Professional Documents
Culture Documents
1 Title Page
8 E-mail: david.aguilera@salud.madrid.org
11 E-mail: luis.escosa@salud.madrid.org
14 E-mail: jesaave@yahoo.es
17 E-mail: emilia.cercenado@salud.madrid.org
20 E-mail: fbaqueroartigao@gmail.com
21
28
29
30 Key words: carbapenem-resistant Enterobacterales, antimicrobial resistance,
31 antibacterial agents, children, epidemiology.
32
35
36
37
38
39
40 ABSTRACT
54
55
56 Carbapenem-Resistant Gram-Negative Infections in Children
57 Introduction
58 Carbapenem-resistant organisms (CRO) are a major global public health threat. Specific
59 pediatric reviews and recommendations have been established for the diagnosis and
66 Microbiology
87 Epidemiology
88 In parallel to what has been observed in adult populations, a progressive increase in the
89 incidence of infections due to multidrug-resistant (MDR) microorganisms has occurred in
90 children. Published studies during the last decade reported that the frequency of carbapenem
91 resistance (CR) in children in the USA increased from 0% in 1999–2000 to 0.47 % in 2010–2011
92 among Enterobacterales isolates(6), from 9.4% in 1999 to 20% in 2012 among P.
98 In most settings, colonization and infection by CRO mainly occur in the adult population. An
99 active surveillance system involving 7 USA states detected a total of 599 isolates of CRE
100 occurring in 481 patients during the 2012-2013 period(10). However, only 3 of those cases
101 (0.6%) were under 18 years of age. In fact, with the exception of some CRE highly-endemic
102 countries (e.g., India or Turkey) where these microorganisms are frequently isolated in
103 pediatric units, CRE are mainly found in adult intensive care units (ICUs) and oncology
104 wards(11)(12)(13)(14). Thus, pediatric data reported worldwide (including countries such as
105 UK, USA, Spain or Italy) has mostly been consistent with sporadic spread and outbreaks of CRE
106 infections (15)(16)(17)(18).
117 Infections due to CRO in children are primarily healthcare-associated (17). A recent Greek
118 report on active surveillance of central line-associated BSIs (CLABSIs) in children conducted
119 from 2016 to 2017 including neonatal ICUs (NICUs), pediatric ICUs (PICUs) and oncology wards,
120 described the impressively high prevalence of CRO: 45% CR among Klebsiella spp, 36% in
121 Enterobacter spp and 38% in P. aeruginosa(23). A study from Lake et al assessed the pathogen
122 distribution and antimicrobial resistance among pediatric healthcare-associated infections
123 reported to the US national healthcare safety network(24). It highlighted the notable
124 difference in the prevalence of CRO isolates among different sources and locations. In the case
125 of pathogens from CLABSIs, CR K. pneumoniae prevalence was higher in the oncology ward
126 (3.3%), whereas CR P. aeruginosa was higher in the general ward (22.2%).
147 Mortality
148 CRO infections in the pediatric population have been shown to increase the risk of mortality 6
149 to 11-fold compared to non-CR infections(31)(32), but on the other hand, a number of studies
150 were unable to document worse outcomes in pediatric patients with CRO infections(33).
151 Mortality rates in children with CRO infections range from 8% to 52% (100% in a small case
152 series) (9)(18) (31)(32)(34)(35)(36)(37)(38)(39) (table 2), depending on the source of infection,
153 underlying diseases and age. One study that evaluated 50 healthcare-acquired CRE BSIs in
154 children from a tertiary hospital in India, mostly by NDM-producing K. pneumoniae, found
155 several significant risk factors for mortality including pediatric PICU admission, intubation,
156 inotropic support, respiratory source and failure to clear bacteremia(35). Meropenem MIC >8
157 mg/L for the isolate considerably increased the risk of mortality (OR: 13.9 [CI 95%: 1.5–125.6],
158 p=0.008) in this study. Regarding the difference in mortality between children and adults, a
159 multicenter national study based in Greece showed higher mortality among infections caused
160 by CRO in adults compared to children (34.9% vs 21.1%)(37).
162 Risk factors for infection or colonization by CRO in children mirror those described for adults.
163 In most cases patients suffer from underlying comorbid conditions or were admitted to the
164 PICU or NICU(37)(16)(40). Specific risk factors for colonization or infection with CRO in children
165 include previous antibiotic exposure (mainly broad spectrum and extended duration), use of
166 medical devices (mostly mechanical ventilation), ICU stay, prior surgery and prolonged
167 hospitalization(18)(31)(33)(41)(25)(42)(43)(44)(45) (table 3).
