A 14-year-old male with a history of congenital hepatic fibrosis with severe portal hypertension and recurrent cholangitis presented with a 1 day history of fever, shoulder pain and acute right upper quadrant (RUQ) tenderness. His history was remarkable for recurrent episodes of cholangitis over the prior 2 years, leading to frequent courses of parenteral antibiotics as well as interval biliary stent placement and subsequent cholecystectomy. He was on longstanding fluoroquinolone prophylaxis, and at the time of admission had a marked leukocytosis of 23,000 cells/ml, elevated C-reactive protein of 19.3 mg/dl and an elevated total bilirubin of 3.1 mg/dl. Alanine transferase and aspartame transferase were within normal limits. A computerized tomography (CT) scan of the abdomen/pelvis demonstrated hepatomegaly with multiple cystic regions within the liver, consistent with dilated biliary ducts/bile lakes and the presence of pneumobilia. Piperacillin-tazobactam was started empirically, but antibiotics were escalated to meropenem due to clinical worsening. Blood culture returned positive for extended spectrum β-lactamase (ESBL) producing Escherichia coli, resistant to ceftriaxone, cefepime, ciprofloxacin, gentamicin, and trimethoprim-sulfamethoxazole (TMP-SMX)—but susceptible to meropenem (minimal inhibitory concentration, MIC, 0.25 μg/ml) and ertapenem (MIC 0.25 μg/ml). He received 10 days of meropenem for bacteremia and cholangitis, with rapid interval clinical resolution and was discharged home.

He required re-admission 1 week after completion of antibiotics due to fever, RUQ pain and shoulder pain, consistent with cholangitis. A repeat CT scan demonstrated persistence of biliary lakes, and cultures from a biliary drain placed into the largest collection demonstrated E. coli, now with resistance to meropenem (MIC 4) as well fluoroquinolones. He received a prolonged 4-week course of ceftazidime-avibactam with subsequent clinical resolution, but symptoms of biliary drainage, fever, and right sided abdominal pain recurred within 1 week of completion of antibiotics. He was listed for liver transplant as treatment for recurrent cholangitis.

A 13-month-old female with a history of relapsed acute myeloblastic leukemia (AML) developed new onset left labial erythema, ulceration and fever, 6 days following allogeneic HCT. She had no prior history of bacterial infection and was on voriconazole, levofloxacin, and pentamidine prophylaxis. On workup, she was profoundly neutropenic (absolute neutrophil count of 20 cells/µl) and was started on vancomycin and piperacillin-tazobactam for febrile neutropenia with concern for labial cellulitis. Fevers continued, and she developed progressive erythema extending to the contralateral labia, multiple ulcerations to the perirectal region, and worsening hypotension on day 9 of illness. Piperacillin-tazobactam was discontinued and meropenem started. A MRI of the abdomen/pelvis demonstrated diffuse severe cellulitis of the perineum with bilateral myositis consistent with necrotizing fasciitis. She was taken to the operating room for fasciotomy, and tissue cultures demonstrated carbapenem-resistant Pseudomonas aeruginosa, also resistant to piperacillin/tazobactam, meropenem, imipenem, ceftazidime, aztreonam, cefepime, and fluoroquinolones but susceptible to amikacin and tobramycin.

Extended antimicrobial susceptibility testing (AST) results later demonstrated a high MIC to ceftazidime-avibactam (MIC 16/4, No Interpretation), but susceptibility to colistin (MIC 1) and ceftolozane-tazobactam (MIC 4). She subsequently showed evidence of clinical improvement with neutrophil recovery on day 11 but ultimately required two additional debridement procedures on days 12 and 14 for control of her infection.

While the incidence of MDRGN infections reported among adult SOT and HCT recipients varies by geography and institution, overall rates of infection have steadily increased in recent years, with estimates showing an almost ten-fold rise in incidence between the early 2000's and 2015 (6). ESBL-producing Enterobacterales infections are the most common cause of MDRGN infections among SOT recipients and account for up to 75% of these infections (10–12). The incidence of CRE infection has ranged from 3% to 10% in areas of endemicity and has been best described in abdominal SOT recipients. Adult data also suggests increased mortality among transplant recipients with MDRGN infections, with 30-day mortality ranging from 14% to 26% with ESBL infection and 18%–41% in CRE infection (10–14).

