, which are partly attributable to PCR inhibitors present in whole blood, as well as limitations in specificity coupled with the extensively MALDI-TOF MS use Although these technologies enabled earlier pathogen identification, several platforms, including LightCycler SeptiFast, T2Bacteria, and IRIDICA BAC BSI, have been withdrawn from the market. InfectID-BSI (Microbio Ltd., Australia) is a recently developed molecular assay that uses 0.4 mL of whole blood to identify the 20 most common bacterial pathogens and 6 Candida species associated with sepsis within approximately 3 hours. In a 2 center study conducted in Australia, InfectID-BSI demonstrated superior detection compared with conventional blood cultures for both Gram-positive bacteria (56 vs 19 detections) and Gram-negative bacteria (46 vs 35 detections), while maintaining a very high negative predictive value (>99.0%). However, the current version of the assay does not include detection of AMR genes. A new NAAT, UllCORE BSI, currently under development by deepull, is designed to identify pathogens (bacterial and fungal pathogens) and selected resistance markers directly from 8 mL of whole blood from patients with suspected infection or sepsis in under 1 hour.

Metagenomic next-generation sequencing (mNGS) has emerged as a rapid, culture-independent approach for pathogen detection in sepsis, demonstrating substantially higher detection rates than conventional blood cultures and enabling identification of a broad spectrum of organisms, including bacteria, viruses, fungi, and atypical pathogens. This technology is particularly valuable for detecting organisms difficult to culture, including viruses, Mycobacterium tuberculosis , atypical pathogens, and polymicrobial infections as it is less affected by antimicrobial use., ^,, Prospective studies report pathogen detection in approximately 67% to 92% of septic patients with mNGS, compared with 19% to 48% using blood cultures. ^,,,,,, Turnaround times range from ∼1.4 to 12 hours ^, Both plasma cell-free DNA and blood cell mNGS can be performed, with combined plasma and blood cell mNGS achieving 85.25% sensitivity, which is significantly higher than plasma alone (67.21%) or blood cultures (36.09%). Plasma mNGS is particularly effective for detecting viruses and cell-free bacterial DNA, while blood cell mNGS better identifies viable intracellular bacteria. ^, However, the high analytical sensitivity of mNGS increases the likelihood of detecting contaminant DNA. Consequently, interpretation requires careful clinical correlation and expert judgment due to challenges distinguishing true pathogens from colonizers or contaminants, and the lack of unified diagnostic criteria. ^, Clinical impact studies show promising but mixed results on mortality. A retrospective study found routine blood culture combined with mNGS reduced 28-day mortality from 62.2% to 47.3% ( P =.043), and early mNGS testing (<1 day) was associated with shorter hospital and ICU stays. A large multicenter prospective study demonstrated that mNGS halved 7-day mortality (hazard ratio 0.44, P =.004) but showed no significant difference in 28-day mortality. mNGS prompted antibiotic modifications in 29% to 60% of patients, including de-escalation in approximately 20% of cases. To definitively assess whether combining mNGS-based diagnostics with standard culture-based methods improves clinical outcomes in patients with sepsis and septic shock, DigiSep-Trial a randomized, controlled, interventional, open-label, multicenter trial is ongoing.

Without structured and timely communication, even the most advanced microbiological results will fail to translate into better patient outcomes. Effective implementation of diagnostic stewardship requires close alignment with antimicrobial stewardship and IPC practices, supported by timely and structured communication among these teams. ^, Active communication can be accomplished by phone calls, paging systems, or secure electronic messaging to enable efficient exchange of microbiological and clinical information, facilitating rapid interpretation of results and prompt therapeutic and infection control decisions. This coordinated approach promotes appropriate antimicrobial use, reduces unnecessary exposure, and contributes to improved patient outcomes and more efficient utilization of health care resources. Ultimately, maximizing clinical success while minimizing the emergence of AMR depends on the integrated application of diagnostic stewardship, antimicrobial stewardship, and IPC strategies. ^,,

The optimal antimicrobial therapy in patients with sepsis caused by antimicrobial-resistant bacteria represents one of the greatest challenges in critical care. Although resistant pathogens encompass a broad range of organisms, both Gram-positive and Gram-negative bacteria, and both community-acquired and health care-associated infections, the burden of AMR severe infections is predominantly driven by Gram-negative bacteria causing health care-associated infections, ^, particularly those recognized by WHO as critical and high-priority antimicrobial-resistant bacteria, that is, CRE, CRAB, and carbapenem-resistant Pseudomonas aeruginosa (CRPA). Multicenter studies confirm that these pathogens account for a large proportion of resistant BSI and have been associated with high mortality rates worldwide. ^,,,, Third-generation CRE, either driven by extended-spectrum β-lactamase (ESBL) or AmpC enzymes also represent a common threat to hospitalized patients. ^,

The therapy for these antimicrobial-resistant Gram-negative bacteria has changed over the last decade due to the launching of many novel beta-lactamase-inhibitors combined with beta-lactams, new cephalosporins, such as ceftolozane and the siderophore cephalosporin, cefiderocol. These new antimicrobials now constitute the first-line therapy for carbapenem-resistant, gram-negative bacteria, as addressed in recent guidelines. ^, However, the access to these new drugs is very limited in low and middle-income countries, where the combination of old and less active agents, such as polymyxins and tigecycline, are still the cornerstone therapies for most of these infections.

