This guideline provides recommendations for target therapy of infections caused by the most frequent MDR-GNB in the hospital settings, and must be interpreted as general recommendations that may be individualized for some specific patients, and according to the availability of the antimicrobials in each setting.

The guideline does not discuss the different criteria for any of the infection sites addressed in the recommendations. Hence, a careful assessment should be performed to ensure that the microorganism recovered in cultures is indeed the cause of the infection and not a colonizer.

Empiric treatment decisions are beyond the scope of this document, and as general principles the choice of the best empiric therapy should be based on: i) previous organisms identified from the patient and associated antimicrobial susceptibility testing (AST) data; ii) previous antibiotic exposure during current or recent hospitalizations; and iii) local AST patterns for the most likely bacteria according the site of infection, unit of infection and, occasionally, locally identified risk factors for acquisition of a MDR-GNB.

Since 2019, Brazilian laboratories have adopted the recommendations of the Brazilian Committee on Antimicrobial Susceptibility Testing (BrCAST),19 a national committee of the European Committee on Antimicrobial Susceptibility Testing (EUCAST), in accordance with Ordinance nº 64 of 12/11/2018.20 In addition, for antimicrobials to which there is no breakpoint available, we defined the conditions in which they may be considered as having in vitro activity (Table 1).

In 2019, EUCAST modified the definition of “Intermediate” category, meaning “Susceptible, Increased Exposure”, while the symbol “I” was retained. This category actually indicates that the susceptibility breakpoints apply to the highest doses of the antimicrobials. This is particularly important in P. aeruginosa and S. maltophilia, in which the former “S” breakpoints values were just changed by the symbol “I”. It means that for almost all antimicrobials, the “I” should be literally read as “S”, provided that the recommended dosing regimen is followed (see Table 2). A single exception should be made for carbapenems, since the current “I” is applied to the same minimal inhibitory concentration (MIC) values (for example, 4 and 8 mg/L of meropenem) of the pre-2019 “I” category. Hence, carbapenems are the single exception in which there is no evidence supporting the use of these drugs in monotherapy for treatment of carbapenemase-producing GNB infections by isolates categorized as “I”, or even “S”.4–11

BrCAST/EUCAST breakpoints also have antimicrobials categorized as (S), “S in parentheses”, which means that there are no breakpoints for monotherapy use of a given antimicrobial. There are also some antimicrobials without species-specific breakpoints, as presented ahead. In these circumstances MICs of antimicrobial agents were considered based on their potential for in vivo rescue of activity through the optimization of dosing regimens as shown in Table 1, which are generally in accordance to the “EUCAST guidance on When there are no breakpoints in breakpoint tables?” document.21

Carbapenemase detection tests may imply additional costs in the routine practice of clinical laboratories. However, the implementation of rapid tests (phenotypic, immunochromatographic or PCR) for carbapenemases detection may provide a first guide to the choice of empirical therapy while awaiting the results of ASTs.

We recommend performing rapid phenotypic tests, such as Carba NP and Blue Carba.35 Isolates of carbapenem-resistant Enterobacterales and carbapenem-resistant P. aeruginosa that are phenotypically identified as carbapenemase producers by these tests can subsequently be submitted for confirmation of the carbapenemase class, preferably using immunochromatographic assays and/or PCR. Phenotypic tests, such as Carba NP or Blue Carba combined with EDTA or newer beta-lactamase inhibitors, can also be employed. Alternatively, other phenotypic methods like CIM and mCIM may be used.35–37 While these tests generally demonstrate high sensitivity and specificity, they do have limitations. Therefore, understanding the local epidemiology, both in terms of prevalent pathogens and the enzymes of interest, is essential to select the most appropriate testing method.

The suggested dosing regimens for the treatment of MDR-GNB infections are based on the specific characteristics of the targeted MDR-GNBs and are presented in Table 2.

