MRGN: New classification for multidrug-resistant Gram-negative bacteria

New MRGN classification

The first KRINKO (Commission for Hospital Hygiene and Infection Prevention) recommendations on hygiene measures in connection with infections or colonization with multidrug-resistant Gram-negative rods were published in 2012 [1]. Such recommendations had been urgently needed. After all, the spread of carbapenemase-producing Gram-negative bacteria has also been observed in Germany [2] and must be contained at all costs. Infections with such strains of bacteria can be treated only with few antibiotics, and go hand in hand with increased mortality [3]. In connection with the new KRINKO recommendations, a new classification of multi-resistance in Gram-negative pathogens was also introduced.

The MRGN classification is an operationalization of the fuzzy concept of “multidrug resistance” and ultimately determines the implementation of hygiene measures, such as single-room isolation. But single-room isolation is an intervention with adverse effects, ranging from the mental impact on affected patients, poorer nursing and medical care and increased workloads for the medical staff to additional financial costs. The great significance of the MRGN classification becomes readily clear from this: it must be unique and has to be a good reflection of the varying hygienic importance of different constellations of multi-resistance, so that the drastic measure of single-room isolation can be used as specifically as possible.

This classification serves only the purposes of hospital hygiene and has no bearing on the treatment. The KRINKO recommendations are silent on how antibiograms are to be reported. Which breakpoints to apply, as well as whether and how a measured value is to be interpreted for the report are issues dealt with by EUCAST and the National Antibiotic Susceptibility Testing Committee, or NAC (http://www.p-e-g.org/econtext/NAK). It must be emphasized in this context that microbiological findings have at least two different objectives: they are to guide treatment and form the basis for recommendations on hospital hygiene.

A definition of multi-resistance in Gram-negative pathogens is difficult for four reasons.

First, multi-resistance in Gram-negative bacteria plays a significant role in around a dozen different bacterial species. This alone makes simple rules for defining multi-resistance impossible. For example, while resistance to cefotaxime and ertapenem in Pseudomonas aeruginosa and Acinetobacter baumannii represents a normal species property, and is thus of lesser importance in terms of hospital hygiene, the same resistance phenotype in Klebsiella pneumoniae would indicate ESBL and/or carbapenemase, thus necessitating a wide range of hospital-hygiene measures.

Second, the multitude of different resistance mechanisms complicates simple definitions. For example, there are hundreds of different beta-lactamases. Extensive resistances of varying degrees are caused by differences in the hydrolysis spectrum. These can limit the therapeutic options, and therefore be of varying significance in terms of hygiene.

Third, when it comes to finding a multi-resistance definition for the purposes of hospital hygiene, it does not make sense to weight all antibiotics equally. Instead, a definition ought to include those antibiotics that are highly significant to the treatment of severe infections. Resistances in connection with such antibiotics do limit the therapeutic options considerably. As a result, hospital-hygiene measures to prevent the transfer of the corresponding strains are, and must be seen as, justifiable. There should be agreement, for example, that, when dealing with K. pneumoniae strains with resistance to meropenem or also cefotaxime, more stringent hygiene measures should be necessary than in connection with K. pneumoniae strains with resistance to sulfamethoxazole-trimethoprim.

Fourth, it makes sense to differentiate for different degrees of severity of multi-resistance in Gram-negative pathogens so as to control hygiene measures better. Thus, the therapeutic options for K. pneumoniae with carbapenemase are significantly more limited than for a strain with ESBL, which should be reflected in different hygiene measures.

There had been definitions of multi-resistance before the KRINKO recommendations, but they often did not account appropriately for the above-mentioned difficulties. According to the preliminary definition of multi-resistance arrived at by an international experts’ group convened by the CDC and ECDC [4], E. coli with resistance to ampicillin, gentamicin and sulfamethoxazole/trimethoprim would be regarded as multi-resistant as KPC-producing K. pneumoniae, which would be sensitive to tigecycline, colistin and fosfomycin. From an infectiological view, this is questionable. This definition would not be very suitable for hospital-hygiene measures, because it does not weight classes of antibiotics at all. As a result, it cannot be used to identify pathogens that really require enhanced hygiene efforts. For this reason, a new multi-resistance definition has been worked out for the KRINKO recommendations.

