Rapid antimicrobial susceptibility testing was not superior to standard testing for 30-day clinical outcomes among patients with gram-negative bloodstream infections, but it was associated with faster antibiotic modification and higher early receipt of effective therapy in the FAST randomized clinical trial published in JAMA.
In the open-label trial, researchers randomized 899 hospitalized pediatric and adult patients at seven centers in Greece, India, Israel, and Spain. The primary analysis included 850 patients assigned to rapid phenotypic antimicrobial susceptibility testing directly from positive blood cultures plus standard testing (n = 413) or standard testing alone (n = 437). Local antimicrobial stewardship teams reviewed all patients and provided treatment recommendations.
The primary end point was desirability of outcome ranking at day 30, an ordinal composite designed to capture both benefit and harm in a single patient-centered rank, integrating mortality, clinical response, discharge outcomes, kidney failure, and acquisition of multidrug-resistant organisms. The probability that outcomes were more favorable with rapid testing was 48.8%, which did not meet the prespecified superiority criterion.
“Among patients with gram-negative bacilli BSIs, rapid blood culture [antimicrobial susceptibility testing] was not superior to standard testing by [desirability of outcome ranking],” wrote lead researcher Ritu Banerjee, MD, of Vanderbilt University Medical Center, and colleagues.
Thirty-day mortality was similar between groups, occurring in 24% of patients in the rapid testing group and 23% in the standard testing group. Median length of hospitalization was 8 days in both groups, and acquisition of multidrug-resistant organisms was also similar.
However, rapid testing was associated with faster treatment optimization. Median time to antibiotic escalation or deescalation was 22 hours with rapid testing vs 36 hours with standard testing. The proportion of patients receiving effective antibiotic therapy within 24 hours was 84% vs 75%, respectively.
The process-of-care differences were more pronounced among patients with resistant infections. Among patients with carbapenem-resistant infections, 72% of those in the rapid testing group received effective antibiotic therapy within 24 hours vs 47% in the standard testing group. Median time to effective antibiotic therapy was 9.5 hours vs 28 hours, respectively. At 30 days, 15% of patients with carbapenem-resistant infections in the rapid testing group remained hospitalized vs 28% in the standard testing group.
These subgroup findings should be interpreted cautiously because the trial was not powered to establish definitive clinical outcome differences in resistant-infection subgroups; this limitation applies to all subgroup comparisons, including time to effective therapy and hospitalization outcomes.
In an accompanying editorial, Arjun Srinivasan, MD, of the Joint Commission and Emory University School of Medicine, wrote that the lack of a mortality benefit may reflect the difficulty of detecting such a benefit in a feasible randomized trial. More than 60% of patients were already receiving effective antibiotics at enrollment, and only approximately one-quarter had septic shock. The subgroup most likely to benefit—patients with septic shock who were receiving ineffective empiric therapy—was likely too small to show outcome differences.
The study researchers also noted several reasons rapid testing may not have translated into clearer clinical outcome improvements. Many patients had antibiotic-susceptible infections, for which rapid testing may be more useful for safe deescalation than for mortality-altering escalation. Access to optimal antibiotics also varied by site, including an aztreonam shortage in Greece and limited availability of newer agents in some settings.
Deescalation may be one of the most clinically important findings for stewardship programs. Among patients receiving unnecessarily broad-spectrum antibiotics, deescalation occurred faster with rapid testing, while fewer than half of patients in the standard testing group underwent deescalation. Srinivasan noted that deescalation may reduce risks associated with prolonged broad-spectrum therapy, including acute kidney injury and emergence of resistance.
The trial had several limitations. It was open label, and the intervention combined rapid testing with antimicrobial stewardship review, so the independent effect of rapid testing could not be isolated. Sites differed in resistance prevalence, standard testing workflows, stewardship resources, antibiotic availability, and local treatment practices. Antibiotic regimens were not standardized. Nearly 20% of enrolled patients had organisms not included on the rapid testing panel, limiting the potential effect of the intervention. The rapid platform also did not provide organism identification, and all sites used MALDI-ToF mass spectrometry separately, which may limit generalizability to laboratories without that capacity.
Cost-effectiveness was not assessed. In the editorial, Srinivasan noted that added costs in US hospitals may be partly offset by New Technology Add-On Payments from the Centers for Medicare & Medicaid Services, although implementation costs and local workflow considerations remain important.
Srinivasan concluded that rapid [antimicrobial susceptibility testing] “should move hospitals firmly in the direction of implementing” the technology, particularly in settings managing resistant infections.
The study was supported by the National Institute of Allergy and Infectious Diseases. bioMérieux provided instruments and reagents and reviewed and approved the manuscript but had no role in study design, conduct, data analysis, or the decision to submit. Several researchers reported NIH or institutional research support and industry relationships. Dr. Srinivasan reported no conflicts of interest.
Source: JAMA