Systematic Evaluation of the Accelerate Pheno System for Susceptibility Testing of Gram-Negative Bacteria Isolated from Blood Cultures

Bacteremia is a major cause of morbidity and mortality. Rapid identification of pathogens for early targeted antimicrobial therapy is crucial for detecting emergence of antibiotic resistance and improving outcomes. However, there are limited data regarding the analytical performance of a rapid identification (ID) and antimicrobial susceptibility testing (AST) method like Accelerate Pheno blood culture detection system compared with the conventional methods routinely used in microbiology laboratories. We undertook a systematic quality improvement (QI) study to compare AST results obtained with Accelerate Pheno system rapid ID/AST system with a standard reference method in a university hospital microbiology laboratory. This was a single center, retrospective (5/10/19 to 8/1/19) and prospective (8/1/19 to 1/31/20) study that evaluated all blood cultures growing Gram-negative rods (GNR). We compared AST results obtained using the reference disk diffusion (DD) susceptibility method with those obtained by the Accelerate Pheno system. We calculated the error rates and categorical agreement between the Accelerate Pheno system and DD for each organism and specific drug tested. We evaluated 355 blood cultures growing GNR, of which 284 met the inclusion criteria. We grouped all Enterobacterales (n = 263) for analysis (156 Escherichia coli, 60 Klebsiella spp., 20 Proteus mirabilis, 17 Enterobacter spp., and 10 Serratia marcescens). Twenty-one Pseudomonas aeruginosa isolates were analyzed separately. For Enterobacterales, categorical agreement (CA) was ≥90% for amikacin (AMK), aztreonam (ATM), cefepime (FEP), ceftriaxone (CRO), ertapenem (ETP), gentamicin (GEN), meropenem (MEM), and tobramycin (TOB); and very major error (VME) was <5% for ampicillin/sulbactam (SAM), GEN, MEM, TOB, CRO, and ceftazidime (CAZ). For ciprofloxacin (CIP), CA was 87% and VME was 8%. For P. aeruginosa, CA was ≥90% for AMK and TOB, and VME was ≥5% for AMK, CAZ, GEN, MEM, piperacillin-tazobactam (TZP), and TOB. Accelerate Pheno rapid ID/AST system for GNR isolated from blood culture (BCs) was reliable for some but not all agents in the panel. Based on the findings from this study, our laboratory reports Accelerate Pheno system AST results only for Enterobacterales, and we limit our reports to CRO, CAZ, TZP, CIP, ATM, and GEN. IMPORTANCE This was an 8-month retrospective and prospective study looking at the analytical performance of the Accelerate Pheno system on clinical isolates obtained from patients seen in our tertiary care hospital. Most of the published literature on the analytical performance of Accelerate Pheno System has been from clinical trials with limited data from clinical microbiology laboratories postimplementation of the system. Here we compare the AST results on 355 blood cultures growing Gram-negative bacteria in Accelerate Pheno system with the CLSI reference disk diffusion (DD) method. The findings from this study highlight the “real-world” performance of the Accelerate Pheno system for Gram-negative bacteria from blood cultures. We provide data to show the reliable susceptibility testing results of Enterobacterales for most of the commonly used antimicrobial agents and significant limitation for susceptibility testing results of Pseudomonas aeruginosa on the Accelerate Pheno system.