Beta-lactam/beta-lactamase inhibitors (BL/BLIs) combinations, either ceftazidime-avibactam or imipenem-relebactam, are the preferred options for the treatment of systemic infections by class A carbapenemase-producing CRE as shown in Table 3. There is no direct comparison between these two agents in terms of clinical efficacy for CRE. There may be resistance to one of these options and susceptibility to the other.38 Imipenem-relebactam has lower in vitro activity than ceftazidime-avibactam against Enterobacterales of the Morganellaceae family (Morganella spp., Proteus spp. and Providencia spp.); therefore, ceftazidime-avibactam should be preferred for the treatment of species of the Morganellaceae family.19

The panel suggests that ceftazidime-avibactam may be preferred due to its narrower antimicrobial spectrum. Clinical studies could not find benefit in adding a second agent to ceftazidime-avibactam in the treatment of KPC-producing CRE.39 Although there is no comparative study with imipenem-relebactam, there is a priori no reason to expect that the addition of a second antimicrobial will improve clinical efficacy of imipenem-relebactam either. Therefore, monotherapy with one of these BL/BLIs combinations is recommended.

If there is susceptibility to other non-beta-lactam antimicrobial, such as fluoroquinolones or TMP-SMX, these drugs may be considered as monotherapy in an individual basis, since there is no clinical studies evaluating the use of these antimicrobials in systemic infections by class A carbapenemase-producing CRE.

Polymyxins in combination with meropenem, tigecycline or aminoglycosides are the preferred alternative regime, when the BL/BLIs are not available. There is no recommendation for the use of any of these agents in monotherapy for systemic infections. The panel suggests that the choice of combination should be based on the susceptibility profile: if the isolates present susceptibility to fluoroquinolones or TMP-SMX, or susceptibility increasing exposure (“I”) to meropenem, one of these agents should be preferentially used in combination with a polymyxin. If the isolate does not present Susceptibility (S or I) to any other agent, the preferred combination is the agent that presents the possibility of PK/PD rescue of activity as presented in Table 1. If more than one agent presents the possibility of rescue of activity through PK/PD optimization, the choice should rely on the availability of the drug, the site of infection and potential toxicities of each option. Although some isolates may present susceptibility within brackets to aminoglycosides, the panel recommends their use in combination with polymyxins in systemic infections should be regarded as the last option due to their higher potential for nephrotoxicity. However, this decision should be individualized, and aminoglycosides may be considered as a combination option, particularly when MIC data for other antimicrobials are unavailable.

The recommendations for the treatment of APN and cLUTI is similar to those for RTI, SSTI, primary BSI and IAI with the difference that other non-beta-lactam antimicrobial, such as fluoroquinolones or TMP-SMX can be used in monotherapy as preferred agents if there is in vitro susceptibility to these drugs. In addition, as alternative agents, aminoglycosides or polymyxins (preferably, colistin) may be considered in monotherapy for complicated but non-severe cystitis, when the potential for nephrotoxicity is considered acceptable. For more severe cLUTI or APN, the panel suggests that these drugs be used in combination with a second antimicrobial agent preferably categorized as S, I or (S), in that order, or with the possibility of rescuing the in vivo action, such as meropenem. Due to its low concentration in the urinary tract, tigecycline should be regarded as the last option for composing a combination scheme.

Ceftazidime-avibactam in combination with aztreonam is the preferred option for the treatment of systemic infections by class B carbapenemase-producing CRE.40 Although in theory aztreonam could be combined with other novel BL/BLI, in vitro studies have shown overall superior activity of aztreonam combined with avibactam compared to relebactam and vaborbactam against class B carbapenemase-producing CRE.41–44 In addition, to date, no clinical study reported the use of aztreonam in combination with imipenem-relebactam for the treatment of CRE.

Notably, in some isolates aztreonam alone may present in vitro activity. In these cases, the panel recommends that aztreonam may be the preferential agent for combination therapy with a polymyxin. As monotherapy, aztreonam may only be considered for stable patients with no severe infections.

The use of other non-BL/BLIs antimicrobials should be considered in the same manner as for class A carbapenemase-producing CRE.

The recommendations for the treatment of APN and cLUTI are similar to those for class A carbapenemase-producing CRE, with the exception that ceftazidime-avibactam in combination with aztreonam should be used instead of ceftazidime-avibactam or imipenem-relebactam-alone. If susceptible, the panel suggests that aztreonam in monotherapy can be used for cLUTI, and for APN in selected stable patients. It may also be the preferred agent for combination therapy for APN or cUTIs in more severe infections.

