Diagnostic challenge

Resistance to carbapenems in Enterobacteriaceae is caused by two principal mechanisms^16,17:

Clinical and epidemiological importance of carbapenemase detection:

The resistance to carbapenems mediated by the production of carbapenemases is of high clinical, therapeutic and epidemiological relevance because

Although the CLSI and the EUCAST do not recommend the routine screening for carbapenemases for therapeutic purpose (only to control its dissemination) ^20,21, the antimicrobial resistance committee of the Colombian Association of Infectious Diseases (ACIN), recommends the routine testing for the presence of this enzymes based on the following key aspects:

In the following section, the most frequently used methods for the detection of these enzymes in clinical isolates and clinical samples will be reviewed. The screening of colonized patients will not be addressed.

Phenotypic tests for the detection of carbapenemases

Production of carbapenemases in Enterobacteriaceae should be suspected and confirmed when intermediate susceptibility or resistance to any carbapenem is observed applying the current CLSI breakpoints²⁵.

Phenotypic laboratory tests used for the detection of these enzymes can be divided into “capture” tests and “differentiation” tests. It is recommended to apply them together based on the arguments presented above²⁶.

Molecular tests for the detection of carbapenemases

Currently, there are several molecular platforms based on polymerase chain reaction (PCR) for the detection of carbapenemases in Enterobacteriaceae and other gram- negative glucose non-fermenting bacteria. Among the advantages of these techniques are^28,29:

Expert rules for reporting antimicrobials when CPE are detected:

Class A carbapenemases (detection of KPC or inhibitory enzymes by boronic acid)

Class D carbapenemases (detection of OXA-48 or enzymes not inhibited by boronic acid and EDTA/dipicolinic acid)

Class B carbapenemases (detection of NDM/VIM/IMP or enzymes inhibited by zinc chelators such as EDTA or dipicolinic acid)

Quality control: To guarantee reliable results that translate into appropriate therapeutic decisions, internal and external quality control is definitive to ensure clinically relevant reports. Quality control must be carried out in accordance with the volume of work of the microbiology section, the endemicity of carbapenemases in the institution, country and/or region, and the type of methods performed by the laboratory. Table 4 suggests some of the quality control strains that can be used to evaluate phenotypic and molecular tests. Discrepancies should be confirmed using molecular test.

Note: The modified Hodge test was removed from the CLSI recommendations for the detection of carbapenemases. However, the ACIN Bacterial Resistance Committee considers that this test is still useful in regions of high endemicity of KPC in K. pneumoniae and E. coli. Due to its limitations, a negative result does not exclude the presence of carbapenemases, which must be screened with an alternative method.

Epidemiological challenge

Due to the fast dissemination of CPE, some countries have implemented national strategies for the containment of these carbapenemases. Israel after the implementation of its national policy in 2007 achieved a decrease in the monthly incidence of nosocomial CPE-infections from 55.5 to 11.7 per 100,000 patient/days. The effective strategy was based on dedicated task force of appointed professionals. The task force paid site visits at acute-care hospitals, evaluated infection-control policies and laboratory methods, supervised adherence to the guidelines via daily census reports on carriers and their conditions of isolation, provided daily feedback on performance to hospital directors, and intervened additionally when necessary³⁰.

In the area of infection control, the lack of randomized studies is one of the limitations when it comes to choose the most effective method in a given hospital because recommendations are not based on a high-level evidence.

Different agencies, societies and countries have published guidelines and recommendations for the control of multidrug resistant gram-negative bacteria. In the last five years, three guidelines have been published which qualify the level of evidence using the methodology described by the Grading of Recommendations Assessment, Development and Evaluation Working Group (GRADE). The first guideline was published in 2014 by the European Society of Microbiology and Infectious Diseases (ESCMID), the second in 2017 by the World Health Organization (WHO) and the third, also published in 2017 by the European Prevention Center and Disease Control (ECDC). The ESCMID Guidelines includes recommendations for extended-spectrum β-lactamases producing Escherichia coli and multi-drug resistant Klebsiella pneumoniae but not specifically for carbapenemase producers. The WHO guideline is specific for CRE and includes studies published until 2016. Finally, the ECDC guideline reviewed studies until 2013 and focuses not only on prevention measures but also on the identification of patients at risk for CRE infections. Table 5 describes the specific recommendations for CRE control included in each of these guidelines^31-33.