• The MRGN classification is based on the following principles: intermediate susceptibility is rated as equal to resistance under the MRGN classification. To improve the readability of this paper, every mention of resistance also implies intermediate susceptibility.

• The classification is based on groups of antibiotics that are especially relevant for therapy: piperacillin as a penicillin derivative, cephalosporin with an extended spectrum, carbapenems and fluoroquinolones. This is to ensure that hospital-hygiene measures are taken particularly in connection with such pathogens as would entail especially severe consequences with respect to treatment if they were transferred to another patient.

• There are not two categories (normal vs. multi-resistant), but three degrees of multi-resistance (normal >3MRGN>4MRGN). This allows for a finer gradation of the resulting hygiene measures. Putting it simply, an isolate would be classified as 3MRGN if three of the four groups of antibiotics mentioned above were to be resistant. Resistance in all four groups of antibiotics would result in a 4MRGN classification. There are indicator substances for each group of antibiotics.

• The terms 3MRGN and 4MRGN are new and not yet assigned. This promotes clarity in communications between the laboratory and the clinic.

• The classification is carried out almost entirely by phenotypic criteria and not according to resistance mechanisms. Thus, no distinction is made as to whether resistance to cephalosporin in Enterobacteriaceae is caused by ESBL or AmpC beta-lactamase. Nor does one distinguish between a mechanism encoded on a plasmid and one encoded chromosomally. But in one point, one deviates from this principle deliberately: the detection of carbapenemases is incorporated directly into the MRGN classification. This is due to the enormous risk that a further spread of carbapenemase-producing strains poses [5]. Moreover, in the case of such strains, an MIC in the resistant range is not always measured for all carbapenems, but these strains would still be of substantial hygienic importance. It follows that every microbiological laboratory must be able to identify carbapenemase-producing strains reliably.

• The MRGN classification is based on the resistance ratings according to the MIC breakpoints. This is important, because in many microbiological laboratories in Germany antibiograms are interpreted, which means that in certain situations an antibiotic measured as sensitive is interpreted as resistant due to certain rules. Such laboratories should use the resistance category before the interpretation to establish the MRGN classification. This approach is to ensure an MRGN classification as uniform as possible, independent of the respective methods of the laboratory with respect to the interpretation of antibiograms. For laboratories that interpret resistances, it may be useful in individual cases to state that in the report in connection with certain constellations in order to make it transparent why no 3MRGN has been assigned even though the antibiogram interpreted and included in the report would require such a classification.

• Exceptions must be taken into account. For example, an increased MIC is normal for imipenem in connection with Proteus spp., Morganella morganii and Providencia spp., and should be ignored for the MRGN classification.

• In footnote 2 of table 5 on p. 1344 of the KRINKO recommendations [1], it is stated that, in neonatology, a sole resistance to third-generation cephalosporins in connection with certain pathogens can entail interdisciplinary considerations about the necessity of a hospital-hygiene intervention. This has to do with the fact that ciprofloxacin is not usually administered to this patient population. In case of a relevant date of birth of the patient and/or hospital department, the laboratory should identify such pathogens in the report, stating that the situation does not warrant 3MRGN as per the definition, but that 3MRGN hygiene measures should be considered for this isolate in this instance.