TMP-SMX or a fluoroquinolone are the preferred agents for the treatment of uLUTI regardless of the mechanisms of resistance to carbapenems, if there is in vitro susceptibility. Fosfomycin and nitrofurantoin are antimicrobials that can be used in uLUTI caused by carbapenem-resistant E. coli. Aminoglycosides may be used in monotherapy for uLUTI by CRE, regardless of the mechanisms of resistance to carbapenems, if there is in vitro susceptibility. Aztreonam in monotherapy may be used for class B carbapenemase-producing CRE.

The recommendations for treating infections caused by CRE producers of OXA-48 or its variants (class D carbapenemases) are generally the same as for class A enzymes across all body sites, with the exception that imipenem-relebactam is not an option. When in vitro susceptibility is confirmed, ceftazidime-avibactam remains the only recently developed BL/BLIs with reliable activity against these pathogens. This is largely because ceftazidime is not efficiently hydrolyzed by OXA-48.4–9

The recommendations for the treatment of non-carbapenemase-producing CRE are the same as those for class A carbapenemase-producing CRE.

The panel recommends the use of a combination of two antimicrobials with in vitro activity (Table 4). Considering the BrCAST/EUCAST breakpoints for CRAB, only ciprofloxacin and TMP-SMX could be a combination of two antimicrobials with the “S” phenotype. However, this profile is rarely present in CRAB isolates, and clinical studies supporting their use in systemic infections by CRAB are scanty. Therefore, the combination regimen will necessarily include antimicrobials with susceptibility within brackets (for definitions, see section 1.2 of this guideline) and/or others considered as having in vitro activity as presented in Table 1. .

There is neither evidence indicating superiority of any specific combination over another, nor of polymyxins-containing schemes over schemes not including these drugs. Some preference for polymyxin-containing combinations is solely based on the susceptibility profile of CRAB isolates, which usually present MIC ≤2 mg/L for polymyxin B or colistin. However, the combination of ampicillin-sulbactam plus tigecycline is an option if MICs are ≤ 8/4 and ≤ 0.5 mg/L, respectively, and is the preferred regimen in the presence of resistance to polymyxins. Given the uncertainty regarding superiority of any regimen for CRAB, therapies should be chosen on an individual basis.

Ampicillin-sulbactam is one of the preferred drugs to be used in combination with a polymyxin due to its relatively low toxicity profile and availability in most hospitals. High dose tigecycline may be a preferred drug for combination with a polymyxin, if the MIC of tigecycline is ≤ 0.5 mg/L. However, either ampicillin-sulbactam or tigecycline MIC of ≤ 8/4 mg/L and ≤ 0.5 mg/L, respectively, are uncommon in CRAB isolates.

If no additional drug to combine with a polymyxin shows in vitro activity based on criteria presented in these guidelines (Table 1), the panel recommends that a drug with potential in vivo activity based on PK/PD profile should be added, with preference for those with the lowest MIC, i.e., ampicillin-sulbactam of 16/8 mg/L or tigecycline of 1 mg/L.

The panel suggests meropenem as an option for combination, if ampicillin-sulbactam and tigecycline presented MICs > 64/32 mg/L and > 2 mg/L, respectively, and the MIC of meropenem is 16 or 32 mg/L. PK/PD studies show that high-dose extended infusion meropenem can achieve free concentration over these values 40 % of dose interval, considered the PK/PD target of meropenem, in a considerable proportion of patients.45–48 In addition, free drug concentrations of meropenem >20 % and <40 % of the time can be achieved in a even higher proportion of patients, when the MIC of meropenem is 16 or 32 mg/L, and this PK/PD target was associated to static to 1 log colony-forming unit bacterial killing in an experimental study with neutropenic mouse.49 Although colistin in combination with meropenem were not beneficial over monotherapy in two RCTs, no evaluation according to the meropenem MIC was performed in these studies.50,51 Moreover, in one trial, meropenem was not used in optimized doses.51 Therefore, it should be regarded as a potential drug for combination in the absence of other options if meropenem is within this MIC range.

