Colistin Dependency among Colistin-Heteroresistant Acinetobacter baumannii Isolates

Abstract

Colistin dependent (CD) isolates are dependent on colistin for optimal growth. Here we aimed to systematically determine the emergence of CD among colistin-heteroresistant carbapenem-resistant Acinetobacter baumannii (CRAB) isolates. We also examined the phenotypic characteristics of CD and the evolution of CD strains to overt resistance. Additionally, we examined whether detection of growth in blood cultures was impaired by CD. Heteroresistant isolates, as determined by population analysis profiling, were exposed to colistin; when the colony count with colistin was significantly higher than without, isolates were suspected to be CD. CD was confirmed by Etest and growth curves. CD strains with colistin minimum inhibitory concentrations > 2 mg/L after growth in colistin-free media were considered colistin-resistant. Of the 65 heteroresistant strains tested, eight became CD after colistin exposure. These strains attained higher colony counts and growth rates with colistin vs. without, and grew adjacent to the colistin Etest strip. CD strains exhibited increased susceptibilities to multiple antibiotics compared to their parent heteroresistant strains. All CD strains tested became colistin-resistant following growth without colistin. CD strains were detected in blood culture bottles, but time to detection was significantly prolonged compared with parent strains, suggesting that CD may lead to delay in detection of CRAB bacteremia.

Keywords: blood culture; carbapenem-resistant Acinetobacter baumannii; colistin dependency; colistin heteroresistance.

Figure 1

Heteroresistance (HR) determination according to population analysis profiling for the A. baumannii isolates. Growth of the 65 colistin-susceptible isolates at 8×MIC indicated HR. The red dots represent the colistin-resistant control, and the green dots represent the colistin-susceptible non-HR control.

Figure 2

(A) Determination of CD. Final colony count (CFU/mL) of CD strains grown with and without colistin. Black bars represent the absence of colistin, and gray bars represent 8×MIC colistin. Results are presented as mean ± SD of two independent repeats. Mean CFU/mL was significantly higher when colistin was present than when it was absent; p < 0.05 for all CD strains (* p < 0.05, ** p < 0.01), as determined by Student’s t-test. (B) Colistin Etest. Etest plates of parent (left), CD (middle) and CR-CD (right) strains. (C) Growth curves. Growth curve of parent (left) and CD (right) strains grown with and without colistin for 24 h. Black lines represent the absence of colistin, and gray lines represent 8xMIC colistin. Results are presented as mean ± SD of three independent tests.

Figure 3

Dendrogram obtained by FT-IR biotyping of colistin-dependent A. baumannii isolates (n = 8). The blue line represents the cutoff value. Clusters composed of two of more isolates are shaded orange. (A) OXA-66-type strains. (B) OXA-71-type strains. Controls–unrelated strains from laboratory collection.

Figure 4

Change in colistin MIC between parent strains and their offspring CD strains.

Figure 5

Bacterial growth in blood culture of eight colistin-dependent isolates. Each black dot represents the difference in TTD between the CD strain and its parent strain. The lines represent the average difference at each colistin concentration.