Background: Various nosocomial illnesses have been linked to Acinetobacter
baumannii, a prominent nosocomial pathogen. For the treatment of infections caused
by multidrug-resistant Gram-negative bacilli, particularly the very resistant A.
baumannii, colistin is now regarded as a treatment of last resort. In a time when MDR
Gram-negative bacteria are becoming more prevalent and colistin resistance is also
rising, a precise technique for measuring colistin's antimicrobial susceptibility is
essential. The aim of this study is to evaluate the phenotypic colistin susceptibility
testing methods in carbapenem-resistant A. baumannii (CRAB) and to detect the
underlying molecular presence of mcr-1 gene alongside these phenotypic methods.
Methods: Gradient diffusion (E-test) and automated systems (Vitek-2, BD Phoenix™)
were used to examine 50 non-repeated A. baumannii isolates, and the outcomes were
compared to the reference broth microdilution (BMD) method which was performed
according to ISO 20776-1 guidelines. Additionally, PCR was used to find the plasmidencoded mcr-1 gene. Results: The reference BMD method yielded 22% resistant
isolates. E-test showed highest relative accuracy with 96% categorical agreement (CA),
86% essential agreement (EA), and 18.18% very major errors (VME), with no major
errors (ME). Vitek-2 compact and BD Phoenix™ showed identical performance with a
CA of 88%, EA of 76%, VME rates of 45.45%, and ME rates of 2.56%. Mcr-1 gene
was found in 12% of isolates and was positive in 45.5% of colistin resistant strains.
Conclusion: To fight against CRAB, laboratories should prioritize MIC-based
accuracy over convenience to guide life-saving therapy. Compared to BMD as the
reference method, E-test offers the highest relative accuracy followed by automated systems.