The main mechanism of resistance is the acquisition of a new penicillin-binding protein, known as PBP2a, which is encoded by the mecA gene.11 In microbiology laboratories, these strains are detected by testing the susceptibility to cefoxitin, by molecular detection of the mecA gene or by detection of PBP2a by immunochromatographic assay. This phenotype entails resistance to all β-lactam agents with the exception of the cephalosporins ceftaroline and ceftobiprole. Historically, there have been differences between the clones involved in community versus health care-associated infections, which differed in virulence and antimicrobial susceptibility, with community clones more frequently exhibiting susceptibility to clindamycin and production of the Panton-Valentine leucocidin toxin. However, in recent years, the introduction of community clones in hospitals and the spread of hospital clones to the community have blurred these differences.11

In Spain, community-acquired MRSA infections in children were first reported in 2006. At present, these strains account for 5%–10% of all S. aureus infections in children, and for as many as 25% of cases of community-acquired pneumonia (CAP).12 Most of the reported cases correspond to skin and soft tissue infections (SSTIs), followed in frequency by pneumonia and bloodstream infections.

Although in the past few years its prevalence has gradually decreased in other European countries, it has remained stable in Spain, with a prevalence of approximately 12% in blood culture isolates in children based on data from the European Centre for Disease Prevention and Control (ECDC).

The COSACO study has found a prevalence of colonization by MRSA in children in Spain of 1.4%, with 4.4% S. aureus isolates exhibiting methicillin resistance.13 It is worth noting that 17% of S. aureus isolates were resistant to clindamycin, a proportion that increased to 26% in the case of MRSA.

The first step should be adequate source control of the infection (Table 3). In case of persistent bacteraemia despite correct treatment, an active search of metastatic foci should be performed.

Glycopeptide antibiotics (vancomycin or teicoplanin) are currently the first-line treatment for MRSA infections. However, their high binding to plasma proteins hinders their diffusion into the tissues. For this reason, it is essential to monitor their plasma concentrations to guarantee values that are adequate for treatment of severe infections. Unfortunately, in recent years there has been evidence of an increase in the minimum inhibitory concentration (MIC) of these antibiotics required for treatment of S. aureus isolates. In addition, there are concerns regarding the increased risk of treatment failure in strains with a vancomycin MIC of 1.5 mg/L or greater,14 so we recommend not using vancomycin if the MIC is 2 mg/L or greater, and in case of a MIC of 1.5 mg/L or greater (determined by Etest), we recommend considering alternatives after consulting with an expert.

Among the available options, clindamycin and linezolid inhibit toxin synthesis, so their use as monotherapy or in combination with other antimicrobials may be useful in case of infection by toxin-producing strains. Linezolid diffuses readily into tissue and allows sequential intravenous-to-oral therapy, although it can cause haematotoxicity and neurotoxicity with prolonged treatment.

Rifampicin exhibits excellent activity against biofilm, and its inclusion in the first-line antibiotherapy strategy is recommended in case of S. aureus infections associated with joint prosthesis.15 Nowadays, cotrimoxazole maintains activity against most MRSA isolates. Furthermore, it is distributed in Spain in the form of oral solution, which is a good option in cases of mild or moderate infections.

In cases of bloodstream infections or endocarditis due to MRSA with a vancomycin MIC of 1.5 mg/L or greater, daptomycin is a suitable alternative. This antimicrobial has bactericidal activity, is approved for use in children aged more than 1 year and is administered as a single dose per day. One of its disadvantages is that it is inhibited by pulmonary surfactant.

Ceftaroline is a fifth-generation cephalosporin approved for use from birth for treatment of SSTIs and CAP, with an excellent safety profile and strong activity against MRSA,16 although the data on its use in children are still scarce.

Enterococcus genus is characterised by an intrinsic resistance to multiple antimicrobials, including nearly all cephalosporins, which makes treatment challenging. Furthermore, they can acquire additional resistance (for instance, to ampicillin or vancomycin),17 further reducing the treatment options.

E. faecalis is the most frequent bactera involved in these infections and it is nearly uniformly susceptible to aminopenicillins, being ampicillin the treatment of choice. E. faecium is second in frequency and is usually resistant to ampicillin and susceptible to vancomycin. E. gallinarum and E. casseliflavus, much less frequent, are intrinsically resistant to vancomycin (although not to teicoplanin), but usually susceptible to aminopenicillins.