Combination of a polymyxin plus an aminoglycoside, if susceptible within brackets, is the last option based on the higher potential for acute kidney injury of this combination. Additionally, it is fundamental to state that there is neither PK/PD nor clinical evidence supporting the current breakpoints of aminoglycosides drugs for A. baumannii,52 in a way that the use of these drugs against CRAB is totally experimental. Nonetheless, it may be considered if none of the other antimicrobials present with MICs within the range of potential in vivo activity. The combination of three of the antimicrobial agents (preferebly including the ampicillin-sulbactam plus meropenem) may be considered in the presence of resistance to polymyxins.78

These infections can be treated in monotherapy with a ciprofloxacin, TMP-SMX if there is in vitro susceptibility, and with ampicillin-sulbactam if the MIC is ≤8/4 mg/L. The combination of two antimicrobials is recommended if these conditions are not present. The panel also recommends the use of combination in severely ill or immunocompromised patients. For combination selection, consider the points presented for infections at non-urinary sites. If a polymyxin is chosen, colistin should be preferred considering its higher concentration in urine compared to that of polymyxin B. Owing to its lower concentration in urine, tigecycline should be regarded as the last option.

These infections can be treated in monotherapy with a single drug according to the results of ASTs. Polymyxins and aminoglycosides are acceptable in monotherapy if the isolate presents susceptibility within brackets, but ciprofloxacin, TMP-SMX and ampicillin-sulbactam are preferred options if in vitro susceptibility is observed, considering the higher potential for acute kidney injury of polymyxins and aminoglycosides.

The other considerations presented for the treatment of APN and cLUTI should be taken into account, when deciding the treatment of patients with uLUTI.

Considering that non-carbapenemase-producing CRPA may present a broad variety of antimicrobial susceptibility profile, including in vitro activity to other beta-lactam agents, the choice of the best antimicrobial therapy may depend on this co-resistance profile (Table 5).

For non-carbapenemase-producing CRPA isolates with in vitro activity to beta-lactams (ceftazidime, cefepime, piperacillin-tazobactam, or aztreonam) or fluoroquinolones (ciprofloxacin or levofloxacin), the panel suggests that monotherapy with one of these narrower-spectrum agents for isolates should only be used in patients with less severe infections, who does neither present inadequate source control nor neutropenia. If susceptible, ceftazidime or cefepime are the preferred agents. Piperacillin-tazobactam may also be an option if the susceptibility test is confirmed by disk-diffusion or broth microdilution. Automated systems and gradient diffusion methods have demonstrated unacceptably high rates of very major errors (false susceptibility).53

For patients with severe infections, especially in critically ill patients, or in those with inadequate source control or neutropenia, the panel suggests that either ceftolozane-tazobactam or ceftazidime-avibactam are the preferred agents. The ceftolozane-tazobactam is preferred if the antimicrobial susceptibility test shows susceptibility to both drugs. Imipenem-relebactam may present in vitro activity against some isolates resistant to both ceftazidime-avibactam and ceftolozane-tazobactam due to mutations in AmpC or acquired OXA-10 variants.54,55 Therefore, imipenem-relebactam may be an option for these cases.

In the absence of one of these agents, a polymyxin combined with another agent with in vitro activity, either a beta-lactam or a fluoroquinolone (e.g., ciprofloxacin or levofloxacin) should be used. Aminoglycosides (amikacin or tobramycin) may be used as combination therapy with a polymyxin if the potential for nephrotoxicity is acceptable, but this class is recommended only if there is no other agent with in vitro activity.

The panel recommends that caution should be taken with the use of a beta-lactam or a fluoroquinolone in monotherapy in isolates with resistance to meropenem, because this phenotype indicates the hyperexpression of efflux pumps that also affects these other agents. Hence, development of resistance during therapy may occur, either to increments in expression of efflux pumps or to derepression of AmpC, in the case of beta-lactams. Although these resistance mechanisms may also affect the activity of both ceftolozane-tazobactam and ceftazidime-avibactam, they are affected to a lesser degree. If there is only resistance to imipenem, but not to meropenem, this phenotype indicates that efflux pumps are not hyper-expressed,56 and the use of beta-lactams or fluoroquinolones might be considered as monotherapy for less severe infections in clinically stable patients.