173 Patients from CRO endemic countries (e.g. India, Turkey or Egypt) (11)(12)(13)(14)(25) may
174 introduce CRO into pediatric institutions of non-endemic countries, suggesting the need for
175 specific interventions such as screening for colonization and specific empirical antimicrobial
176 treatment in newly admitted patients from these countries with infectious
177 diseases(36)(40)(42).
183 Treatment
184 Introduction
185 Specific recommendations in antibiotic therapy for infections due to CRO in pediatric patients
186 are based on studies including only adults(3) (table 4). The availability of pediatric
187 pharmacokinetics and safety data is the cornerstone to broaden the use of proposed agents in
188 adults into the pediatric population. Only scarce observational studies, case series and single
189 reports describe the efficacy of these treatments in infected
190 children(17)(18)(32)(34)(35)(36)(38)(40)(26)(51)(52)(53) (table 2). Furthermore, the regulatory
191 approval status of antibiotics in under 18-year-old patients makes the treatment of infections
192 caused by CRO more complicated in children. Antimicrobial treatment has to be individualized
200 Several studies conducted in adults have evaluated the combination of classic antibiotics with
201 different mechanism of action (e.g., polymyxins, aminoglycosides, fosfomycin, carbapenems,
202 etc) for CRO. Although combination therapies have better outcomes in some groups of adults
203 with CRE infections(55)(56), this has not been thoroughly evaluated in children. Additionally,
204 some data from adult studies suggest that combination treatment may not be beneficial in all
205 settings(57). Notably, the lack of enough knowledge regarding the optimal dose of several
206 antibiotics prescribed for the treatment of CRO infections in children raises the concern about
207 suboptimal treatment if only one antibiotic with limited documented experience is prescribed.
208 Furthermore, inappropriate empirical treatment has been associated with higher mortality
209 among CRE BSIs in pediatric patients (58). Therefore, the implementation of better methods to
210 rapidly identify CRE may improve outcomes(59).
211 Over the past few years, the use of combined treatment for CRO infections has increased in
212 children(38). An observational study in India assessing risk factors for mortality in CRE BSIs in
213 children found that mortality rates were significantly lower in the univariate analysis when ≥2
214 effective drugs were used in combination(35). However, given the small number of patients
215 included in the study and the lack of details regarding the monotherapy regimens provided, it
216 is difficult to draw any conclusions.
217 An Italian multicenter study that included 34 children with CRE infections evaluating the
218 influence of combined treatment on survival rate did not find any difference between patients
219 treated with monotherapy versus combination therapy(18). Nevertheless, the heterogeneity
220 of the sample could not exclude if patients with more severe infections may have benefited
221 from combination therapy. Indeed, studies in adults have proven that only high-risk patients
222 can benefit from combined treatment(55)(60). Unlike adult patients, mortality risk indexes
223 (e.g., Pitt or Charlson score in adults) are not broadly used in children, making this approach
224 more difficult to apply.
225 Data from clinical trials and observational studies conducted in adults have demonstrated the
231 Taking into account this knowledge regarding treatment of non-severe infections, and
232 according to the source of the infection, monotherapy could be an option in certain
233 circumstances, while avoiding monotherapy with potentially suboptimal drugs (e.g., tigecycline
234 or colistin). However, due to the limitations of most of the published pediatric studies, such as
235 small sample sizes or selection bias, further studies in pediatric patients are needed.
236 Carbapenems
237 Several studies have evaluated the safety and pharmacokinetics of meropenem in pediatric
238 patients(62)(63). Due to the higher probability of reaching the pharmacodynamic target for
239 isolates with MIC 8 mg/L when meropenem is administered at a higher dose by extended
240 infusion in adults(64), this approach has been extrapolated to pediatric patients with CRO
241 infections(2). Extended infusion and higher dosing (40 mg/kg/8h) strategies have proven to
242 increase the likelihood of obtaining bactericidal targets for meropenem in Monte Carlo
243 simulated pediatric population (from preterm neonates to adolescents)(65)(66).
244 Similarly to clinical data from adults(67), infections with isolates with meropenem MIC >8 mg/L
245 were associated with increased mortality in two pediatric studies(35)(68), and all children with
246 MIC ≥32 mg/L died(35). Mortality in children who received meropenem for an isolate with
247 MIC >8 was higher than when MIC ≤8 mg/L in this study (100% vs 45.5%, p=0.014).
248 Additionally, some concerns have arisen regarding a lower probability of reaching the target
249 for Gram-negative organisms with MICs 4-8 mg/L in critically ill pediatric patients(66).
