Discrepancies in susceptibility testing to colistin in Acinetobacter baumannii: The influence of slow growth and heteroresistance.

The increasing use of polymyxins as last-resort drugs for managing infections by Acinetobacter baumannii has led to the emergence of resistance. This study aimed to determine the resistance mechanisms in Acinetobacter baumannii isolates with colistin MIC ≥ 4 mg/L and to relate the mechanisms of resistance with the difficulties in detecting them. Absolute agreement among the different methodologies (Phoenix automatized system, broth and agar dilution, and a rapid colorimetric test) in the 140 colistin-susceptible isolates was observed; whereas in the 25 resistant isolates, the performance varied according to the colistin MIC value. Most of the discrepancies (irrespective of the methodology that was used) were observed in isolates with an MIC value close to the breakpoint. The number of errors in each method in the resistant isolates was as follows: rapid test, four of 25 (16%); agar dilution, eight of 25 (32%); Phoenix system, 13 of 25 (52%) and its manual reading at 24 h, eight of 25 (32%). Categorical errors were detected in 13 isolates: slow growth was the main reason in five isolates, whereas in the remaining eight isolates, slow growth was detected together with a low proportion of colistin-resistant subpopulations and the colistin MIC value was close to the breakpoint value. To understand the probable reason for the observed MIC values, sequencing of genes associated with colistin resistance was performed. Mutations at lpxA, lpxC, and pmrB genes were detected and it was observed that isolates carrying mutations in lpxC presented slow growth at killing curves.