Yet despite the increasing attention on MDR infections among adult transplant recipients, knowledge of the epidemiology of these infections in children remains limited. MDR infections have been increasingly identified in children and adolescents, although data, particularly in immunocompromised children, requires further study (1–4). While the prevalence of CRE infections in US children remains low overall, with the Surveillance Network (TSN) database-USA estimating 0.08% CRE vs. 0.47% ESBL in available isolates from children, rates continue to increase (4). Such infections have been associated with poor outcomes such as ICU admission and death in children (1, 2).

Research directly examining MDRGN infections in pediatric SOT or HCT recipients is scarce, but available data suggests the epidemiology, risk factors, and outcomes in the pediatric population may reflect trends observed among adults, with increased rates of MDRGN infection observed in the pediatric immunocompromised and transplant population. Among pediatric liver transplant recipients, the estimated incidence of MDR in a single-center retrospective study of 118 children in Thailand between 2010 and 2018 found 62.4% (58/93) of culture-proven bacterial infections post-transplant were due to MDRGN isolates, with a predominance of Klebsiella pneumoniae and Escherichia coli (15). These were largely intra-abdominal infections, accounting for 47.9% of infections. Another recent single-center retrospective study of pediatric liver transplant recipients from the United Kingdom found a lower rate, noting 27% of patients had colonization with MDRGN with an infection rate of 16.6% (16). MDR organisms have also been identified as a significant cause of severe sepsis for the pediatric liver transplant population as well, accounting for 47.6% (20/42) of organisms causing bacterial sepsis in a tertiary pediatric intensive care unit in the United States. It should be noted though that MDR organisms in this study were defined by resistance to one or more classes of antibiotics and also included cases of vancomycin resistant Enterococcus (accounting for 9 of the 20 episodes of MDR organisms) (17). Similar to adult studies, ESBL infections accounted for the majority of MDR Gram-negative infections in these pediatric liver transplant recipients, and notably, CRE infection rates were found to range from 3% to 17% (15–17).

Knowledge of the incidence and outcomes of MDRGN infections in children receiving treatment for malignancy or after HCT is also limited. In a multi-national, multicenter retrospective study of 1,291 bloodstream infections in pediatric patients with allogeneic HCT or those treated with chemotherapy, more than 25% of Gram-negative infections were resistant to ceftazidime, cefepime, piperacillin-tazobactam, and ciprofloxacin. CRE represented 9% of infections. This study found that bloodstream infection with MDRGN infection was significantly associated with ICU admission and death (18).

Risk factors for MDR infections in immunocompromised adults have been well described and include prior exposure to antimicrobial drugs, critical illness, prolonged and frequent exposure to health care, residence in long term facilities, pre-transplant colonization, and longer length of stay prior to infection (11). In children, previous exposure to antibiotics, particularly third generation cephalosporins and carbapenems, has been significantly associated with MDRGN bloodstream infection and post-transplant infection in pediatric patients with hematologic malignancy, HCT, and liver transplantation (15, 18). Other noted factors in Phichaphop, et al. included operation duration and length of ICU stay (15).

The molecular and geographic epidemiology of β-lactamase producing organisms, particularly carbapenemases, is an important consideration when approaching diagnostic and empiric treatment strategies. In the United States, the K. pneumoniae carbapenemase (KPC) enzyme is the predominant carbapenemase found in CP-CRE isolates infecting adults and children, while other carbapenemases are infrequently identified (8, 19–21). The class of metallo-β-lactamase (MBL) carbapenemases are unique as they are not inhibited by avibactam, vaborbactam, or relebactam. Given this has a significant impact on the available treatment options, clinicians must utilize caution and have clinical suspicion if carbapenem resistance is detected in isolates from children with epidemiologic links to MBL-endemic regions. Additionally, sporadic domestic acquisition of other MBLs remain possible (8, 22, 23).