250 Considering the lack of pediatric indications for many of the new drugs due to the scarce
251 experience in pediatric patients with most of them, along with the good clinical outcomes for
252 isolates with a meropenem MIC 8 mg/L in adults studies, we consider that meropenem at a
253 high dose by extended infusion is still the recommended treatment backbone until more
254 pediatric evidence is attained with new antibiotics. Backbone alternatives may be considered
255 with a meropenem MIC of 4-8 mg/L, especially in critically ill patients.
256 Ceftazidime-avibactam
257 Avibactam, a β-lactamase inhibitor, inhibits class A β-lactamases (e.g., KPC), class C β-
268 The safety profile in one phase I and two phase II trials including children from 3 months to 18
269 years was similar to that observed in the adult population(72)(73)(74). Based on the results
270 from these trials, the FDA has recently approved its use for pediatric patients >3 months of age
271 with complicated intraabdominal infections (cIAIs), in conjunction with an anti-anaerobic
272 agent, and also for urinary tract infections (UTIs). Currently, clinical data in the pediatric
273 population for the treatment of CRE infections with ceftacidime-avibactam is limited to a few
274 case reports(32)(75)(76)(77)(78).
279 Ceftolozane-tazobactam
305 Meropenem-vaborbactam
316 Polymyxins
317 Since the advent of a growing number of MDR infections, polymyxins have been studied in
318 different pediatric ages(99)(100). Nephrotoxicity is one of the main concerns(99), with a higher
319 risk among adolescents(101). A huge diversity of doses among studies and resulting
320 recommendations makes it hard to establish the optimal dose for children.
321 Although colistin (polymyxin E) is more broadly used, polymyxin B could offer some
322 advantages(102), such as lower risk of nephrotoxicity. Although both antibiotics may need
323 drug adjustment in patients with renal impairment, since polymyxin B clearance is not affected
328 Most of the observational studies in children described a favorable outcome in >70% of
329 pediatric patients treated with polymyxins for MDR infections(103)(104)(101). Colistin has also
330 shown to be effective and safe against MDR infections in neonates(104).
331 Some studies suggest that traditional lower doses of colistin could be suboptimal(99). This
332 seems especially relevant when treating MDR infections located in tissues where penetration
333 of colistin is low (e.g., lung). In the case of central nervous system infections, other routes for
334 colistin administration (e.g., intraventricular/intrathecal) may be considered(105). In critically
335 ill patients with life-threatening infections, it is recommended to start treatment with a
336 loading dose in order to achieve therapeutic concentrations more quickly(100)(106), although
337 there is no data that supports this recommendation in children.
338 In a pediatric study, patients receiving an antibiotic regimen containing colistin for CRE
339 infections had an increased risk of mortality(18). This together with the adult
340 evidence(55)(107) (86), makes it recommendable to use colistin as an alternative to β-lactams
341 only in combination treatment for infections caused by CRO until enough evidence regarding
342 the optimal dose in children could be ascertained through appropriate clinical trials.
343 Emerging polymyxin resistance is considered a serious public health problem. Polymyxin
344 resistance is mainly mediated by modifications in lipopolysaccharides, which can be acquired
345 chromosomally or as a transferable mechanism (plasmid-mediated mcr gene)(108).
346 Tigecycline
347 Tigecycline has excellent in vitro activity against a significant proportion of pediatric CRO
348 isolates(17)(18)(37)(42)(53)(104)(109)(110), but some concerns exist about poorer outcomes
349 compared to other alternatives(111). Some reasons given have been the pharmacokinetics of
350 the drug, which has a large volume of distribution with low serum levels(112). However, some
351 studies have suggested that these therapeutic failures may be due to tigecycline
352 underdosing(113).
353 Due to a higher risk of mortality in adults when prescribed in monotherapy(114), it is mostly
354 used in combination therapy in children. Alarmingly, mortality among children who received
355 tigecycline for MDR/XDR infections was 86% in BSIs and 24% in non-bacteremic
364 Fosfomycin
365 Fosfomycin retains activity against numerous pediatric CRO isolates(15)(40)(119)(120) and
366 should be considered in the absence of other alternatives based on some experience for the
367 treatment of CRO infections in children(26)(120). Interestingly, fosfomycin has shown
368 synergistic in vitro activity with meropenem against MBL-producing P. aeruginosa(121). All 28
369 neonates from a center in China who were treated with fosfomycin-containing regimens in
370 combination with meropenem for CR-producing K. pneumoniae infections (mediated by NDM-
371 1) survived(26).