Screening for CPE carriers

It is recommended to perform CPE carrier screening during outbreaks or high endemicity within a hospital, in all patients that meet one or more of the following risk factors:

Microbiological samples can be obtained from stool, rectum or perianal area or any other site where an infectious process is evidenced.

Regarding the duration of contact precautions, the ECDC and SHEA (Society for Healthcare Epidemiology of America) guidelines recommend that contact precautions should be maintained throughout the hospitalization, particularly in cases of XDR and PDR. In contrast, the WHO did not issue any recommendation in its guideline. Banach et al. recently published an article proposing the elimination of contact precautions 6 months after the initial positive CPE culture, when a patient has two negative control rectal or perirectal CPE cultures taken with one-week interval³⁴.

Therapeutic challenge

Due to the reduced number of randomized clinical trials, the therapeutic approaches for CPE-infections are mainly based on accumulating clinical experience. A wide heterogeneity of most of the published studies (different types of infections, different groups of patients, a variety of treatment regimens, different methods and outcome definitions) can be a serious obstacle in comparing data, which precludes the possibility of a rigorous meta-analysis. Therefore, it is difficult to recommend a single therapeutic approach⁸. It is important to mention that some of the studies included the term CPE, whereas other studies included carbapenem resistant Enterobacteriaceae (CRE). The CDC initially defined CRE as those Enterobacteriaceae, which were non-susceptible to ≥ 1 carbapenem and were resistant to 3rd generation cephalosporins. In their November 2015 update, this definition was revised and CRE are now defined as any Enterobacteriaceae which are resistant (excluding intermediate resistance) to any carbapenem or are documented to produce a carbapenemase. In addition, for those Enterobacteriaceae, which may have intrinsic reduced susceptibility to imipenem such as Proteus mirabilis, resistance to a non-imipenem carbapenem is required. The CDC acknowledges that this definition lacks specificity for CPE, especially in low-prevalence areas³⁵.

In 2017, Trecarichi and Tumbarello performed a literature review of the therapeutic options for CRE and/or CPE infections. The review points out the difficulty of finding comparable studies with a design that allows the identification of the best therapeutic options for patients with these infections³⁶. In this review, the authors analyze 11 available studies. However, there are some fundamental limitations when it comes to compare these studies and draw conclusions that is important to take into account when making a therapeutic decision:

All but one of the studies included K. pneumoniae-infections. All studies were observational cohorts, which implies certain variability in the clinical conditions of patients, as well as differences in the severity of the infection and the type of CPE/CRE. There is also variation in the site of infection in the different studies, with multiple clinical samples (including blood in variable percentages) and a non-systematic use of the MIC since there are different breakpoints of susceptibility or resistance according to the region of the world (CLSI vs. EUCAST). Furthermore, a detailed definition of “monotherapy” and “combination therapy” is not included in most of the articles. In addition, there is a small percentage of reports of Verona integron-encoded metallo-β-lactamases (VIM) alone or in combination with KPC strains making it impossible to extrapolate the clinical or therapeutic impact of each enzyme. To make the comparison even more difficult, in several studies the isolated strains that expressed carbapenemases were carbapenem-susceptible in vitro based on the EUCAST cutoff point. All studies reported different antimicrobial schemes and their association with mortality rates. The different therapeutic approaches were included in the analysis of independent risk factors for mortality in most of the studies; however, there is no adjustment by multivariate analysis for appropriate antibiotic treatment, combined therapy and removal of the source of infection, making the analysis for risk factors and/or mortality associated with CRE/CPE difficult.

Finally, the characteristics of the patient population included in the multivariate models of risk factors for mortality could be divided into two categories: (a) those whose cohort included all patients diagnosed with CPE infections, regardless of whether they received active antimicrobial treatment for a small period of time, and (b) those that included only patients with CPE infections who were treated for at least 48 hours with an effective antibiotic treatment (defined as one or more active drugs due to in vitro sensitivity against CPE/CRE strains). This important difference of cohorts should be considered, especially in the group of patients who did not receive enough treatment time for a definitive adequate therapy for CPE/CRE-infection, as may represent a bias when analyzing the impact of different treatments on mortality rates.