Although less frequent than antimicrobial-resistant gram-negative bacteria, gram-positive isolates, particularly, MRSA continues to cause severe infections with substantial mortality. ^,, Globally, MRSA has shown the largest increases in resistance-associated mortality over recent decades.

We will present an overview of the optimal antimicrobial therapy for the treatment of patients with sepsis caused by major resistant gram-negative bacteria, including carbapenem-resistant isolates, extended-spectrum β-lactamase-producing and AmpC-producing Enterobacterales , as well as MRSA and vancomycin-resistant enterococci (VRE), which together represent the pathogens of greatest therapeutic concern in severe health care-associated infections.

The antimicrobial therapy for CRE, of which Klebsiella pneumoniae isolates are by far the most common species, depends largely on the underlying resistance mechanism. KPC-producing isolates are typically susceptible to newer β-lactam/β-lactamase inhibitor combinations such as ceftazidime–avibactam, imipenem-relebactam or meropenem–vaborbactam, which are considered preferred agents for severe infections by these pathogens. Although only two small clinical trials comparing these agents against the best available therapy have been conducted, ^, several observational studies have reported lower mortality and toxicity rates with newer β-lactam/β-lactamase inhibitor, especially with ceftazidime-avibactam versus other older antimicrobials. ^, Comparative effectiveness of these antimicrobials is scarce and limited ceftazidime-avibactam versus meropenem–vaborbactam, which showed no difference in clinical effectiveness, but ceftazidime-avibactam may be preferred for its narrow spectrum. However, although still very uncommon, resistance to these agents may be drug-specific and the appropriate choice requires an antimicrobial susceptibility test. ^, No benefit has been shown with the combination of other drugs to ceftazidime-avibactam.

Metallo-β-lactamase-producing CRE isolates, of which NDM-type is the most common, are resistant to these agents, and treatment often relies on aztreonam-based strategies combined with avibactam-containing regimens. ^,, Combinations of aztreonam with either vaborbactam or relebactam have been shown to present inferior in vitro activity against metallo-β-lactamase isolates compared with avibactam. ^, Hence, either aztreonam combined to ceftazidime-avibactam or the recently launched aztreonam-avibactam are the therapy of choice for patients with sepsis by metallo-β-lactamase-producing CRE isolates.

CRAB is particularly difficult to treat and is associated with some of the highest mortality rates among resistant pathogens. ^,, The development of sulbactam–durlobactam represents an important advance, as this combination restores activity against many CRAB isolates by inhibiting class D carbapenemases, the major mechanism of carbapenem resistance in A baumannii In the seminal clinical trial, sulbactam–durlobactam administered with imipenem demonstrated improved outcomes compared with colistin plus imipenem-based therapy and is increasingly viewed as a preferred option when available. Nevertheless, access remains limited in many regions, and alternative regimens, including high-dose sulbactam-containing combinations, polymyxins, or tigecycline derivatives, are still commonly employed.

Pseudomonas aeruginosa poses additional therapeutic challenges because resistance mechanisms are multifactorial, involving β-lactamases, efflux pumps, porin loss, and target modifications, and even new antimicrobial agents may be commonly affected. Ceftolozane–tazobactam has been the main therapeutic option against non-carbapenemase-producing CRPA and other difficult-to-treat P aeruginosa i solates, ^,, which improved clinical effectiveness compared with polymyxins and aminoglycosides. Ceftazidime–avibactam has been shown to present similar clinical effectiveness and is also an adequate option, which may also be active against class A (majorly KPC-type)-carbapenemase-producing isolates. Imipenem-relebactam is also an option for CRPA, but comparative studies with other options are lacking. These agents may allow effective monotherapy when susceptibility is documented. Metallo-β-lactamases are the most common type of carbapenemase in P aeruginosa , but differently to CRE, the activity of aztreonam-avibactam may be affected by other non-enzymatic resistance mechanisms. Although with a promising role in metallo-β-lactamase-producing P aeruginosa , currently there is no susceptibility breakpoint and reported clinical experience is still limited. Polymyxins-based regimes combined with other antimicrobial agents have been used against carbapenemase-producing isolates or where these new drugs are not available. ^,,