Infection due to a MDRGN organism should be suspected in an immunocompromised or transplant recipient patient when there is a lack of clinical response to antimicrobial therapy, risk factors present for possible MDR organism exposure, and/or previous isolation or colonization of such pathogen. Early involvement of pediatric transplant ID specialists in these complex cases is important to assist with additional clinical evaluation, which should be guided by the patient's presentation, signs, and symptoms. Aggressive evaluation should be pursued to identify the source of infection, and relevant imaging studies may be necessary. Sampling with bacterial cultures from blood as well as any other suspected sites of infection should be obtained as clinically feasible to guide therapy. Source control, including drainage of infected fluid collections or removal of infected devices or hardware, are important mainstays of therapy, especially given the limited antimicrobial options for MDR Gram-negative infections.

If a clinical isolate is available, the first step in evaluation of antimicrobial resistance is examining the AST panel results to assess for the possible mechanism(s) of resistance via implementation of interpretive criteria from the Clinical & Laboratory Standards Institute (CLSI) or European Committee on Antimicrobial Susceptibility Testing (EUCAST) (24, 25). Often additional testing may be necessary (such as requesting additional local or send-out susceptibilities to novel agents), and isolates should be evaluated in laboratories experienced with recognition of complex resistance mechanisms.

Ceftriaxone-resistance is used as a common proxy to identify ESBL-producing bacteria. Routine ESBL testing is not performed by most clinical microbiology laboratories as current CLSI and EUCAST breakpoints should detect most ESBL isolates, but testing for ESBL confirmation and characterization may be warranted for infection control or epidemiologic purposes.

For identification of CRE, the current CLSI or EUCAST breakpoints are used to identify nonsusceptibility to imipenem, meropenem, doripenem, or ertapenem. Enterobacterales that are resistant to at least one of the carbapenems listed are considered CRE based on the CDC definition, but this definition could include both CP-CRE and non-CP-CRE as noted previously (9). Both CLSI and EUCAST define Enterobacterales susceptibility to ertapenem and doripenem the same (MICs of ≤0.5 μg/ml and ≤1 μg/ml, respectively), but there is some variation in the breakpoints for Enterobacterales and meropenem/imipenem. CLSI defines susceptibility to meropenem and imipenem with MICs of ≤1 μg/ml, while meropenem and imipenem MICs ≤2 μg/ml are considered susceptible by EUCAST (24, 25).

Although the presence of carbapenemase does not influence the categorization of susceptibility, carbapenemase detection and characterization may be recommended for public health and infection control purposes. Additionally, knowledge of whether a CRE isolate is carbapenemase-producing and the specific carbapenemase is also important to guide clinical decisions (which will be discussed further in the following treatment section). Lab tests are available to identify the presence or absence of a carbapenemase, including the Carba NP assay or the modified carbapenem inactivation method (mCIM). Molecular tests can be used to identify specific carbapenemase genes. It is important to note that “negative” results on molecular carbapenemase testing should not be interpreted as confirmation of the absence of carbapenemase (as other carbapenemase genes may be present that are not included in the assay used) or carbapenem susceptibility (as the isolate may be a non-CP CRE). The CLSI does recommend use of the Carba NP assay, mCIM, and/or a molecular assay when Enterobacterales isolates have an imipenem or meropenem MIC of 2–4 μg/ml or ertapenem MIC of 2 μg/ml.

The recommended diagnostic approach for identification of ESBL and carbapenemases are the focus of this review and are summarized in Table 1. A few other notable MDRGN organisms include Pseudomonas aeruginosa with difficult-to-treat resistance (DTR-P. aeruginosa; defined as non-susceptibility to piperacillin-tazobactam, ceftazidime, cefepime, aztreonam, meropenem, imipenem-cilastatin, ciprofloxacin, and levofloxacin; illustrated in Case 2), carbapenem-resistant Acinetobacter baumannii (CRAB), and Stenotrophomonas maltophilia. CLSI and EUCAST interpretative criteria for antimicrobial susceptibilities are also utilized for these organisms. It is useful to note that carbapenemase production is a rare cause of carbapenem resistance in P. aeruginosa isolates in the US and generally carbapenemase testing of clinical isolates of P. aeruginosa is not as critical as for CRE clinical isolates in US hospitals. That said, carbapenemases have been identified in up to one-fifth of carbapenem-resistant P. aeruginosa in other regions of the world and current IDSA guidance encourages performing additional susceptibility testing for novel β-lactam agents (7).