The MRGN classification is very complex. A correct MRGN classification requires reliable detection of carbapenemases, knowledge of the resistance ratings according to the MIC breakpoints, and, not least, microbiological expertise. It can therefore only be carried out by a microbiology laboratory. The special role of the microbiology lab is also emphasized in the KRINKO recommendations: the diagnostic laboratory should report the classification as 3MRGN or 4MRGN in the findings, and also communicate the confirmation of 4MRGN by phone. It is recommended that text modules be used that contain the following information:

• 3MRGN! (see corresponding KRINKO recommendation)

• 4MRGN! (see corresponding KRINKO recommendation)

• Isolate with specific resistances. In some areas (e.g., neonatology), it should be treated like 3MRGN (e.g., in case of K. pneumoniae with ESBL, but with sensitivity to ciprofloxacin).

Laboratories that interpret antibiograms, and thus report a formally sensitive result as resistant under certain circumstances, can indicate this by means of text modules in order to make it transparent why, for example, the 3MRGN classification was discarded.

The MRGN classification is generally based on table 2 of the KRINKO recommendations. Since a single table can hardly map the complexity of the MRGN classification, the KRINKO recommendations contain further information on pp. 1313–1316 and in table 3 [1]. In addition, after consulting with Robert Koch Institute (RKI), a document with frequently asked questions about MRGN classification was prepared and posted on the website of the National Reference Center (NRZ) for nosocomial Gram-negative pathogens (http://memiserf.medmikro.ruhr-uni-bochum.de/nrz/), which must also be taken into account. The following notes are intended to provide further guidance on the correct MRGN classification:

• Where the section on MRGN classification speaks of A. baumannii, read A. baumannii complex. This means that the procedure is the same for Acinetobacter pittii (formerly Acinetobacter genomospecies 3) and Acinetobacter nosocomialis. Other Acinetobacter spp., however, are not assigned an MRGN classification.

• Only Enterobacteriaceae,P. aeruginosa and species of the A. baumannii complex are assigned an MRGN classification. Accordingly, Pseudomonas spp. except P. aeruginosa,Stenotrophomonas maltophilia,Burkholderia cepacia,Achromobacter xylosoxidans and other nonfermenters are not assigned an MRGN classification. This does not mean that special hospital-hygienic measures may never be indicated in those cases, but these species are not subject to the KRINKO recommendations.

• In table 2 of the KRINKO recommendations, the type of logical link for the lead substances has been left open (“and/or”). With Enterobacteriaceae and the A. baumannii complex it should be used as an OR link; with P. aeruginosa as an AND operation.

• Laboratories are often faced with the question whether the lead substance for acyl ureidopenicillin for classification purposes should be piperacillin or piperacillin/tazobactam, because an announcement of the new MRGN classification still listed piperacillin/tazobactam (Epidemiologisches Bulletin 36/2011). It is true that piperacillin is to be considered according to table 2 of the current KRINKO recommendations. In this case, a deliberate decision was made to deviate from the principle of using antibiotics of special therapeutic relevance for the classification – this was indicated by a lot of good reasons:

  • Piperacillin/tazobactam is in vitro sensitive for many ESBL strains. But such bacterial strains are not diagnosed uniformly in Germany. While some laboratories report it as sensitive according to EUCAST recommendations, others interpret it as resistant. Arguments can be made in favor of either approach. A consistent approach in this matter is not expected any time soon.

  • The resistance testing for piperacillin/tazobactam and thus also the determination of the resistance category is heavily dependent on the methods [6, 7].

  • If one were to apply to the classification the test result for piperacillin/tazobactam according to the in vitro results without any interpretation, only 10–30% of all ESBL-K. pneumoniae and E. coli would be classified as 3MRGN [7]. In view of the fact that the 3MRGN classification requires fluoroquinolone resistance, leaving only approx. 80% of all ESBL-K. pneumoniae and E. coli in this category, and considering that special hygienic measures are taken only in at-risk areas of hospitals in connection with these strains, a further narrowing of the 3MRGN category by including the in vitro result of piperacillin/tazobactam would run counter to the fundamental statement of the recommendations, and would also differ greatly from the recommendations in other European countries.

• It is not always necessary that laboratories actually test all lead substances for the purposes of the MRGN classification.

