AGREEMENT AND ERROR RATES IN ANTIMICROBIAL SUSCEPTIBILITY TESTING FOR THREE COMMERCIAL AUTOMATED SYSTEMS: A SYSTEMATIC LITERATURE REVIEW AND META-ANALYSIS

ABSTRACT BACKGROUND Automated antimicrobial susceptibility testing (AST) systems are crucial for accurate, timely detection of drug-resistant microbial isolates. This meta-analysis assessed the performance of the BD Phoenix™ (“Phoenix”, BD Diagnostic Solutions), Vitek ® 2 (“Vitek 2”, bioMérieux), and DxM MicroScan WalkAway (“MicroScan”, Beckman Coulter, Inc.) AST systems relative to common reference methodology. METHODS A systematic literature search in Ovid (MEDLINE and Embase) yielded 275 unique (not duplicated) records, with 44 additional records retrieved from handsearching; 39 studies met inclusion criteria. Categorical agreement (CA), essential agreement (EA), very major errors (VMEs), and major errors (MEs) for the three instruments were compared to a common reference method. Ratios of proportions were analyzed using random-effect meta-regression. RESULTS The instruments did not differ significantly in CA, EA, or ME. Vitek 2 showed a higher overall VME rate than Phoenix (∼44% higher; Vitek 2-to-Phoenix ratio = 1.44; p =0.062 [approaching significance]) and MicroScan (74% higher; ratio = 1.74; p =0.045). No appreciable difference was observed for VME between Phoenix and MicroScan. Subgroup analyses should be interpreted cautiously due to limited overall significance indicating varying performance across systems. Vitek 2 generally had higher relative VMEs for gram-negative organisms and lower relative VMEs for gram-positive organisms, whereas Phoenix showed the opposite pattern. MicroScan had relatively low VMEs when stratified by Clinical and Laboratory Standards Institute (CLSI) criteria; no differences in VMEs were observed using European Committee on Antimicrobial Susceptibility Testing (EUCAST) criteria. CONCLUSION Although some VME differences were noted, overall performance of the three systems was comparable. Organism- and drug-specific VME patterns—and updates to CLSI criteria over time—highlight the importance of continued monitoring of current breakpoints for all three instruments.