Automated systems used to test the antibiotic susceptibility of a group of highly resistant bacterial species are less than foolproof, suggesting that clinical labs should consider using an additional test to validate results, according to a new report from Chinese researchers.
The Chinese team examined the accuracy of three automated systems used to test the susceptibility of carbapenem-resistant Enterobacteriaceae (CRE) to two classes of antibiotics, aminoglycosides and fluoroquinolones, according to their report yesterday in Antimicrobial Resistance & Infection Control. It said the three systems (Phoenix 100, Vitek-2 Compact, and MicroScan WalkAway 96 Plus) are commonly used in Chinese hospitals.
CRE, sometimes referred to as "nightmare bacteria," are resistant to many types of antibiotics and are associated with high mortality. The report notes that aminoglycosides are often used in combination regimens to treat CRE infections, while fluoroquinolones are sometimes used to treat urinary tract infections caused by CRE.
The research team used polymerase chain reaction to identify resistance determinants in 75 CRE isolates. Using the three automated test systems, they determined the minimum inhibitory concentrations (MICs) of two aminoglycosides (gentamicin and amikacin) and two fluoroquinolones (ciprofloxacin and levofloxacin) for the CRE isolates. These results were then compared with the MICs determined by a reference method (agar dilution).
The authors found that 97% (73 of 75) of the CRE isolates harbored quinolone resistance determinants, while 23% (17 of 75) had aminoglycoside resistance determinants.
Almost 10% error rate
Out of 900 MIC results in the study, the overall error rate for the automated systems as compared with the reference method was 9.89%, the report says. That included 28 "very major" errors (3.11%), 22 "major" errors (2.44%), and 39 "minor" errors (4.33%).
The very major errors were almost evenly distributed between results for fluoroquinolones and aminoglycosides, while the major and minor errors were more common with the fluoroquinolones than with the aminoglycosides.
"Our study shows that testing difficulties in susceptibility testing do exist in automated systems," the authors conclude. "We suggest clinical laboratories using automated systems should consider using a second, independent antimicrobial susceptibility testing method to validate aminoglycosides and fluoroquinolones susceptibility."
Aug 1 Antimicrob Resist Infect Control report