Distribution of Genotypes between CTX-M-Producing and Non-Extended-Spectrum-β-Lactamase-Producing Klebsiella pneumoniae Isolates

LETTER

We read with great interest the study by Shin and Ko (1) reporting on the comparison of the distribution of genotypes and virulence between CTX-M-producing and non-extended-spectrum-β-lactamase (non-ESBL)-producing Klebsiella pneumoniae isolates. We believe that the authors have raised an important issue that must be investigated. However, we would like to offer some comments on the interpretation of the genotype distribution results.

The authors concluded that most CTX-M-producing K. pneumoniae isolates in South Korea did not occur by transfer of the bla_CTX-M gene into susceptible strains. This conclusion was based on their observation that only a minority of their CTX-M-producing isolates represented sequence types (STs) that also occurred among their non-ESBL-producing isolates; in detail, by multilocus ST analysis, a total of 52 different STs were identified among the 98 K. pneumoniae isolates, and only 3 STs (ST11, ST15, and ST148) were detected in both CTX-M-producing and non-ESBL-producing isolates.

However, as the number of isolates varied quite a bit across the STs, it might be more appropriate to consider the number of isolates in each ST as a weight in their analysis and interpretation. Analyzing the data in Table 2 of their article, as many as 28 (28.6%) out of the 98 isolates belonged to the three STs (ST11, ST15, and ST48) and the proportion of the isolates that belong to the three STs was significantly higher in CTX-M producers than in non-ESBL producers (14/33 [42.4%] versus 14/65 [21.5%]; P = 0.03 by chi-square test). Moreover, even among the three STs, the proportion of CTX-M producers was significantly different: 31.3% (5/16) in ST11, 87.5% (7/8) in ST15, and 50.0% in ST48 (P = 0.02 by Fisher's exact test).

This finding suggests that the CTX-M-producing population seems more highly clonal than the non-ESBL-producing population and specific clones acquire bla_CTX-M more frequently. This is in line with a previous study (2) with Escherichia coli, where the genotypic diversity was lower in drug-resistant isolates and specific clones (e.g., ST131 and ST69) get antibiotic resistance more easily than others. In addition, considering the clonally distributed differences in bla_CTX-M-flanking regions (K. pneumoniae ST15 isolates contained intact ISEcp1, while ST11 and ST48 isolates harbored ISEcp1 truncated by IS26) (3), the difference in the proportion of CTX-M producers between clones might be due to the difference in the genetic environment.

Taken together, the authors' observations actually better support the concept of clonal expansion and spread than a special tendency of these particular clones to repeatedly acquire bla_CTX-M de novo. The fact that Shin and Ko could not identify within their isolate collection a susceptible variant of most of the STs in which they encountered CTX-M production does not mean that such a variant never existed or does not still exist somewhere. Current evolutionary understanding holds that bla_CTX-M moved into Klebsiella from a phylogenetically distant source, most likely Kluyvera, possibly passing through intermediate genera along the way (4). Such a transfer would have converted non-ESBL-producing Klebsiella lineages into ESBL producers. Presumably, every CTX-M-producing Klebsiella today derives from a susceptible ancestor that acquired bla_CTX-M by lateral transfer.

We applaud the authors for identifying the genotype distribution among the ESBL-producing and non-ESBL-producing K. pneumoniae genotypes and offer an alternative explanation.