Editorial 

A Multinational Trial of Rapid Antimicrobial Susceptibility Testing: Is FASTer Better?

Arjun Srinivasan

JAMA Published Online: April 18, 2026

doi: 10.1001/jama.2026.5504

Selecting the right antibiotic to treat an infection is the cornerstone of effective antibacterial therapy and lifesaving in severe infections.¹ The growing epidemic of antimicrobial resistance is making decisions about empiric treatment increasingly difficult and elevating the importance of having pathogen susceptibility data.² Unfortunately, the time to get this vital information remains agonizingly long. In most microbiology laboratories, organisms must be grown and then tested after isolation from blood culture samples, a process that usually takes more than 24 hours. While awaiting this crucial information, clinicians are faced with a difficult dilemma: choose potentially excessively broad-spectrum antibiotics and risk potentially preventable adverse events or risk ineffective therapy. In nearly all cases, clinicians will choose the former, helping fuel the downward spiral of antimicrobial resistance because broad-spectrum antibiotics increase the selective pressure for resistant organisms, which, in turn, require even broader-spectrum antibiotics.

Faster results on antibiotic susceptibility could provide at least a partial off-ramp. Knowing the right antibiotic in hours, rather than days, would allow clinicians to optimize therapy earlier in the treatment course, perhaps in time to change outcomes for the better. Fortunately, the technology to make this possible is not science fiction. Rapid susceptibility testing systems are commercially available and have been associated with lower lengths of stay and even mortality reductions in some studies.³ However, that literature has been challenging to interpret. The evidence for shorter length of stay and mortality benefits has come mostly from observational studies. Prospective trials have shown benefits in decreased time to optimal therapy, but not mortality or length of stay, although all have been conducted in settings with low rates of antibiotic resistance.

Against this backdrop, we now have the results of the Fast Antimicrobial Susceptibility Testing for Gram-Negative Bacteremia Trial (FAST) published in this issue of JAMA.³ In a multinational randomized clinical trial, Banerjee and colleagues evaluated whether rapid antimicrobial susceptibility testing (AST) performed directly from positive blood culture bottles improves clinical outcomes for gram-negative bacteremia compared with standard AST methods. The FAST trial included 850 hospitalized adults and children across 7 centers in Greece, India, Israel, and Spain. Cephalosporin or carbapenem resistance was detected in 44% of patients overall.

Participants were randomized to undergo either rapid AST using the VITEK REVEAL system plus standard AST or standard AST alone. All patients were evaluated by local antimicrobial stewardship programs, which issued treatment recommendations based on initial Gram stain results, organism identification, and susceptibility findings. The trial’s primary end point—the desirability of outcome ranking (DOOR) at day 30—integrated mortality, clinical response, discharge outcomes, and deleterious events, such as kidney failure or acquisition of multidrug-resistant organisms. The DOOR methodology has been used in many of the trials run by the National Institutes of Health–funded Antibacterial Resistance Leadership Group, and its strengths and limitations have been described elsewhere.⁴ The authors also examined several secondary outcomes and explored outcomes in the subsets of patients with cephalosporin- and carbapenem-resistant infections.

Rapid AST did not achieve superiority over standard AST. The probability that a patient in the rapid testing group would have a more desirable DOOR outcome than a patient in the standard group was 48.8% (95% CI, 45.3%-52.4%), narrowly failing to exceed the prespecified 50% threshold for superiority. However, rapid AST was associated with important benefits in duration of hospitalization. Although overall length of stay was similar between the rapid and standard AST groups, the percentage of patients who remained hospitalized at 30 days post enrollment was 4% lower in the rapid than standard AST group and this difference increased to 13% among patients with carbapenem-resistant infections. Likewise, the median number of days hospitalized after enrollment was 3 days shorter in patients with cephalosporin-resistant infections.

There were also important process of care benefits in the group randomized to undergo rapid AST. Patients undergoing rapid AST had significantly shorter time to first antibiotic modification, with a median of 22 hours compared with 36 hours in the standard group. This advantage was magnified among patients with cephalosporin- or carbapenem-resistant infections, where antibiotic changes were made 23 and 29 hours faster, respectively. Additionally, and perhaps most important, in patients receiving undertreatment, rapid AST led to antibiotic escalation a full 24 hours earlier than standard AST.

The FAST trial presents the same interpretation conundrum as some of the other trials of rapid AST: if patients were given proper therapy faster, why were there not clearer benefits related to mortality and length of stay? This could partly be a sample size challenge. One would expect to see a mortality benefit from rapid AST primarily in the subset of patients who had severe infections, such as sepsis, who were also receiving ineffective empiric therapy. Despite the size of the FAST trial, only approximately one-quarter of patients had septic shock. Additionally, despite the high rates of resistance, more than 60% of patients were already receiving effective antibiotics when they were enrolled. It is likely that the subset of patients with septic shock and receiving ineffective therapy enrolled in the FAST trial was too small to assess outcome differences between rapid and standard AST. Attempting to conduct a study that enrolled only patients with septic shock and receiving ineffective empiric therapy would require an enrollment size that is likely well beyond what could be realistically supported with current research resources. The authors also point out that in some cases, the optimal antibiotic was not available because of various access issues, such as a national shortage of aztreonam in Greece during the study. Rapid AST would obviously have had no benefit in cases where the agent identified by the test was not available to the patient. Issues with access to effective antibiotics, especially newer agents, remain an important challenge in resource-limited settings.⁵

Unlike previous rapid AST trials, the FAST study did show some important benefits in duration of hospitalization, especially among patients with resistant infections. It appears that the length of stay benefits occurred primarily in patients with long durations of hospitalization; for example, patients still hospitalized 30 days after enrollment. Considering this finding, it was interesting to see the major difference in antibiotic deescalation. Among patients receiving unnecessarily broad-spectrum antibiotics, the median time to deescalation was faster in the rapid testing group because less than 50% of patients in the standard group underwent any deescalation. Studies have shown important benefits to deescalation that can reduce the risk of prolonged hospitalization, including reduced risks of acute kidney injury⁶ and development of antibiotic resistance.⁷

Banerjee and colleagues should be commended for undertaking this challenging study, which was not only evaluating a rare outcome, but also posed substantial implementation challenges regarding workflow. It would have been interesting to see details on the cost differences between rapid and standard AST, but these might not have been available because the technology was implemented as part of a trial. In US hospitals, the added costs of adopting rapid AST could be at least partially offset by New Technology Add-On Payments from the Centers for Medicare & Medicaid Services, for which the system studied in the FAST trial does appear to qualify.⁸ Additionally, as with most new technologies, it is likely that the price will drop over time.

The FAST trial should move hospitals firmly in the direction of implementing rapid AST. Certainly, the benefits are likely to be greater in patients with resistant-organism infections and septic shock. However, the former is never known at the time of presentation and the latter can easily develop after blood cultures are obtained and before results are known. Therefore, trying to focus rapid AST on a subset of patients where it might have clear mortality benefit seems impractical and unnecessary. Nearly every hospital in the US, and many globally (including all the hospitals in the FAST trial), have antibiotic stewardship programs focused on ensuring that patients receive optimal antibiotic therapy.⁹ If we have a tool that will help antibiotic stewardship programs support clinicians in optimizing therapy faster, an outcome that is consistent across all studies of rapid AST, should we not work toward making that tool available? The answer from a patient-centric perspective is clearly yes.