372 Due to the concern of developing resistance while on therapy(122), it is recommended to use
373 it in combination therapy(2). Recommended doses for fosfomycin in pediatric patients vary
374 widely. A PK study concluded that the currently recommended doses could be suboptimal in
375 children and neonates, suggesting the need for more frequent dosing intervals and higher
376 dosages(123). Oral fosfomycin has been proposed for MDR Gram-negative uncomplicated
377 lower UTIs due to fosfomycin-susceptible organisms in older children and adolescents(1)(124).
378 Imipenem-cilastatin-relebactam
379 Imipenem-cilastatin-relebactam has recently been approved by the FDA in adults. It has good
380 activity against KPC carbapenemases but poor activity against OXA-48-like carbapenemases
381 and lacks activity against MBLs(125)(126). This combination enhances the activity of imipenem
382 against P. aeruginosa isolates(127). Notably, it restores activity against mutated KPC-3
383 carbapanemase, which increases the MIC of ceftazidime-avibactam(125). The RESTORE-IMI 1
384 trial demonstrated better clinical response and lower nephrotoxicity for imipenem-cilastatin-
385 relebactam compared to imipenem combined with colistin in patients with imipenem-non
386 susceptible bacterial infections(107). A phase I study and a phase II/III study are currently
387 investigating imipenem-cilastatin-relebactam for children <18 years of age(128)(129).
396 Limited experience with fluoroquinolones for CRO infections makes this option less
397 recommendable. Eravacycline is a novel fluorocycline with in vitro activity against
398 carbapenemase-producing Enterobacterales and CR A. baumannii, but not against P.
399 aeruginosa (70)(130).
400 Aztreonam is not efficiently hydrolyzed by MBLs; however, these strains usually coproduce
401 other β-lactamases (i.e., ESBLs, AmpC), making them non-susceptible to this antibiotic. The
402 combination with avibactam frequently restores its activity(131)(132) and is an alternative
403 against MBLs-producing Enterobacterales when meropenem MIC is >8 mg/L due to the lack of
404 therapeutic possibilities. The combination of aztreonam-avibactam is currently undergoing
405 two phase III clinical trials in adults evaluating the treatment of serious Gram-negative
406 bacterial infections(133)(134). Several case reports combining aztreonam with ceftazidime-
407 avibactam has shown promising results(135). However, there is no published experience with
408 this combination in children.
409 Cefiderocol is a siderophore cephalosporin with broad activity against broad spectrum MDR
410 Gram-negative bacteria, including serine carbapenemases (i.e., KPC or OXA-48-like
411 carbapenemases) and MBL-producing Enterobacterales, P. aeruginosa and A.
412 baumannii(136)(137). It has recently been approved by the FDA for the treatment of cUTI in
413 adults. There are some concerns due to higher mortality reported in a clinical trial (CREDIBLE-
414 CR) for BSIs and pneumonia; therefore, caution should be used when it is prescribed for severe
415 infections(138). To date, the company (Shionogi Inc.) has provided information about the
416 compassionate use of cefiderocol in 9 cases of patients under 18-years-old who did not have
417 other options, with six of them surviving(138). Pediatric trials whit cefiderocol have yet to be
418 initiated and no specific pharmacokinetic information is known in this population.
420 Some inhaled antibiotics, mainly tobramycin and colistin, have been used successfully in
429 Most of the data about treatment of CRO infections is focused on CRE and the majority of
430 information from this review may apply to this group of microorganisms. Notably, almost all of
431 the new developed antibiotics are focused on the treatment of infections due to
432 Enterobacterales. However, epidemiological surveillances will be useful to assay potential
433 clonal changes and to improve the development of new drugs, as the emergence of resistance
434 to newer antibiotics will probably occur when these new agents are used more broadly.
436 P. aeruginosa exhibits a huge diversity in its world distribution, with different circulating clones
437 that harbor a high diversity of the mechanisms involved in resistance against carbapenems.
438 This means an even higher challenge for treatment. The prevalence of susceptible in vitro
439 isolates to different drugs shows a broad variability. Ceftolozane-tazobazam has good in vitro
440 and clinical activity against several CR P. aeruginosa isolates not mediated by carbapenemases
441 (e.g., OprD deficiency). Ceftazidime-avibactam is a good alternative for CR P. aeruginosa
442 harboring class A carbapenemases (e.g., GES enzymes). Other newer β-lactam-β-lactamase
443 inhibitors, such as imipenem-cilastatin-relebactam, may play an important role against CR P.