  • Piperacillin can be considered resistant for the purposes of classification in connection with all Enterobacteriaceae with resistance to cefotaxime or ceftazidime, because beta-lactamases that hydrolyze cephalosporins with an extended spectrum (ESBL or AmpC) are also active against piperacillin. In connection with species of the A. baumannii complex, piperacillin can generally be considered intrinsically resistant.

  • In connection with species of the A. baumannii complex, cefotaxime can be considered intrinsically resistant for the purposes of classification.

• Cefepime may be considered for the MRGN classification with P. aeruginosa. Cefepime is not expressly mentioned in table 2 of the KRINKO recommendations [1], but implicitly as “4th generation cephalosporin”. For a P. aeruginosa isolate that was measured as ceftazidime-intermediary and cefepime-sensitive, therefore, the antibiotic group of 3rd/4th generation cephalosporins would still be considered effective with respect to the MRGN classification. This follows from the fact that the link is established via a logical AND for P. aeruginosa.

• Whether cefepime is considered for Enterobacteriaceae or the A. baumannii complex is irrelevant in practice: in fact, in the case of Enterobacteriaceae, resistance to cefepime with simultaneous sensitivity for cefotaxime and ceftazidime is not expected to occur. In the A. baumannii complex, the group of 3rd/4th generation cephalosporins can be considered ineffective for the purposes of the MRGN classification per se.

• It must be pointed out emphatically that the KRINKO recommendations can only be implemented properly if the laboratory can detect carbapenemase-producing Enterobacteriaceae reliably. According to the KRINKO recommendations (p. 1315, B) [1], Enterobacteriaceae are classified as 4MRGN once carbapenemase has been detected – that is, independently of the antibiogram. Carbapenemase-producing Enterobacteriaceae do not always exhibit MICs in connection with imipenem and meropenem that fall within the intermediary or resistant range. Testing of ertapenem is recommended in order to find carbapenemase with sufficient sensitivity in the case of K. pneumoniae, K. oxytoca, E. coli and Proteus spp. For the MRGN classification, ertapenem was deliberately left out as a lead substance in Table 2, because ertapenem resistance has too low a specificity for the detection of carbapenemase in connection with E. cloacae, E. aerogenes, C. freundii and other Enterobacteriaceae with chromosomal AmpC. What is more, with P. aeruginosa and the A. baumannii complex, it is intrinsically resistant. Therefore, it is all the more important that each laboratory can detect carbapenemases reliably. The website of the NRZ for nosocomial Gram-negative pathogens (http://memiserf.medmikro.ruhr-uni-bochum.de/nrz/) contains detailed information on efficient carbapenemase diagnostics, which any laboratory can implement. The NRZ also offers testing of suspicious isolates.

• According to the KRINKO recommendations (p. 1313, A) [1], an isolate of Enterobacteriaceae or the A. baumannii complex should be classified as 4MRGN if a carbapenem resistance is present. Since the rating is intermediary and resistant for the purposes of this classification, and an OR link is applied to Enterobacteriaceae, such an isolate would formally have to be rated as 4MRGN. In this respect, this is a borderline case, as most experts would rate such an isolate as less significant hygienically than a carbapenemase-producing strain with the same phenotype. But from a microbiological point of view, and in spite of the 4MRGN classification, a simplified approach could be recommended for the de-isolation of the patient in the further course.

In summary, it can be stated that the new MRGN classification comes quite close to the objective of identifying pathogens of particularly relevant pathogens in terms of hygiene and marking them out for the responsible hospital staff. The MRGN classification is complex by its very nature. Simplistic multi-resistance definitions are of no use when it comes to the meaningful control of hospital-hygienic interventions, such as single-room isolation. For a correct MRGN classification, the expertise of the microbiological laboratory is indispensable. In particular, carbapenemases must be detected reliably. In addressing the looming crisis caused by multidrug-resistant, Gram-negative bacteria, microbiological laboratories play a crucial role.