Another relevant aspect of Enterococcus spp. is their intrinsic low-level resistance to aminoglycosides due to a deficient transport to the cytoplasm of the bacteria. However, when combined with another antimicrobial that acts on the cell wall (e.g. a β-lactam or glycopeptide), there is an increase in the penetration of the aminoglycoside, with a synergistic bactericidal effect, that is necessary for the treatment of severe infections (such as endocarditis, meningitis and bloodstream infections). Enterococcus spp. can acquire high-level resistance to aminoglycosides, in which case the synergistic effect disappears. The addition of ampicillin and ceftriaxone can also have a synergistic bactericidal effect, although only against E. faecalis.

Resistance to linezolid and daptomycin is currently very rare in Enterococcus spp. and mainly related to outbreaks in health care facilities or previous exposure to them.18

The prevalence of vancomycin resistance in E. faecium isolates from blood cultures in patients aged less than 18 years in the 2011–2012 period in 12 European countries was of 8.3%.19 However, based on data reported by the ECDC, in recent years the prevalence of resistance to vancomycin in Spain in invasive isolates from all age groups has ranged from 0.1% to 0.3% for E. faecalis and from 1.8% to 2.5% for E. faecium, remaining stable.

Whenever the isolate is susceptible to ampicillin, this should be the first-line treatment (Table 4).4 The alternative would be a glycopeptide (vancomycin or teicoplanin). In case of severe infections, such as endocarditis, meningitis or bloodstream infections, we recommend the combination of 2 antibiotics to seek bactericidal synergy.20 The most frequently used combinations are ampicillin with gentamycin or ceftriaxone for infections by E. faecalis and vancomycin with gentamycin for infections by E. faecium.

If the selected combination includes aminoglycosides, in vitro testing must be performed before starting treatment to assess for a potential high-level resistance by determining the MIC. In case of high-level resistance (gentamycin MIC ≥ 500 mg/L or streptomycin MIC ≥ 2000 mg/L), alternative agents must be selected, such as ampicillin + ceftriaxone in case of E. faecalis. If the strain is resistant to ampicillin and glycopeptides, possible alternatives include daptomycin or linezolid.

E. faecalis and E. faecium can develop resistance to vancomycin and teicoplanin mediated by the gene vanA, or to vancomycin not to teicoplanin mediated by the gene vanB. However, there have been reports of resistance to teicoplanin emerging during treatment with this antimicrobial, and therefore its use is not recommended for treatment of these cases. The resistance to vancomycin of E. gallinarum, E. casseliflavus and E. flavescens is intrinsic and is mediated by the gene vanC, which only confers resistance to vancomycin but not to teicoplanin, so the latter could be used for treatment against these species.17

Extended spectrum β-lactamases (ESBLs) are enzymes that can hydrolyse and thus cause resistance or reduced susceptibility to penicillins, oximino- β-lactams (cefotaxime, ceftriaxone, ceftazidime and cefepime) and monobactams (aztreonam), but not to cephamycins (cefoxitin) or carbapenems. They are generally inhibited in vitro by different β-lactamase inhibitors (such as clavulanic acid, tazobactam or avibactam), which helps differentiate them phenotypically from AmpC β-lactamases, which are only inhibited by avibactam.

Currently, the most common ESBLs are CTX-M type. Extended spectrum β-lactamase-producing strains usually contain other genes that confer resistance to aminoglycosides (mainly to gentamycin, and less frequently to amikacin), cotrimoxazole or fluoroquinolones, which further restricts therapeutic options.

In recent years, there has been evidence of a substantial increase in the prevalence of ESBL-producing Enterobacteralesin both hospital and community settings. A recent systematic review on ESBL-producing Enterobacterales isolates from paediatric blood cultures worldwide found an increase in the prevalence from 3.5% to 8% between 1996 and 2013.21

The prevalence in the paediatric age group in Spain exhibits a similar trend to the one observed globally. A study that analysed the epidemiology of community-acquired urinary tract infections (UTIs) in children aged less than 14 years in Spain in 2016 found a prevalence of ESBL-producing Enterobacterales of 3.2%.22 There are few data on infections outside the urinary tract in the paediatric population of Spain.

The most important factors associated to infection by ESBL-producing bacteria are previous hospitalization, recent surgery or recent antibiotic therapy, especially with third-generation cephalosporins, fluoroquinolones or carbapenems, and the presence of chronic disease.23 Furthermore, colonization by these strains could increase the risk of infection by them.24 A relevant phenomenon that is becoming increasingly known is the vertical transmission in mothers colonised by ESBL-producing Enterobacterales.25

The selection of the antibiotic agent is based on the source and severity of the infection and the immune status of the patient (Table 5).8 Traditionally, carbapenems have been the antibiotics of choice for infections with this resistance profile. However, the increase in the prevalence of CR bacteria has underscored the need to identify carbapenem-sparing strategies to ensure a similar effectiveness without promoting the selection of resistant strains.