444 aeruginosa infections, but little data is currently published against MDR/XDR P. aeruginosa
445 isolates.
446 For resistant isolates to all β-lactams, classic drugs such as polymyxins or aminoglycosides, and
447 synergistic combinations (142)(143), may be considered. From the agents commented in this
448 review, tigecycline and eravacycline do not exhibit relevant activity against P. aeruginosa.
449 Although high-dose extended infusion of carbapenems combined with a second active
450 antibiotic may be considered for the treatment of CR P. aeruginosa
451 infections(2)(144)(142)(143), less is known compared to CRE, making this option less
452 attractive.
463 Conclusions
464 The rate of CRO infections has greatly increased in the last years, representing a major public
465 health problem. These infections are difficult to treat, leading to high mortality. The evaluation
466 of risk factors for developing a CRO infection could provide individualized empirical broad-
467 spectrum antibiotic therapy, according to local epidemiology. Clinical evidence regarding
468 treatment of CRO infections remains scarce, and it mainly comes from observational studies,
469 which are even more limited in children. New antibiotics may open the door to highly effective
470 treatments, but the delay in conducting pediatric trials is leading to off-label use in this
471 population. Antibiotic stewardship programs remain a key element to preserving current
472 antibiotic activity through a rational approach in antimicrobial treatment.
473
474 ACKNOWLEDGMENTS
475 We thank Cindy McCoig, MD, for kindly reviewing the manuscript.
476
479
480 FUNDING
481 This study did not receive any funding. DAA is funded by the Spanish Ministry of Health
482 – Instituto de Salud Carlos III (ISCIII) and co-funded by the European Union (FEDER) [Contrato
483 Río Hortega CM18/00100].
485
486 REFERENCES
1121 * Antibiotics referred as previous exposure, with variable time criteria among studies.
1122 BSI, bloodstream infections; CR, Carbapenem-resistant; CRE, Carbapenem-resistant Enterobacterales; CS, Carbapenem-sensitive; CSE,
1123 Carbapenem-sensitive Enterobacterales; CVC, Central vascular catheter; ESBL, Extended-spectrum β-lactamase; GNB, Gram-negative bacteria;
1124 LOS, Late-onset neonatal sepsis, NICU, Neonatal intensive care unit; PICU, Pediatric intensive care unit.
1125 Table 4. Proposed treatment for carbapenem-resistant Gram-negative infections in children.
1126 Adapted with permission from Rodríguez-Baño J(3).
Backbone recommended: Meropenem in extended and high dose infusion
Meropenem
Backbone alternatives (consider Ceftazidime-avibactam3, meropenem-
MIC ≤8 mg/L
specially when meropenem MIC 4-8)1,2: vaborbactam4, ceftolozane-tazobactam5,
Susceptible to a
Meropenem aztreonam6, ceftazidime7, ampicillin-sulbactam8
β-lactam Backbone recommended:
MIC >8 mg/L or imipenem-cilastatin-relebactam9
36
1152 Table 5. Recommended drugs according to infection source for carbapenem-resistant Gram-
1153 negative infections in children
Accompanying Alternative accompanying Other alternatives
Source Backbone
drug drug (1)
Fluoroquinolone or
CLABSI/BSIs β-lactam Polymyxin (4) Fosfomycin or aminoglycoside
tigecycline (2)
37
1171 Table 6. Proposed dosing for the most frequently used drugs against carbapenem-resistant
1172 Gram-negative bacilli in children.
Drug Dose Regulatory approval status Comments
Meropenem 40 mg/kg/dose IV over 3 hours q8h (max: 2 EMA: >3 m
g/dose) FDA: All ages (<3 m only for IAI)
Neonates <32 wk gestation and <2 wk old:
same dose q12h
Ertapenem ≥13 y and adults: 1 g q24h EMA and FDA: >3 m Combined ertapenem-
38
Ceftazidime- 2-hour IV infusion q8h: FDA: >3 m Consider 3-hour infusion for
avibactam 1) 6 months to <18 years: 50 mg/kg EMA: not approved <18 y severe infections.
ceftazidime (max. 2 g/dose)/12.5 mg/kg
avibactam (max. 0.5 g/dose)
2) 3 to <6 months: 40 mg/kg ceftazidime/10
mg/kg avibactam
Ceftolozane- 1-hour IV infusion q8h of ceftolozane 20 FDA and EMA: not approved <18 For severe lung infections
tazobactam mg/kg/dose (max. 1 g/dose) and tazobactam y consider ceftolozane 40
39