If no antimicrobial other than a polymyxin presents in vitro activity, the panel recommends the use of a polymyxin combined with meropenem due to potential synergism, and the numerically higher survival rate observed in a double-blind randomized clinical trial.51

For patients with APN presenting with sepsis, the recommendations are the same as those for LRT, SSTI, PBSI and IAI.

For APN in stable patients who are not critically ill, as for cLUTIs, either a beta-lactam or a fluoroquinolone in monotherapy may be used as the first-choice therapy. Aminoglycosides in monotherapy are regarded as second line agents, but it is an alternative if the potential for nephrotoxicity is considered acceptable in stable patients with cLUTIs.

If no antimicrobial other than a polymyxin presents in vitro activity, the use of a polymyxin, preferably colistin, in monotherapy may be considered for cLUTIs in stable patients. In APN or in cLUTI with sepsis the panel recommends the combination with meropenem.

For uLUTI monotherapy with an antimicrobial presenting in vitro activity can be used as the first option. A beta-lactam or a fluoroquinolone should be preferred, while aminoglycosides and polymyxins should be regarded as alternative agents, if other options did not present in vitro activity.

Although newer ceftolozane-tazobactam and ceftazidime-avibactam are recognized as first-line options for various CRPA infections, their use in uLUTI should be judicious and reserved for patients with renal dysfunction infected by isolates resistant to other beta-lactams and fluoroquinolones.

In cases of recurrence, the panel recommends that combination therapy with a polymyxin and another agent may be used, and the choice of the second agent should follow the same principles presented for systemic infections.

BL/BLIs combinations, either ceftazidime-avibactam or imipenem-relebactam, are considered the preferred options for treating systemic infections by class A carbapenemase-producing CRPA if in vitro susceptibility is demonstrated. Both agents are recommended as monotherapy, as no study has yet addressed the potential benefit of combining them with other antimicrobial classes. No direct clinical comparison is currently available between these two agents. Susceptibility testing for both BL/BLIs is essential to guide therapy, because there may be resistance to one and susceptibility to another.

If new BL-BLIs are not available, the panel recommends the use of a polymyxin combined with another agent with in vitro activity, either a fluoroquinolone or an aminoglycoside (amikacin or tobramycin). The potential for nephrotoxicity using aminoglycosides and polymyxins should be carefully evaluated.

If no antimicrobial other than a polymyxin presents in vitro activity, the panel recommends the use of a polymyxin combined with meropenem due to potential synergism, and the numerically higher survival rate observed in a double-blind randomized clinical trial.51

If fluoroquinolones (ciprofloxacin or levofloxacin) present in vitro activity, the panel suggests that monotherapy with any of these agents should only be used in patients with less severe infections, who does neither present inadequate source control nor neutropenia. However, this susceptibility phenotype is uncommon among carbapenemase-producing CRPA, and clinical evidence supporting fluoroquinolone monotherapy in this setting is limited.

For the treatment of APN and cLUTI caused by class A carbapenemase-producing CRPA, the recommendations align with those for other systemic infections. Monotherapy with either ceftazidime-avibactam or imipenem-relebactam is preferred. Monotherapy with ciprofloxacin may be selected in case-by-case selection as previously mentioned for systemic infections.

As alternative treatment options, monotherapy with once-daily aminoglycosides (amikacin or tobramycin) or polymyxins (colistin preferentially) may be considered for cLUTI in stable non-neutropenic patients.

The panel recommends that a polymyxin combined with ceftazidime-avibactam plus aztreonam is the preferred regimen for severe systemic infections, particularly in critically ill or neutropenic patients.

Although the combination of ceftazidime-avibactam plus aztreonam may demonstrate in vitro activity against class B carbapenemase-producing CRPA, owing to the current paucity of clinical studies evaluating these combinations for P. aeruginosa,57 the panel suggests the use of these drugs without the combination of a polymyxin should be conditioned to the following situations: i) In less severe infections and stable patients if MIC testing for aztreonam-avibactam of ≤4/4 mg/L, based on current breakpoint for Enterobacterales; or ii) In less severe infections and stable patients if MIC testing for aztreonam-avibactam of 8/4 mg/L, and exceptionally of 16/4 mg/L, based on current aztreonam alone breakpoints, in patients who have failed to the treatment with or is unable tolerate a polymyxin. There is no study correlating in vitro synergy tests neither with lower MICs of aztreonam-avibactam nor with clinical outcomes. Therefore, the panel does not recommend that a synergy testing may substitute the MIC determination for the evaluation of the potential in vitro activity. In vitro activity of aztreonam-avibactam may be compromised in the presence of penicillin-binding protein 3 (PBP3) mutations, or depending on the specific beta-lactamases produced by the CRPA isolate.58,59 Therefore, in the absence of MIC determination for aztreonam-avibactam, the use of ceftazidime-avibactam plus aztreonam without a polymyxin, or occasionally of another potential drug showing in vitro activity (such as a fluoroquinolone or an aminoglycoside) should only be considered for non-severe infections in patients whose the anticipated consequences of a therapeutic failure to the initial scheme are tolerable.

