Fast antimicrobial susceptibility testing: What to know (FAST trial)
FAST is faster and improves the quality of antibiotic treatment. As for outcomes, well ...
For most patients who’ve been in the ICU more than a few days who then develop a new suspected bacteremia, empiric treatment with very broad-spectrum antibiotics like piperacillin-tazobactam, cefepime, or carbapenems is almost a clinical reflex. After that, an organism’s sensitivity usually isn’t known for at least a few days, during which broad coverage is continued in most patients.
A new generation of rapid antimicrobial sensitivity tests may reduce that timeframe. Evidence is emerging that these assays can direct optimized antibiotic therapy earlier, in either direction (broader, for resistant organisms, or narrower, when resistance is unlikely).
There’s no proof they improve hard outcomes— but could insisting on that level of certainty be a disservice to patients and the antibiotic stewardship project?
FAST: It’s Faster (But Not That Fast)
The status quo
First, a recap of traditional sensitivity testing. After standard blood cultures have incubated in broth media (in the bottles) and “turned positive” (usually by detection of high CO2 production inside the bottle), the broth is inoculated onto agar plates. Once enough colonies grow, the organism can be identified through conventional methods and complementary molecular techniques.
Antibiotics can then be applied directly to the agar to determine sensitivity, with other more advanced techniques often used (e.g., dilution of the broth into multiple wells containing antibiotics).
Conventional sensitivity testing takes about 24 to 48 hours after the blood cultures turn positive, often two to five days after blood cultures are drawn.
Rapid susceptibility testing
Rapid susceptibility tests are performed on the blood culture broth in the inoculated bottles after the cultures have returned as positive—rapidly, without the necessity of waiting for growth on agar or other media.
They use techniques not dissimilar to conventional susceptibility testing (e.g., diluting broth across a matrix of antibiotic-impregnated microwells), but identify signs of resistance faster (e.g., detecting volatile metabolites from the wells, or microscopic optical changes, rather than waiting for the traditional slower signals).
Rapid AST assays can return susceptibility results in less than 8 hours.
It’s important to distinguish rapid AST (which test an actual cultured bacteria for observable resistance) from inference-based rapid molecular genetic tests (e.g., BioFire™, Cepheid Expert™) that identify bacterial DNA and resistance genes without actually observing resistance or susceptibility.
Six randomized trials have tested rapid AST in patients with bacteremia:
Banerjee et al JAMA 2026, (the FAST trial, n=850)
All found that use of rapid AST identified organisms and antimicrobial resistance faster, resulting in faster delivery of optimal antimicrobial therapy—usually by 12 to 48 hours faster than conventional sensitivity testing.
This usually led to more accurate targeting of antibiotics and less use of broad spectrum agents.
Rapid testing had high agreement (e.g., ~94%) with conventional testing.
However, none of the six RCTs found that use of rapid AST reduced mortality, hospital lengths of stay, or other “hard” clinically relevant outcomes.
The FAST Trial
As the first five trials were conducted in countries with relatively low rates of antibiotic-resistant organisms, the FAST trial was performed specifically to test the use of rapid AST at hospitals in nations with high antibiotic resistance.
At 7 medical centers in 4 countries with a high prevalence of multidrug-resistant gram-negative bacteria (Greece, India, Israel, and Spain), 899 patients with confirmed gram-negative bacteremia were randomized to standard susceptibility testing or fast AST using Biomerieux’s system (VITEK Reveal).
There was even less support for rapid AST in this trial than its predecessors: similar proportions of patients received appropriate antibiotics at enrollment, 24 hours, and at 3 days.
However, patients with resistant organisms had a shorter time to optimal targeting of their antimicrobials (escalation or de-escalation). For cephalopsporin-resistant organisms, rapid AST led to changes in ~12 vs. 25 hours with conventional testing; for carbapenem-resistant bugs, tailoring occurred after 13 vs. 47 hours.
Why No Benefit?
Rapid susceptibility testing occurs well after the critical decision and time window has passed for treatment of severe infections with antibiotics. Most patients with positive blood cultures have already received antibiotics for the two days while their cultures were incubating. Empiric therapy, not downstream tailoring does almost all the work in saving lives from sepsis.
Patients in clinical trials are often already receiving high-quality care at advanced centers, which for patients with suspected bacteremia in the ICU includes broad-spectrum antibiotics; this likely led to good outcomes in the control arms of these trials.
It’s also possible that most patients don’t suffer bad outcomes from being on suboptimal antibiotics for 1-2 days, on average.
Mortality is driven by many other factors besides antibiotic susceptibility. Anecdotally, the patient whose death is intuitively attributable to undertreatment with broad spectrum antibiotics is a rarity; overtreatment is by far the more common scenario in modern ICUs.
It’s Worth Being Bullish on Rapid AST Anyway
Accumulated evidence strongly suggests that rapid susceptibility testing can lead to faster appropriate antibiotic de-escalation, and likely leads to more rapid escalation to cover resistant organisms as well. It’s rare to find a new test that works so consistently well on an important, clinically-relevant measure.
Although pricing is opaque, rapid susceptibility testing appears to be expensive, both on a platform and per-unit basis, and in the absence of demonstrable clinical or financial benefit, adoption may be slow. This is an area of ongoing innovation and competition among vendors of the competing diagnostic platforms and assays.
For the practicing clinician, an important distinction at this stage of the game is between assays that test actual resistance on live, cultured bacteria (e.g. the rapid AST test we’ve been discussing and its competitors), which have been demonstrated to have a high concordance with conventional susceptibility testing — versus the molecular DNA-based tests that only infer or predict resistance based on genetic profiles. The latter class of tests can provide a lot of value in early targeting decisions, but must be complemented with phenotypic resistance testing (rapid and/or conventional AST).
References
Banerjee et al JAMA 2026, (the FAST trial, n=850)
In my mind, if I can find out the bacteria faster, I can de-escalate faster, then the patient gets better faster and then patient goes home faster. If the studies don’t show that, then what’s the point. And if they don’t reduce antibiotic resistance then further what’s the point?