Although ESBL-producing bacteria often exhibit in vitro susceptibility to piperacillin-tazobactam, there have been reports of treatment failure, especially in cases of high-inoculum infections (pneumonia, undrained abscesses, etc). Data from observational studies have demonstrated the effectiveness of piperacillin-tazobactam, mainly in low-inoculum and non-severe infections and using an extended-infusion dosing strategy.8

The MERINO trial, which assessed mortality at 30 days from initiation of treatment for bloodstream infecions caused by Enterobacterales resistant to third generation cephalosporins (Escherichia coli and Klebsiella pneumoniae) did not demonstrate the non-inferiority of piperacillin-tazobactam compared to meropenem.26 However, several limitations of the study (such as deaths being mainly due to non-infectious causes, administration of piperacillin-tazobactam with a standard infusion protocol, etc) limit the generalizability of the results. A subsequent subanalysis found a similar mortality in case of bloodstream infections from a urinary source and non-severe infections, which was consistent with the findings of previous studies.8

Other antimicrobials (e.g. aminoglycosides, cotrimoxazole, amoxicillin-clavulanate, fosfomycin, cefoxitin or fluoroquinolones) may be an option in case of non-severe UTI and as sequential intravenous-to-oral therapy with a switch to the oral route after achieving control of the source of infection. Ertapenem is also a good alternative to meropenem in some cases of infection by ESBL-producing Enterobacterales, as it decreases antibiotic pressure on P. aeruginosa in addition to the advantages involved in the longer intervals between doses.

AmpC β-lactamases can be chromosomal (intrinsic resistance) or encoded in plasmids (acquired resistance). Enterobacter spp. Citrobacter freundii complex, Serratia spp. Providencia spp, Morganella morganii and the species recently named Klebsiella aerogenes (previously known as Enterobacter aerogenes) all contain the gene for production of AmpC in their chromosomes and have intrinsic resistance to ampicillin, amoxicillin-clavulanate and first and second generation cephalosporins, but remain susceptible in vitro to third and fourth generation cephalosporins and piperacillin. The risk of induction of AmpC β-lactamase production is highest in Enterobacter species.27

In case of exposure to β-lactams, mainly cephalosporins and penicillins, the gene may be derepressed, leading to overexpression of the enzyme and producing the characteristic AmpC phenotype, which confers resistance to all β-lactams (except for cefepime, carbapenems and combinations with novel β-lactamase inhibitors, such as ceftazidime-avibactam). Plasmid AmpC β-lactamases can also be transferred between different Enterobacterales species, including those that do not have the AmpC gene in their chromosomes.

In case of severe infection by AmpC-producing bacteria, due to insufficient evidence, carbapenems are currently considered the most suitable option (Table 6).27 When feasible, targeted treatment with ertapenem could reduce antibiotic pressure on P. aeruginosa. Some non-carbapenem β-lactams, such as cefepime or piperacillin-tazobactam, are considered weak inducers, so they may be an alternative in certain circumstances with the use of high doses delivered through extended perfusion.28 We recommend against the use of piperacillin-tazobactam in high-inoculum infections or if the piperacillin-tazobactam MIC is greater than 8 mg/L, and against the use of cefepime in isolates for which the cefepime MIC is greater than 1 mg/L.29 Cefepime is usually not hydrolysed by AmpC β-lactamases, making it an excellent option. Several alternatives could be considered for treatment of non-severe UTI depending on the results of antimicrobial susceptibility testing (such as aminoglycosides, cotrimoxazole, fluoroquinolones, etc).

The development of carbapenem resistance is one of the most alarming phenomena within antimicrobial resistance, as it involves the loss of activity of a group of drugs with one of the broadest spectrums available. Furthermore, CR is frequently associated with resistance to other groups of antimicrobials (such as aminoglycosides or fluoroquinolones). In case of Enterobacterales, this resistance is usually due to production of carbapenemases, while the resistance of P. aeruginosa in Spain is usually due to non-enzymatic mechanisms (e.g. inactivation of porins and overexpression of efflux pumps).

Recent studies conducted in the United States have found an increase in the prevalence of infections by CR Gram-negative bacilli (GNB) in children, from 0% in 1999–2000 to 0.47% in 2010–2011 in the case of Enterobacterales,30 and from 9.4% in 1999 to 20% in 2012 in the case of P. aeruginosa.31

Currently, the prevalence of CR Enterobacterales in Spain is low, amounting to 4% of all invasive isolates of K. pneumoniae in 2017 (ECDC). Most of these cases occur in adults, and in children they mainly occur in the context of hospital outbreaks. The most prevalent carbapenemase in Spain is currently OXA-48-type, followed by metallo-β-lactamases (MBLs) and KPC β-lactamases.32 Due to the low circulation of these strains in children, their epidemiology has not been accurately established, although there is evidence of a predominance of MBLs of VIM class.33,34