Evidence for combined therapy

Despite the limitations of design in the studies, some studies have shown lower mortality rates using combinations of two or even three antimicrobial agents. In absence of the ideal study, those with the greatest clinical impact will be reviewed.

In a study by Tumbarello et al., the mortality of patients with bacteremia caused by KPC-producing K. pneumoniae who received monotherapy was 54% compared to combined therapy with a mortality rate of 34% at 30 days in 125 patients. Interestingly, patients who received a combination of tigecycline plus colistin and meropenem showed an independent association with higher survival rates at 30 days (Odds Ratio [OR] 0.11) ⁽³⁷. Similarly, Daikos et al. analyzed the clinical outcomes of 205 patients with bacteremia caused by K. pneumoniae producing KPC or VIM, demonstrating an independent association with higher survival rates in patients who received a combined antimicrobial regimen. Combination therapy had the greatest effect on survival in patients with severe sepsis, septic shock, and rapidly fatal underlying diseases³⁸. Qureshi et al. also evaluated the results of 41 patients with bacteremia caused by KPC-producing K. pneumoniae, finding that combined therapy was independently associated with survival (OR 0.07), and a significantly lower 28- day mortality rate (13.3%) in comparison with monotherapy (57.8%)³⁹. In a studied executed by Falcone et al. which involved 141 patients in the ICU with KPC-producing K. pneumoniae infections and septic shock, treatment with colistin (Hazard ratio [HR] 0.21) and the use of ≥ 2 active in vitro antibiotics as definitive therapy, was associated with favorable results (HR 0.08) in terms of mortality rates⁴⁰. And finally, a study by Trecarichi et al. in 149 patients with hematologic malignancies and K. pneumoniae bacteremia who had received ≥48 hours of adequate antibiotic therapy, demonstrated that combined therapy was independently associated with higher survival (HR 0.32) ⁽⁴¹.

Nevertheless, there are several publications, although in a lesser amount, which have not found a survival advantage by giving combination therapy. Some of these published studies are:

Gomez-Simmonds et al. evaluated the clinical outcomes of 141 patients with bacteremia caused by carbapenem resistant K. pneumoniae according to the number of active in vitro agents received and whether an extended-spectrum β-lactam antibiotic had been administered previously. There was no association between the individual treatment characteristics and clinical outcomes, regardless of the use of monotherapy or combined therapy or a β-lactam. Interestingly, different schemes of meropenem were used: 500 mg every 6 hours, 2 g every 8 hours, 2 g in extended infusion over 3 hours every 8 hours, but no stratification was done for each of the treatment schemes⁴². Likewise, no association was found between combined therapy and survival in their cohort of 118 patients with KPC-producing Enterobacteriaceae infections (of which 78 were bacteremia) by De Oliveira et al. The use of polymyxins in this study was an independent predictor of mortality. However, one of the biases of this study was that during the first 10 days of treatment for KPC-producing Enterobacteriaceae, 47 patients (40%) had 51 infections due to other etiological agents such as other bacteria (77%), fungi (21%), or virus (2%)⁴³.

Regarding therapy and the different antibiotic options, there is more evidence for KPC producing Enterobacteriaceae than for other bacteria, enzyme types or therapies. The antibiotics with the best evidence will be discussed below.

Evidence for carbapenems in CPE infections

Some studies showed a benefit of meropenem in combined therapy by an association of lower mortality rates; therefore, carbapenems have been one of the most important pillars of combination therapy in CPE. In the following section, we will discuss the literature that has evidence in favor of carbapenems and their association with the minimal inhibitory concentration (MIC) of the antibiotic.

In an article published in 2011, Daikos and Markogiannakis reviewed the clinical data of the treatment and the results of patients infected with carbapenemase producing K. pneumoniae, reporting 44 patients who had received carbapenems as monotherapy. The analysis of the clinical result according to the MIC of carbapenem showed a survival rate of 29% with a MIC of > 8 mg/L, 60% for a MIC of 8 mg/L, and 69% for a MIC of 4 mg/L or less. Notably, the mortality rates in the group of a MIC of 4 mg/L or less are similar to those reported for K. pneumoniae infections that do not produce carbapenemase, or even CPE infections that have received appropriate combined therapy with a carbapenem. The authors concluded that carbapenems could be a therapeutic option for CPE when the MIC is ≤ 4 mg/L if it is administered at high doses and prolonged infusion and in combination with another active antibiotic⁴⁴.