Selection of an appropriate empiric antibiotic regimen is essential to reduce morbidity and mortality risk following transplantation, with inappropriate empiric antimicrobial regimens associated with up to a three-fold increased mortality risk in adult HCT recipients (27). Factors that should be considered in choosing an empiric antibiotic regimen include local resistance patterns (particularly ESBL and regional carbapenemases prevalence), prior patient history of infection or colonization, recent antimicrobial exposure, and severity of illness. Empiric antimicrobial regimens following transplantation should demonstrate robust activity against Pseudomonas aeruginosa. Carbapenem agents, such as meropenem or imipenem, should be considered in settings of high local ESBL prevalence, known prior history of ESBL infection or colonization, and/or critical illness. The empiric use of novel BLBLIs (such as ceftazidime-avibactam, ceftolozane-tazobactam, imipenem-cilastatin-relebactam) or other novel agents is generally avoided if possible but should be considered in the setting of known or suspected risk factors for CRE.

For bloodstream and other serious non-urinary tract infections caused by ESBL-E, treatment with carbapenem therapy should be utilized, with meropenem and imipenem preferred over ertapenem in cases of critical illness. Newer BLBLIs and cefiderocol are active against ESBL-producing organisms but are not utilized as first line if other options are available to preferentially reserve these agents for carbapenem resistant organisms.

The results of carbapenem susceptibility testing should guide treatment selection, although antibiotic choice should generally not be stratified based on the presence of carbapenemase production (CP-CRE vs. non-CP-CRE). To start, carbapenem susceptibility testing can initially provide guidance on whether an extended infusion carbapenem can be used vs. a newer BLBLI (22). For isolates that exhibit susceptibility to meropenem and imipenem but are not susceptible to ertapenem, the use of extended-infusion meropenem (or imipenem-cilastatin) is suggested, assuming no carbapenemase has been identified. Extended-infusion β-lactam (typically over 3 h) optimizes killing time and increases the likelihood of achieving target drug levels. This approach is often used in settings of high inoculum or refractory infection as well.

If the CRE isolate has a meropenem MIC ≥4 µg/ml (such as in Case 1), the use of a novel BLBLI (ceftazidime-avibactam, meropenem-vaborbactam, or imipenem-cilastatin-relebactam) is preferred. If specific carbapenemase testing is available and identifies a specific carbapenemase, this can further inform the selection of the optimal BLBLI given differential activity based on carbapenemase type. For example, while these three agents all have a high likelihood of activity against KPC-producing organisms as well as non-CP-CRE, meropenem-vaborbactam and imipenem-cilastatin-relebactam do not retain activity against OXA-48-like carbapenemases. MBL-producing organisms present further challenges in treatment as neither ceftazidime-avibactam, meropenem-vaborbactam, nor imipenem-cilastatin-relebactam have activity.

For patients with suspected or confirmed infection due to MBL-producing organisms or with previous clinical or surveillance cultures for MBL, preferred treatment options include the combination of ceftazidime-avibactam plus aztreonam (used in conjunction to create “aztreonam-avibactam”) or cefiderocol monotherapy. MBLs, such as NDM, hydrolyze penicillins, cephalosporins, and carbapenems—but not aztreonam. Aztreonam is active against MBLs but may not be effective as monotherapy due to co-production of ESBLs, AmpC β-lactamases, or other carbapenemases such as KPC or OXA-48-like. Avibactam does not have an effect on MBL itself but can “protect” aztreonam from hydrolysis due to ability to inhibit these β-lactamases co-produced by most MBL isolates. Of note, avibactam restores activity of aztreonam in MBL-producing Enterobacterales, but not MBL-producing P. aeruginosa due to additional non-β-lactam mechanisms of resistance in P. aeruginosa (28). Specific synergy antimicrobial susceptibility testing of this combination therapy is available in very limited clinical labs, but no specific lab method is endorsed by CLSI. There is clinical experience with combination of ceftazidime-avibactam plus aztreonam but it is not FDA-approved and requires careful attention to the logistics of co-administration (29, 30). Another MBL-producing CRE option is cefiderocol, which exhibits activity against any of the major carbapenemase enzymes as well as non-CP-CRE. Generally cefiderocol is preserved for cases where other β-lactams cannot be used, such as MBL.