Cefiderocol, a siderophore cephalosporin with broad activity against carbapenem-resistant gram-negative organisms, offers an additional option for infections caused by DTR P aeruginosa and CRAB, ^, despite the fact that this is a newer drug that is not recommended as a first-line therapy in severe infections by carbapenem-resistant gram-negative bacteria. In fact, results of clinical trials have raised concerns regarding outcomes against CRAB and other gram-negative bacteria. However, cefiderocol remains an alternative, particularly when other active agents are unavailable or in combination schemes, with overall similar or superior results compared with best available therapy for CRAB infections in observational studies. ^, It may also be useful as part of combination schemes against CRE and CRPA, when first-line agents are not available. ^,

ESBL-producing and Enterobacterales with inducible or derepressed AmpC β-lactamases are among the most frequently encountered resistant pathogens, particularly in BSIs originating from urinary or intra-abdominal sources. Although both mechanisms confer resistance to third-generation cephalosporins, their therapeutic implications differ. For ESBL-producing Enterobacterales , carbapenems remain as first choice therapy in severe infections, ^,, mostly relying on the results of the MERINO trial. In addition, some concerns regarding ESBL inhibitory activity of tazobactam and antimicrobial susceptibility tests in third-generation CRE have also supported the recommendation for avoiding piperacillin-tazobactam in patients with sepsis by ESBL-producing Enterobacterales , despite some favorable results in some observational studies. ^, Cefepime-enmetazobactam is a newly approved combination with promising activity against ESBL-producing Enterobacterales ; however, its role as a carbapenem-sparing strategy in severe infections still needs to be determined.

AmpC-producing Enterobacterales present a distinct challenge because resistance may emerge during therapy through derepression of chromosomal AmpC enzymes. Cefepime is generally favored over third-generation cephalosporins and piperacillin-tazobactam due to greater stability against AmpC hydrolysis and lower risk of selecting resistant subpopulations. Carbapenems remain highly active and are commonly used in critically ill patients, ^,, and has been the preferred agent particularly if the isolate is non-susceptible to one or more third-generation cephalosporins. However, cefepime is a reliable option for infections by these pathogens, with effectiveness compared with meropenem in observational studies.

Although multiple agents demonstrate vitro efficacy against MRSA and VRE, treating invasive infections due to these organisms continues to be challenging. ^, Optimal management of MRSA infections requires both appropriate antimicrobial therapy and effective source control. MRSA is resistant to most β-lactams, except ceftaroline and ceftobiprole, primarily due to alterations in penicillin-binding protein 2 encoded by mec genes. Several newer agents demonstrate in vitro activity against MRSA, including daptomycin, linezolid, tetracycline derivatives (eg, tigecycline and eravacycline), and long acting lipoglycopeptides. However, only vancomycin, daptomycin, and ceftobiprole are approved by the FDA/EMA for treatment of MRSA BSI. ^,

Vancomycin remains widely used as first-line therapy because of its availability, large clinical experience, and relatively low cost, although its use is limited by the need for therapeutic drug monitoring and risk of toxicity. ^, Daptomycin is an effective alternative for bacteremia, and higher doses (8–10 mg/kg/day) are associated with improved clinical success in complicated MRSA BSIs In the ERADICATE trial, ceftobiprole was noninferior to daptomycin for complicated MRSA bacteremia, with comparable microbiological outcomes and no significant difference in treatment success. Drug-related adverse events were more frequent with ceftobiprole, whereas serious adverse events were slightly more common with daptomycin

Routine combination therapy for S aureus bacteremia is not supported by current evidence but may be considered in persistent bacteremia despite adequate therapy and source control Observational studies suggest that combinations such as daptomycin plus ceftaroline may accelerate microbiological clearance and reduce clinical failure ^,, Additional salvage strategies include adding cefazolin, fosfomycin, or ceftobiprole to vancomycin or daptomycin. ^,,

Among long-acting lipoglycopeptides, dalbavancin has the strongest supporting clinical evidence, with studies suggesting a role as step-down therapy in selected patients after bloodstream clearance. ^,, Oritavancin also demonstrates potent in vitro activity and offers simplified dosing due to its prolonged half-life, although clinical experience in invasive infections remains limited ^, Both agents may be considered as consolidation therapy when prolonged intravenous access is undesirable, but neither is recommended for initial treatment of MRSA bacteremia.