Notably, in some isolates aztreonam alone may present in vitro activity. In these cases, the panel suggests that aztreonam may be the preferential agent for combination therapy with a polymyxin. As monotherapy, aztreonam may only be considered for stable patients with no severe infections.

If ceftazidime-avibactam plus aztreonam are not available, the panel recommends the use of a polymyxin combined with another agent with in vitro activity, either a fluoroquinolone or an aminoglycoside (amikacin or tobramycin). The potential for nephrotoxicity using aminoglycosides and polymyxins should be carefully evaluated. If no antimicrobial other than a polymyxin presents in vitro activity, the panel recommends the use of a polymyxin combined with meropenem for the same reason discussed for other CRPA mechanisms.51

If fluoroquinolones (ciprofloxacin or levofloxacin) present in vitro activity, the panel suggests that monotherapy with any of these agents should only be used in patients with less severe infections, who does neither present inadequate source control nor neutropenia. However, as for CRPA with other carbapenem resistance mechanisms, this susceptibility phenotype is uncommon and clinical evidence supporting fluoroquinolone monotherapy in this setting is limited.

For patients with APN presenting with sepsis, the recommendations are the same as those for LRT, SSTI, PBSI and IAI.

For APN in stable patients who are not critically ill, as for cLUTIs, the panel suggests that either aztreonam or a fluoroquinolone in monotherapy may be used as the first-choice therapy. For these patients, ceftazidime-avibactam plus aztreonam without the addition of another agent may be an option even without MIC testing for aztreonam-avibactam, if the isolate is resistant to fluoroquinolones and aztreonam alone. However, the addition of a polymyxin (preferably, colistin) or an aminoglycoside may be a better alternative in more severe infections.

If no antimicrobial other than a polymyxin presents in vitro activity, and ceftazidime-avibactam plus aztreonam is not available, the use of a polymyxin, preferably colistin, in monotherapy may be considered for cLUTIs in stable patients. In APN or in cLUTI with sepsis the panel recommends the combination of a polymyxin with meropenem.

As alternative treatment options, monotherapy with once-daily aminoglycosides (amikacin or tobramycin) or polymyxins (colistin preferentially) may be considered for cLUTI in stable non-neutropenic patients.

For uLUTI monotherapy with an antimicrobial presenting in vitro activity can be used as the first option. The panel recommends that aztreonam (for class B carbapenemase-producing) or a fluoroquinolone (regardless of the mechanism) should be preferred, while aminoglycosides and polymyxins should be regarded as alternative agents, if other options did not present in vitro activity.

Although ceftazidime-avibactam alone or with aztreonam are recognized as first-line options for various CRPA infections caused by isolates producers of class A or Class B carbapenemases, respectively, their use in uLUTI should be judicious and reserved for patients with renal dysfunction infected by isolates resistant to aztreonam alone (in class B carbapenemase-producing isolates).

In cases of recurrence, the panel suggests that combination therapy with a polymyxin and another agent may be used, and the choice of the second agent should follow the same principles presented for systemic infections.

Carbapenems are the drugs of choice for the treatment of systemic infections, including LRT, SSTI, PBSI, and IAA, caused by ESBL-E and AmpC-E (Table 6). In critically ill patients or those with hypoalbuminemia, meropenem or imipenem are preferred over ertapenem. Cefepime should be preferred over carbapenems for systemic infections caused by AmpC-E isolates that are susceptible to third-generation cephalosporins. The panel suggests that cefepime may also be occasionally an option in less severe infections caused by AmpC-E isolates resistant to third-generation cephalosporins (Table 6); however, close monitoring of possible selection of fully derepressed AmpC is warranted. The panel emphasizes that cefepime is not an alternative for the treatment of infections caused by ESBL-E.