If carbapenemase testing is not available or is negative, novel BLBLIs remain the preferred treatment option, unless there is a possible epidemiologic link/exposure to MBL-producing organism or prior clinical or surveillance cultures with an MBL-producing organism.

Other agents to note for management of CRE include aminoglycosides, polymyxins (such as polymyxin B or colistin), plazomicin, tigecycline, eravacycline, and omadacycline. Aminoglycosides and polymyxins were historically the backbone for treatment of CRE, but the newer BLBLIs are favored now due to clinical success, less complex dosing regimens, and less risk of nephrotoxicity. Aminoglycoside monotherapy is generally reserved for urinary tract infections. Tigecycline and eravacycline are alternative options for CRE infections not involving the bloodstream or urinary tract and their activity is independent of type of carbapenemase.

MDR-P. aeruginosa isolates demonstrating carbapenem resistance but susceptibility to traditional anti-pseudomonal agents can usually be treated effectively with traditional beta-lactams via extended infusion. In cases of critical illness or poor source control, novel agents such as ceftolozane-tazobactam, ceftazidime-avibactam, and imipenem-cilastatin-relebactam are viable alternatives, although this is not the otherwise preferred option to preserve effectiveness of these agents for future use in increasingly antibiotic-resistant infections.

DTR-P. aeruginosa should be treated with ceftolozane-tazobactam, ceftazidime-avibactam, or imipenem-cilastatin-relebactam. Ceftolozane and ceftazidime have a similar structure, but a benefit of ceftolozane is the decreased impact by hydrolysis and porin loss compared to ceftazidime. Ceftolozane has independent anti-pseudomonal activity that does not rely on tazobactam and the combination of ceftolozane-tazobactam is relatively stable to all pseudomonal β-lactam resistance mechanisms. The patient in Case 2 was transitioned to ceftolozane-tazobactam for definitive therapy.

Cefiderocol is an alternative treatment option that is reserved for when there is inactivity, intolerance, or unavailability of the newer BLBLIs. In the case of infection with a DTR-P. aeruginosa isolate that is MBL-producing though, cefiderocol would be the drug of choice.

The possible benefit of combination therapy with multiple agents has been explored, but ongoing use once susceptibility testing is available is controversial. Empiric combination antimicrobial therapy, such as use of β-lactams or BLBLIs with aminoglycosides or possibly polymyxins, is often deployed initially to increase the likelihood that at least one agent administered is active against infection in patients at high risk for MDRGN infections, especially CRE and MDR/DTR-P. aeruginosa. However, once a β-lactam agent demonstrates activity, there is no clear established benefit of continued combination therapy over monotherapy in CRE or DTR-P. aeruginosa infections. There is no randomized trial data comparing monotherapy vs. combination therapy, but the current IDSA panel guidance suggests against the use of combination therapy in these cases of CRE or DTR-P. aeruginosa infection given the lack of clear benefit in available studies and risk of antibiotic associated adverse events with multiple antibiotics (7).

In contrast, combination therapy is suggested for infections due to CRAB and MDR Stenotrophomonas maltophilia. While the incidence of CRAB infection remains low among pediatric transplant recipients, the general approach for treatment would include use of high-dose ampicillin-sulbactam (regardless of susceptibility results) in combination with at least one other active agent, given the limited clinical data supporting any single antibiotic. The production of OXA carbapenemases leads to resistance to β-lactams in most CRAB isolates, but these isolates may also produce MBLs and other serine carbapenemases. Sulbactam-durlobactam is a recently FDA-approved agent with activity against CRAB. Durlobactam inhibits class A, C, and D β-lactamase enzymes commonly produced by CRAB, restoring activity of sulbactam. Specific pediatric data on this agent is not available at this time. S. maltophilia infections pose similar challenges to CRAB infections, and treatment selection is hindered by resistance often caused by MBL and other β-lactamases, intrinsic resistance, and accumulation of efflux pumps. These infections are usually addressed with trimethoprim-sulfamethoxazole in combination with another agent. Another possibility is use of the ceftazidime-avibactam and aztreonam combination in cases of critical illness, intolerance, or inactivity of other agents.