Despite not being approved by FDA/EMA for treatment of VRE infections, daptomycin has been recommended for treatment of infections caused by Enterococcus spp. resistant ampicillin and vancomycin, and along with linezolid constitute the main therapeutic options for treatment of VRE BSI. Enterococci exhibit higher daptomycin MICs and intrinsic tolerance, and inoculum effects may further reduce efficacy, necessitating higher daptomycin doses to achieve PK/PD targets. Clinical evidence supports a dose–response relationship, with standard dosing (6 mg/kg/day) associated with worse outcomes and higher doses (≥9–12 mg/kg/day) linked to improved survival and microbiological response. ^,, Outcomes worsen with increasing MIC, and optimal exposure may remain difficult when MICs are greater than or equal to 2 μg/mL. ^, Linezolid provides established clinical efficacy with advantages including reliable oral bioavailability and lower cost. Comparative studies of linezolid and daptomycin show mixed results ^, outcomes appear similar when adequate daptomycin exposure is achieved, with excess mortality being associated with PK/PD target failure. ^, Tedizolid demonstrates potent in vitro activity, but it is not recommended for VRE bacteremia due to bacteriostatic activity, suboptimal pharmacodynamic exposure, and limited clinical evidence. ^,

Teicoplanin may retain activity against vanB VRE, but resistance emergence and inconsistent clinical outcomes limit its therapeutic role. ^, Among lipoglycopeptides, only oritavancin demonstrates activity against vanA -mediated and vanB -mediated VRE, with limited clinical evidence from case reports of successful use in complex hematologic patients. ^,,

Among useful strategies to combat the threat posed by AMR in septic patients, the development of new drugs is paramount. Alas it is quite slow and, despite regulatory and financial incentives for drug companies, the availability of these newer compounds is not as fast as necessary.

In a setting of widespread AMR prevalence, the empirical use of newer, broad spectrum antibiotics is often necessary to reduce the risk of initially inappropriate therapy that has long been associated with increased mortality in several different syndromes. ^,,,,, Effective communication with the microbiology laboratory is paramount. Also, the judicious use of novel diagnostic techniques to minimize the time until appropriate therapy and for early de-escalation seems relevant. For example, the use of rapid molecular testing in BSI is associated with decreased mortality only when used within the context of an ASP. Thus, optimal use of these limited resource is critical, both for maximizing their effectiveness and to reduce antibiotic pressure and, consequently, diminish the emergence of resistant strains as evidence suggests that, at a population level, AM use seems to have thresholds below which the emergence of resistance happens.

As for effectiveness, the first consideration needs to be the diagnosis of the source of infection. This will allow for the selection of drugs with adequate penetration and activity and to assess the need for source control procedures. We should strive to obtain a microbiological diagnosis whenever possible, which has been shown to correlate with mortality.

The optimal use of antimicrobials in critically ill patients should aim for ideal Pk/Pd parameters with adequate loading doses (regardless of ongoing organ dysfunction) and dosing intervals that are consistent with the Pd profile of the drug being used. The concept of “dose modulation” in critically ill patients as originally proposed in 2013 encompasses these ideas. This is well established for drugs with a narrow therapeutic window like aminoglycosides or vancomycin ^, for which serum level monitoring is widely available.

Beta-lactam ABs are time-dependent drugs, commonly used as first-line options for severe infections, and data regarding the use of prolonged or continuous infusions to optimize their efficacy continue growing. The recent Bling III trial, focused specifically on this issue and added significant evidence in favor of these strategies.

In this open-label RCT, ICU patients receiving meropenem or piperacillin/tazobactam were randomized to either continuous (CI; n = 3498) or intermittent (II; n = 3533) AM infusions. The primary end-point of 90 day all-cause mortality was met in 24.9% of patients in the CI group versus 26.8% in the II group (OR 0.91 [95%CI 0.81–1.01]) and the secondary goal of clinical cure of the infection was significantly more frequent in the CI group (55.7 vs 50.0% OR 1.26 [95%CI 1.15–1.38]; P <.001). The incidence of adverse events was low (0.3 vs 0.2% in the CI and II groups).This raises the issue of the variability of serum levels of antimicrobials in critically ill patients that can be at risk of underdosing and treatment failure but also of toxicity. Interestingly, in the Eurobact II cohort, the availability of TDM (for aminoglycosides and vancomycin) or of a clinical pharmacist were associated with decreased mortality. Thus, as is accepted standard of care for drugs such as voriconazole, the way forward is the use of TDM to allow for maximal, safe dosing of antibiotics in patients with sepsis.

Current evidence shows that the solution to reduce the growing rate of AMR in patients with sepsis requires an approach that expands from optimizing the individual bedside ICU care to a multi-disciplinary approach that encompasses the cohesive work of intensivists, infectious disease clinicians, nurses, pharmacists, microbiologists, infection prevention and antibiotic stewardship teams.

We propose that the combination of improved bedside microbiological diagnostics, more efficient PK/PD utilization of therapeutics by simultaneously reducing the utilization of unnecessary antibiotics and increasing patients’ colonization resistance, in combination with effective and sustainable communication among the ICU and multidisciplinary teams is the most efficient strategy to prevent the further dissemination of AMR infections in patients with sepsis.