Other studies have also investigated the correlation between treatment with a carbapenem, its MIC and the clinical outcomes. In a study by Tumbarello et al., which included 36 patients with bacteremia caused by carbapenemase-producing K. pneumoniae, treated with a combined therapy with meropenem, the patients were stratified according to the MIC: in K. pneumoniae with a meropenem MIC of ≤ 8 mg/L, all-cause mortality at 30 days was 15.8%, vs 35.2% for patients with strains with a MIC of ≥ 16 mg/L³⁷.

A mortality of 19.3% was observed by Daikos et al. in patients with carbapenem-resistant K. pneumoniae bacteremia treated with a combination that included a carbapenem, when the MIC for carbapenems was ≤ 8 mg/L, in contrast to an increase in mortality of 35.5% in patients who received a carbapenem in combination therapy when the MIC was > 8 mg/L for carbapenems38. A similar effect was observed by Tumbarello et al., describing lower mortality rates at 14 days in patients who received combined therapy that included a carbapenem and its MIC was ≤ 8 mg/L¹⁴. In all these studies, carbapenems were administered at high doses14,37,38 and in the studies by Tumbarello et al. prolonged infusion was included^14,37.

Evidence for polymyxins in CPE infections

In January 2019, the first International Consensus Guidelines for the Optimal Use of the Polymyxins was published ⁴⁵. In this guideline, the question about the treatment of CPE infections using polymyxins in monotherapy or combination therapy is addressed. However, no consensus was obtained due to the heterogeneity in the methodology of the published studies. The authors decided to issue recommendations based on voting, therefore, the level of evidence is low. The recommendations based on the previous methodology are:

Ceftazidime/avibactam as a new therapeutic option

Ceftazidime/avibactam is the combination of the third-generation cephalosporin ceftazidime with a novel β-lactamase inhibitor called avibactam. The initial studies and the clinical conditions for which the drug was approved by the FDA were urinary tract infections, complicated intra-abdominal infections and ventilator-associated pneumonia. However, due to its potent in vitro activity against CPE, especially KPC producing Enterobacteriaceae, and, partial activity against enzymes of the OXA-48 family, in addition to its high stability against other enzymes such as ESBL and AmpC, this antibiotic presents an excellent profile as a therapeutic option for infections caused by KPC producing Enterobacteriaceae. It is important to highlight that this antimicrobial agent has no activity against metallo-enzymes⁴⁶.

Shields et al. reported a series of cases, including 37 patients with CRE treated with ceftazidime/avibactam, observing a 76% survival (28 out of 37 of patients) and a clinical cure of 59% (22 out of 37 patients), although this rate did not reach statistical significance in patients who received monotherapy (58% [15/26]) versus combination therapy (64% [7/11]). However, there was a recurrence of CRE-infections in 23% between patients with initial clinical success. Microbiological failure was diagnosed in ten of 37 patients (27%) and in three of the ten patients (30%) resistance to ceftazidime/avibactam (MIC > 8 mg/L) was detected⁴⁷.

In a study conducted by Temkin et al. that involved 38 patients with infections caused by CRE and in some cases Pseudomonas aeruginosa, 65.8% of the participants received a second antimicrobial agent with in vitro non-resistance of the microorganism. Of these patients, 73.7% had clinical and/or microbiological cure⁴⁸. Carmeli et al. published a study in 16 countries including patients with complicated urinary tract and intra-abdominal infections caused by Enterobacteriaceae or P. aeruginosa resistant to ceftazidime. Patients randomly received ceftazidime/avibactam or the best available treatment chosen by the physician. The results were comparable, with 91% success in both arms⁴⁹.

Finally, van Duin et al. compared the efficacy of ceftazidime/avibactam versus colistin against CRE, finding that in the adjusted analysis for greater probability of response, patients treated with ceftazidime/avibactam had a lower mortality rate (9%) and a 64% greater probability of better clinical results than those treated with colistin (with a mortality rate of 32%)⁵⁰. On the other hand, some of the concerns are the likelihood of selection of resistance with KPC-3 due to lower affinity to avibactam⁵¹, which is why, in severe infections, the possibility of combination therapy is raised, which still needs to be studied.