A summary of the available antimicrobial agents and general treatment recommendations by organism are also described in Tables 2, 3. Ultimately the final antibiotic regimen will require consideration of susceptibility results, ability to obtain adequate drug levels at site of infection, side effect profile, route of administration, and cost/formulary considerations. Children in need of transplantation may have chronic and prolonged bacterial infections in addition to their underlying organ dysfunction and remain at risk for recurrent and relapsing infections following transplantation (9). As such, in addition to effective antimicrobial regimens, adequate source control, awareness of post-transplant co-morbidities, and consideration of overall host immune defenses are critical in efforts to control bacterial infection following SOT/HCT. Standard recommended antimicrobial durations in SOT and HCT recipients are often confounded by impaired immunologic clearance, high risk of recurrence, and persisting sources of infection, and as such, durations such should be guided by interval clinical improvement, restoration of immune function and demonstrated source control.

The patient in Case 1 required several prolonged courses of antibiotics which would improve his symptoms temporarily, so he was ultimately listed for transplant as treatment for recurrent cholangitis. He underwent successful deceased donor liver transplant 1 month later with subsequent resolution of his infections. The patient in Case 2 underwent successful skin graft procedure on day 21 of illness, and ceftolozane-tazobactam was continued for an additional week post-debridement with subsequent clinical resolution and wound healing.

Patient-specific factors include prolonged hospitalization in the pre- and peri-transplant setting due to poor functional status or complications related to the underlying diagnosis, indwelling medical devices such as central venous catheters and ventricular-assist devices, and repeated antibiotic treatment courses (11). Such factors may lead to selection for resistant organisms and persistent colonization with these organisms, which may drive increasingly broad antibiotic selection in the setting of fevers and other signs of infection. Antimicrobial stewardship interventions to optimize antibiotic use in both the pre- and post-transplant setting are essential, and the lack of antibiotic de-escalation and use of expansive broad-spectrum empiric antibiotics for prolonged durations represent opportunities for improvement in transplant patients (11). In addition, efforts to remove indwelling medical devices as soon as clinically feasible can reduce risk of infection. Improving the functional status of patients, addressing nutrition and growth needs, and timely progression to transplant to alleviate complications associated with a failing organ or attain immunologic recovery are also critical for preventing infections, but many competing factors can influence the ability to achieve these goals.

Hospital-level infection control and prevention practices can ensure that MDR gram-negative infections are not transmitted within hospital settings. The Centers for Disease Control (CDC) developed a toolkit for healthcare facilities on the prevention of transmission of MDR organisms, specifically for CRE (summarized in Table 4), and the Cystic Fibrosis Foundation Infection Prevention and Control Guidelines should be implemented where transplant recipients with cystic fibrosis access care (32, 33). The benefit of routine surveillance of all patients with MDR organisms, including CRE, in healthcare settings or even on transplant floors is unclear. Only a small proportion of patients colonized with MDR organisms go on to develop clinically significant infections, and optimal infection control measures for transplant candidates or recipients known to be colonized with MDR organisms is uncertain. There is limited evidence of support surveillance outside of the outbreak setting, and intestinal decolonization for patients with known ESBL and CRE carriage is not recommended (11).

Targeted screening of known contacts of patients with MDR infections and active surveillance of patients who meet specific risk criteria for MDR infections may be a more useful approach, particularly to identify the need for contact precautions and patient cohorting (33). For transplant candidates and recipients with known colonization with MDR organisms, pre-emptive treatment or ongoing prophylaxis, in the absence of evidence of clinical infection, is not recommended. Such use of prophylaxis has not demonstrated benefit for clearing colonization and raises concern for development of additional resistance through selective pressure.

Transplant candidates with known colonization with or history of MDR infections, including CRE, should not be excluded from transplantation based on this feature alone, with rare exceptions. SOT donors with MDR infection or colonization should also not be excluded, but communication between donor and recipient healthcare facilities is essential. If a donor is bacteremic or if there is evidence of infection of the allograft being transplanted, recipients will require appropriate post-transplant antibiotic treatment based on susceptibility testing of the identified organism.