Fluoroquinolones may be considered alternative agents for the treatment of milder infections in stable patients caused by ESBL-E and AmpC-E, or used as a step-down option for oral therapy in patients showing clinical response to initial treatment. Although, ceftolozane-tazobactam may be effective against ESBL-E infections,60 the panel recommends reserving this drug for the treatment of P. aeruginosa or co-infections with ESBL-E. A clinical trial was designed to better define the role of ceftolozane-tazobactam compared to meropenem for the treatment of BSI caused by third-generation cephalosporin-non-susceptible Enterobacterales or known chromosomal AmpC-producing Enterobacterales. However, this trial was withdrawn. In this way, it is important to note that ceftolozane-tazobactam has not been recommended against AmpC-E.

Although ceftazidime-avibactam and imipenem-relebactam are active against ESBL-E and AmpC-E isolates, these drugs should not be used to treat such bacteria and should instead be reserved for the treatment of carbapenem-resistant Gram-negative organisms.

In the treatment of APN or cLUTI, fluoroquinolones, either ciprofloxacin or levofloxacin, can be used whenever susceptibility is confirmed against ESBL-E and AmpC-E isolates. However, especially in critically ill or hemodynamically unstable patients, meropenem may be preferred. Likewise, cefepime can be used for AmpC-E infections in critically ill patients when isolates are susceptible to third generation cephalosporins. TMP-SMX may be used in cLUTI cases with confirmed susceptibility, and occasionally in non-critical patients with acute pyelonephritis (Table 6).

Aminoglycosides are reserved as alternatives to first-line drugs, even in the treatment of UTIs, whenever the potential for nephrotoxicity is considered acceptable, since patients with UTIs treated with these drugs in monotherapy tended to present higher clinical failure and had significantly higher microbiological failures compared to other drugs, in a meta-analysis of randomized clinical trials.61

It is important to note that ceftazidime may appear susceptible in ESBL-producing isolates, particularly those producing CTX-M enzymes. However, there are currently no clinical data supporting its use in this setting. The panel suggests that ceftazidime may be considered as a carbapenem-sparing option in non-critically ill patients with cUTIs, if other options are unavailable or not feasible for the patient. Ceftazidime may also be used for non-severe cUTI caused by AmpC-E isolates, provided they are susceptible to both third-generation cephalosporins. Ceftazidime is preferred over ceftriaxone due to its urinary excretion profile. However, the panel does not recommend third-generation cephalosporins for patients with UTIs whenever abscesses are present. The panel emphasizes that third-generation cephalosporins are not recommended for AmpC-E isolates causing systemic infections.

The panel considers that piperacillin-tazobactam may be an alternative for non-critically ill, stable patients with urinary tract infections caused by ESBL-E. However, it is not recommended for patients with more severe infections. This combination is also not recommended for the treatment of AmpC-E infections, and alternative therapies should always be sought.62 However, if it is considered in a specific case as a therapeutic option for the treatment of severe ESBL- or AmpC-producing infections, the panel recommends confirming piperacillin-tazobactam MICs by using broth microdilution when this agent is intended to be prescribed for the treatment of severe ESBL- or AmpC-producing infections. False susceptibility to piperacillin-tazobactam has been observed with automated systems, gradient strips, and disk diffusion methods compared to broth microdilution, particularly in ESBL producers that co-produce OXA-1 (a narrow-spectrum oxacillinase) and in AmpC producers. Additionally, a significant increase in piperacillin-tazobactam MICs has been reported when testing high inocula against both ESBL and non-ESBL producers.63,64

For uncomplicated urinary tract infections, fosfomycin and nitrofurantoin can be used for E. coli, and preference should be given to oral agents such as ciprofloxacin, levofloxacin, or TMP-SMX.

Alternative agents are the same as those for cLUTIs. The panel considers that oral beta-lactamase inhibitor combinations, such as amoxicillin-clavulanate, may be used as an option for ambulatorial patients (Table 6).