Evaluation of automated systems for aminoglycosides and fluoroquinolones susceptibility testing for Carbapenem-resistant Enterobacteriaceae

BACKGROUND: Automated systems (MicroScan WalkAway 96 Plus, Phoenix 100, and Vitek 2 Compact) are widely used in clinical laboratories nowadays. The aim of this study is to evaluate the performance of these three systems for susceptibility testing of aminoglycosides and fluoroquinolones against Carbapenem-resistant Enterobacteriaceae (CRE). METHODS: A total of 75 CRE isolates were used in this study. Quinolone resistance determinants (QRDs) (qnrA, qnrB, qnrC, qnrD, qnrS, aac(6′)-Ib-cr, oqxAB and qepA) and aminoglycoside resistance determinants (ARDs) (aac(6′)-Ib, armA, npmA, rmtA, rmtB, rmtC, rmtD and rmtE) of these CRE were screened by PCR. The MICs of aminoglycosides (gentamicin and amikacin) and fluoroquinolones (ciprofloxacin and levofloxacin) to CRE obtained with the automated systems were compared with the reference method (agar dilution method). RESULTS: Totally, 97.3% (73/75) of CRE harbored QRDs. The qnr gene was the most common QRD determinant identified in 68 (96.7%), followed by aac (6′)-Ib-cr in 56 (74.7%), oqxAB in 23 (30.7%), and qepA in 2 (2.7%), respectively. 22.7% (17/75) of CRE harbored ARD determinants. rmtA, rmtB and npmA were identified among these isolates in 6 (8.0%), 6 (8.0%) and 5 (6.7%), respectively. A total of 900 results were obtained in this study. Overall, the total error rate was 9.89%. Twenty-eight very major errors (3.11%), 22 major errors (2.44%) and 39 minor errors (4.33%) were identified against agar dilution method. The very major errors were almost evenly distributed between results for fluoroquinolones (2.89%) and aminoglycosides (3.33%), while the major errors and minor errors were more commonly found in the results of fluoroquinolones (3.11% and 6.44%, respectively) than aminoglycosides (1.78% and 2.22%, respectively). CONCLUSIONS: Our study shows that testing difficulties in susceptibility testing do exist in automated systems. We suggest clinical laboratories using automated systems should consider using a second, independent antimicrobial susceptibility testing method to validate aminoglycosides and fluoroquinolones susceptibility.