The expert rules laid down by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) suggest that if an isolate of the family Enterobacteriaceae is intermediate or resistant to tobramycin and susceptible to gentamicin and amikacin, then its amikacin susceptibility status should be revised from "susceptible" to "intermediate" because production of acquired AAC(6')-I may not confer phenotypic amikacin resistance. In a previous study, we observed that over 40% of the Enterobacter cloacae isolates had the aac(6')-Ib gene [4]; however, many of the isolates with this gene were susceptible to amikacin [4]. Therefore, in this study, we determined the aminoglycoside susceptibility profiles of aac(6')-Ib-harboring E. cloacae isolates. Further, we investigated the aac(6')-Ib mutations (Leu119Ser, Leu120Ser, Glu167Ala, Phe171Ala, and Tyr166Ala) that are known to be associated with the loss of amikacin resistance [5-7].

We had previously collected 178 consecutive, non-duplicate isolates of E. cloacae from specimens obtained at 12 clinical microbiology laboratories in Korea between March 2005 and July 2005. The aac(6')-Ib gene was PCR amplified using 2 primers-5'-TTGCGATGCTCTATGGGCTA-'3 and 5'-CTCGAATGCCTGGCGTGTTT-'3-to obtain a 482-bp product [8]. In our previous study, all the 178 isolates were analyzed for the presence of aac(6')-Ib-cr [4]. Of the 74 E. cloacae isolates that were found to be positive for aac(6')-Ib, 58 were available for this study. For these 58 isolates, the minimum inhibitory concentrations (MICs) of amikacin (2-256 mg/L), kanamycin (2-256 mg/L), tobramycin (0.5-64 mg/L), and gentamicin (0.5-64 mg/L) were determined by the agar dilution method according to the CLSI guidelines [9].

To detect the mutations associated with the loss of amikacin resistance, the 482-bp PCR products were sequenced using a DNA Analyzer (Applied Biosystems, Foster City, CA, USA). To investigate the clonal relatedness of the isolates, pulsed-field gel electrophoresis (PFGE) was performed using a CHEF Mapper system (Bio-Rad Laboratories, Hercules, CA, USA). The whole-cell DNA was digested with XbaI, and the results were interpreted according to the criteria proposed by Tenover et al. [10].

Table 1 shows the antibiotic susceptibilities for the isolates. Of the 58 isolates harboring the aac(6')-Ib gene, 49 (84.5%) were susceptible to amikacin (≤16 mg/L); 2 (3.4%), to kanamycin (≤16 mg/L); 2 (3.4%), to tobramycin (≤4 mg/L); and 17 (29.3%), to gentamicin (≤4 mg/L) according to the CLSI breakpoints. According to the EUCAST breakpoints, 32 (52.2%) isolates were susceptible to amikacin (≤8 mg/L); 2 (3.4%), to tobramycin (≤2 mg/L); and 11 (19%), to gentamicin (≤2 mg/L). The distributions of the MICs of different aminoglycosides for the 58 E. cloacae isolates are shown in Fig. 1 and 2. Of the 49 amikacin-susceptible isolates (MIC ≤16 mg/L), only 2 isolates-KN7 and SO15-had the Leu119Ser mutation in the aac(6')-Ib gene. The KN7 isolate had 1 more amino acid change (Arg173Lys) in this gene. Findings of PFGE showed 10 different clones and no clonal relatedness among isolates from different clinical microbiology laboratories. However, clonal relatedness was observed among isolates from same hospitals: 2 out of 8, 2 out of 4, 2 out of 6, 3 and 2 out of 7, 2 out of 6, 2 and 2 out of 8, 4 out of 5, and 2 out of 2 isolates collected from 8 clinical microbiology laboratories. The isolates belonging to the same clone revealed identical PFGE pattern, but one isolate from hospital GM was related. We observed that 19 isolates belonging to 8 different clones were susceptible to amikacin (MIC ≤16 mg/L) and had different MICs for other aminoglycosides. However, 4 isolates belonging to the remaining 2 clones were resistant to all the tested aminoglycosides (MIC for amikacin and kanamycin, ≥256 mg/L; MIC for tobramycin and gentamicin, ≥64 mg/L) (data not shown).

A very high percentage of the aac(6')-Ib-harboring isolates were susceptible to amikacin (84.5% and 55.2% according to the CLSI and EUCAST breakpoints, respectively). However, among the isolates that were shown to be amikacin-susceptible according to the CLSI and EUCAST criteria, the phenotypes of only 30.6% (15/49) and 18.8% (6/32) isolates, respectively, could be revised according to the EUCAST expert rules because many of these amikacin-susceptible isolates were resistant to gentamicin. The gentamicin resistance may be because of gentamicin-modifying enzymes such as AAC(3)-I, AAC(3)-VI, AAC(2')-I, AAC(3)-IV, Ant (2')-I, and ACC(3)-II or impermeability. Since only 2 isolates had mutations associated with the loss of amikacin resistance, we think that the remaining isolates might have produced low-levels of AAC(6')-Ib. Low levels of AAC(6')-Ib could not confer resistance to amikacin and isepamicin in vitro but were able to efficiently modify the small amounts of drug entering the bacterial cell and thereby conferred resistance in vivo. Therefore, a previous study did not report any noticeable difference in the in vivo efficacies of amikacin or isepamicin for low- and high-level AAC(6')-Ib-producing organisms [11].

Our results show that the MICs of amikacin for many AAC(6')-Ib-producing isolates were below the susceptibility breakpoints recommended by the CLSI or EUCAST and that most of these susceptible isolates could not be considered to have intermediate amikacin resistance according to the above-mentioned EUCAST expert rules. Data on the epidemiology and optimal therapy of infections caused by these strains are scarce; therefore, further studies are needed to characterize these infections.