Seven isolates of hetero-VISA from clinical specimens were used in this study. Four isolates were from Korea (K1272, K1299, K193, and K237), two from Japan (Mu3 and J51), and one from India (I93). Two vancomycin-susceptible MRSA strains (MRSA120 and MRSA202) were also tested. All nine strains were resistant to oxacillin. Reduced susceptibility to vancomycin of S. aureus isolates was confirmed by the method of population analysis as previously described (1). Hetero-VISA was defined as a strain that contained subpopulations of cells that grew on the 4 mg/L vancomycin plate at a frequency of 10^-6 or higher (1).

Six antimicrobial agents were used in the in vitro susceptibility test and time-kill assay ; arbekacin (Meiji-Seika Co. Ltd., Tokyo, Japan), vancomycin (Sigma, St. Louis, MO., U.S.A.), rifampin (Sigma), ampicillin-sulbactam (Pfizer Pharmaceuticals Korea, Ltd., Seoul, Korea), teicoplanin (Sigma), and quinupristin-dalfopristin (Rhone-Poulenc Rorer, PE, U.K.).

MICs were determined by broth microdilution method of the National Committee for Clinical Laboratory Standards (NCCLS) (21). MIC determinations were performed using cation-adjusted Muller-Hinton broth (CAMHB). The minimal bactericidal concentrations (MBCs) were determined by subculture of wells with no visible growth after MIC determination. From each microtiter wells, 0.1 mL aliquots were cultured on blood agar plates (Becton-Dickinson, Sparks, MD, U.S.A.) and colonies were counted after 18-24 hr incubation at 37℃. The MBC was defined as the lowest concentrations of antibiotics that reduced the inoculums by ≥99.9% (21). All assays were performed in duplicate.

Time-kill assay was performed with the modified method of Watanabe et al. (14). For time-kill assay, antimicrobial agents were used at concentrations of 0.5× and 1× MIC. Time-kill assay was performed in CAMHB with isolates of 1.5×10⁶ colony-forming unit (CFU)/mL. A 0.1 mL suspension of each isolates was added to 5 mL of CAMHB with each antibiotics. Bacterial culture tubes were incubated at 37℃ with constant shaking for 24 hr. Arbekacin and other antimicrobial agents were tested alone, or in combination, at concentrations of 0.5× and 1× MIC. Teicoplanin and quinupristin-dalfopristin were not tested against 2 MRSA strains. Aliquots (0.1 mL) of bacterial culture were removed from cultures at 0, 4, 8, and 24 hr. Each aliquot was serially diluted in sterile saline and plated on to blood agar plates; colonies were counted on plates yielding 10-100 colonies after incubation at 35℃ for 24 hr. The minimum detection limit when plating 0.1 mL of bacterial culture is about 2 log₁₀ CFU/mL. Tests were performed in duplicate; results are expressed as mean log₁₀ CFU/mL. Synergy and additivity/indifference were defined, respectively, as a ≥2 log₁₀ CFU/mL decrease and a <2 log₁₀ CFU/mL change in the average of viable count at 24 hr for organisms treated with the combination, in comparison with the most active single drug. Antagonism is a negative interaction; the combined effect of the drug being examined is significantly less than their independent effect (22). The killing activities of various antibiotic regimens were expressed as log₁₀ CFU/mL changes in the number of surviving bacteria after incubation for 0, 4, 8, and 24 hr. Serial dilution of plated samples coupled with filtration using a 0.45 micron filter was performed to minimize antimicrobial carryover effect (23).

Mean bacterial concentrations in each regimen were compared by one-way analysis of variance with the post-hoc test for multiple comparisons (SPSS, release 11.0; SPSS Inc., Chicago, IL, U.S.A.). A p-value of <0.05 was considered significant.

The MICs and MBCs of arbekacin, vancomycin, rifampin, ampicillin-sulbactam, teicoplanin, and quinupristin-dalfopristin for nine stains are represented in Table 1. All nine strains were oxacillin-resistant and their mecA genes were confirmed by PCR method (data not shown). The MICs and MBCs of arbekacin ranged from 0.25 to 4 mg/L and from 1 to 32 mg/L, respectively. The MIC : MBC ratios of arbekacin ranged from 2 to 8, indicating no antimicrobial tolerance. No tolerant strains for the other antimicrobials were found.

Single regimen of arbekacin resulted in re-growth of 4 hetero-VISA (K193, Mu3, J51, and I93) and 2 MRSA strains after 8 hr. The combination regimens of arbekacin with vancomycin, ampicillin-sulbactam, or teicoplanin were synergistic against strains of MRSA and hetero-VISA either at both concentrations (0.5× and 1× MIC) or at 1× MIC (Table 2). Combination of arbekacin and vancomycin showed the synergistic killing effect in all strains of hetero-VISA and MRSA except one (MRSA120). Combination of arbekacin and rifampin showed the synergistic killing effect in only three hetero-VISA (K1272, K1299, and Mu3) strains and one MRSA strain (MRSA120). Combination of arbekacin and teicoplanin or ampicillin-sulbactam was synergistic against 4 strains out of 7 hetero-VISA strains. The combination of arbekacin and quinupristin-dalfopristin was not synergistic against all 7 hetero-VISA. Fig. 1 showed the bacterial killing effect of arbekacin based combination regimens at 0.5× and 1× MIC concentrations against a representative hetero-VISA strain, Mu3. Antagonistic interaction was not observed in any combination regimens. The combination of arbekacin and ampicillin-sulbactam was the most effective significantly at 1× MIC concentration against 5 strains out of 7 hetero-VISA strains (p<0.05). At 0.5× MIC concentration, however, the combination of arbekacin and vancomycin was the most effective significantly against 4 hetero-VISA strains (p<0.05) (Table 2).