Infections by these microorganisms have a higher mortality,3 and the risk factors for contracting them are similar to those observed in case of ESBL-producing GNB.35

Currently, it is recommended combination therapy with 2 or more antibiotics with demonstrated activity against the isolated microorganism in a majority of cases (Fig. 1). The INCREMENT study, conducted in adults, propose the use of monotherapy in infections with a lower risk of severe disease.36 However, due to several factors, including the scales used or the lesser knowledge of the pharmacokinetics and pharmacodynamics of many antibiotics in children, the extrapolation of these findings to the paediatric population poses a challenge. Table 7 presents the combinations that may be used depending on the source of infection. For non-severe infections monotherapy could be individually considered.

Figure 1.

Treatment of infections by carbapenem-resistant Gram-negative bacilli in children. Adapted from Hsu AJ.40

aPossible antibiotics: aminoglycoside, colistin, fosfomycin, fluoroquinolone or tigecycline. Choose the antimicrobial considering the source of infection and the susceptibility of the isolate (Table 7).

bConsider in case of a meropenem MIC > 4 mg/L, especially in severe infections.

cAztreonam may be used against GNB that produce MBLs (VIM, NDM or IMP) or OXA-48-type carbapenemases, if coproduction of ESBL or AmpC is not detected; if either of the latter are detected, consider use of aztreonam in combination with ceftazidime-avibactam if there is evidence of a synergistic in vitro effect.

dCeftazidime may be used in infection by OXA-48-producing GNB if coproduction of ESBL or AmpC is not detected.

eCeftazidime-avibactam may be used in infections by GNB with carpapenem resistance that is not mediated by MBLs.

fCurrently, ceftazidime-avibactam or ceftolozane-tazobactam, whose use would be off-label in children, should be restricted to severe infections in which the isolate has not exhibited susceptibility to other β-lactams, or in case of infection by a pandrug-resistant microorganism. Data from adults suggest that combined therapy is not neccesary when using these new antibiotics.

gCeftolozane-tazobactam may be used in infections by P. aeruginosa with carbapenem resistance that is not mediated by carbapenemases.

hSome combinations with which there is evidence of in vitro synergistic effect, but little clinical experience: ceftazidime-avibactam + aztreonam, meropenem + ertapenem, meropenem + fosfomycin, ceftazidime-avibactam + fosfomycin.

(0.22MB).

The use of optimised dosage of meropenem (double dose with extended infusion) achieves the desired pharmacodynamic target in isolates with a meropenem MIC of up to 8 mg/L,37 and has exhibited adequate effectiveness in observational studies.38,39 Therefore, we consider it the current treatment of choice, to be given in combination with a second antibiotic with demonstrated in vitro activity.7,40 In the case of isolates with meropenem MIC greater than 8 mg/L, the alternative of choice would be a β-lactam with demonstrated activity in antimicrobial susceptibility testing. If none of them are active against the isolate, other possible options include colistin, aminoglycosides, fosfomycin, fluoroquinolones or tigecycline.

The new β-lactam-β-lactamase inhibitor combinations currently distributed in Spain (ceftazidime-avibactam and ceftolozane-tazobactam) have not been authorised for use in the paediatric population. However, phase II clinical trials have proven its safety and efficacy in children, and ceftazidime-avibactam has been recently approved by the Food and Drug Administration (FDA) for use in children aged more than 3 months. Consequently, its compassionate use could be considered in patients with isolates with meropenem MIC greater than 8 mg/L with demonstrated susceptibility to this antibiotic, especially in cases of severe infection. Ceftazidime-avibactam has demonstrated activity against OXA-48- and KPC-producing bacteria, but not against those that produce MBL carbapenemases, while ceftolozane-tazobactam is not active against any carbapenemase-producing bacteria, but is active against P. aeruginosa with carbapenem resistance that is not mediated by carbapenemases.

In case of extremely resistant isolates, with few treatment options, we recommend that the microbiology laboratory evaluate in vitro the synergistic activity of different combinations (e.g., aztreonam-avibactam, meropenem-ertapenem, meropenem-fosfomycin).7 The combination of aztreonam and avibactam (currently, of aztreonam with ceftazidime-avibactam, as the combination has yet to be marketed), has proven effective against some MBL-producing strains that are resistant to aztreonam. Thus, this could be an option for treatment against pathogens with this type of carbapenemases in the absence of other alternative treatments.7

Although the current evidence on the use of inhaled antibiotics (e.g. colistin or tobramycin) is limited, they could be considered as an adjuvant to systemic antibiotic therapy in lung infections caused by CR GNB.