PK/PD Target Interpreter: Which Index for Which Drug

Matching an antibiotic class to the wrong pharmacodynamic index is the error this category exists to prevent. Choose the class and the endpoint and this names the index, the published target, the model and organism it came from, and whether it is free or total drug.

Last updated 8 Oct 2026

Which PK/PD target applies

Class + endpoint → index

Antibiotic classThe class decides the index, and the index is the thing most often got wrong. A Cmax/MIC of 10 says nothing useful about a beta-lactam, and an fT>MIC of 70% says nothing useful about gentamicin.

Which published endpointMost published targets are animal-model endpoints, and the number depends on which endpoint was chosen: stasis needs roughly half the exposure a 2-log10 kill does. The clinical option gives the guideline or human-study target instead, which is sometimes a different index entirely.

60 to 70 percent fT>MICExample

Cephalosporins, a 1-log10 kill or greater in the animal model

The three indices, and what each one says

time-dependent: percent fT>MIC · concentration-dependent: Cmax/MIC · exposure-dependent: AUC/MIC

percent fT>MIC: the fraction of the dosing interval during which FREE drug exceeds the MIC. Belongs to the beta-lactams. Improved by shortening the interval or lengthening the infusion, barely at all by raising the individual dose
Cmax/MIC: the post-distribution peak divided by the MIC. The classical aminoglycoside index, and almost uniquely among PK/PD targets it comes from human cohorts rather than mice. Improved only by raising the peak, which toxicity limits
AUC/MIC: 24-hour exposure divided by the MIC. Belongs to vancomycin and the fluoroquinolones, and increasingly to the aminoglycosides too. Depends only on the total daily dose and the clearance, so it is indifferent to how the daily dose is split up
free against total: the single most common source of confusion. EUCAST’s rationale documents and all the beta-lactam fT>MIC figures are FREE drug; the vancomycin 400-600 window and Forrest’s quinolone AUC of 125 are TOTAL drug. Convert on the free drug concentration calculator before comparing
what probability of target attainment means: a published PTA curve is the proportion of a simulated population that reaches a stated target at each MIC, computed by Monte Carlo simulation over a population pharmacokinetic distribution — thousands of draws. This site cannot and does not draw one: its calculator engine has no loops, no arrays and no random numbers, and a fitted curve standing in for a simulation would be a plausible wrong answer. Read the curve in the paper that drew it; use these pages for the closed-form arithmetic underneath it

Worked example

Cephalosporins, a 1-log10 kill or greater in the animal model
Cephalosporins are time-dependent, so the index is the percentage of the dosing interval with FREE drug above the MIC
Berry and Kuti 2022 report 60 to 70% fT>MIC for a 1 to 2 log10 CFU reduction, attributing the figure to Craig 1998 — the highest requirement of the three beta-lactam classes
Change the endpoint to stasis and the answer falls to 30 to 40%; roughly half the exposure, for a target of no net growth rather than killing
Change the class to carbapenems and the answer falls to about 40%, usually attributed to a longer post-antibiotic effect
Change it to aminoglycosides with the clinical endpoint and the INDEX changes, not just the number: Cmax/MIC 8 to 10, with EUCAST using fAUC/MIC instead. That is the error this page exists to prevent

Index, target, model and basis — everything on one page

Read the last column before the third. Two rows for the same class can differ by an order of magnitude because one is free drug and the other total, and a target lifted out of this table without its basis is a number with no meaning.
Class	Index	Published target	Model and basis
Penicillins	percent fT>MIC	50-70% for maximal effect	Neutropenic murine models, free drug (Berry and Kuti 2022, attributing Turnidge 1998)
Cephalosporins	percent fT>MIC	30-40% stasis, 60-70% kill	Neutropenic murine thigh, free drug (Berry and Kuti 2022, attributing Craig 1998)
Carbapenems	percent fT>MIC	about 20% stasis, about 40% kill	Neutropenic murine thigh, free drug (Berry and Kuti 2022; doripenem)
Aminoglycosides, classical	Cmax/MIC	8-10	Human cohorts, total drug (Moore 1987 n=236; Kashuba 1999 n=78, via Bland 2018)
Aminoglycosides, EUCAST	fAUC/MIC	21.4-55.4 stasis, 62.5-90.4 kill	Neutropenic mouse thigh, free drug (EUCAST amikacin v3.1, USCAST 2019)
Fluoroquinolones, clinical	AUC/MIC	125	Human, total drug (Forrest 1993, via EUCAST ciprofloxacin v2.0)
Fluoroquinolones, EUCAST	fAUC/MIC	13.1-35.8 stasis, 21.0-68.7 kill	Neutropenic mouse thigh, free drug (EUCAST ciprofloxacin v2.0)
Vancomycin	AUC24/MIC	400-600	Human observational, total drug, MIC by broth microdilution (2020 consensus guideline)
Any beta-lactam, critically ill	free trough / MIC	4-8 × MIC for 100% of the interval	Expert suggestion on animal, in-vitro and observational data (SFPT/SFAR 2018 R2.2)

What the human evidence does and does not support

This is the table most PK/PD pages leave out. The murine thresholds are precise and reproducible; the human evidence that crossing them changes outcome is mixed, and the last row is a direct challenge to the whole enterprise from the people who set the breakpoints.
Question	What was found
Does crossing 50% fT>MIC change outcome?	In DALI, as reported by Berry and Kuti, patients who did not reach at least 50% fT>MIC were 32% less likely to have a positive clinical outcome — but the ROC analysis showed no clear difference between the 50% and 100% targets
Does a higher target do better?	In EXPAT, 63.3% reached 100% fT>MIC and 36.7% reached 100% fT>4×MIC, and neither was significantly associated with 30-day survival. In the ceftaroline and cefiderocol phase 3 programmes no relationship with any endpoint was seen even at 100% fT>4×MIC
Is there an upper limit?	In the TARGET trial, as reported by Berry and Kuti, mortality was higher at steady-state concentrations below 32 mg/L and above 96 mg/L, which suggests one
Does prolonging the infusion save lives?	BLING III: 90-day mortality 24.9% against 26.8% in 7,031 patients, unadjusted odds ratio 0.91 (0.81-1.01, P = 0.08), prespecified adjusted 0.89 (0.79-0.99, P = 0.04). Companion meta-analysis of 18 trials and 9,104 participants: relative risk 0.86 (credible interval 0.72-0.98)
Should a single MIC drive a dose change?	Mouton and colleagues say no: a log2 standard deviation of 0.3-0.5 of a dilution within one laboratory and 0.5-1 or more between laboratories, and their stated conclusion is that using an individual MIC value to modify a dosing regimen is not justified

Three indices, and the error of matching the wrong one

Antibacterials do not all kill the same way, and the pharmacodynamic index that predicts activity differs accordingly. The beta-lactams are time-dependent: their killing saturates a few multiples above the MIC, so what matters is how long free drug stays above it. The aminoglycosides are concentration-dependent: killing accelerates with concentration and continues for a while after the concentration falls, so what matters is how high the peak went. Vancomycin and the fluoroquinolones sit in between and are judged on total daily exposure, the area under the curve divided by the MIC. Three drugs, three indices, and three different levers — the interval and the infusion time for the first, the size of the single dose for the second, the total daily dose for the third.

Applying the wrong index is not a harmless approximation; it points dosing in the wrong direction. Judging a beta-lactam on its peak suggests giving larger doses less often, which is exactly the change that lowers fT>MIC. Judging an aminoglycoside on time above the MIC suggests frequent small doses, which both lowers the peak and raises the trough — the trough being where aminoglycoside nephrotoxicity and ototoxicity accumulate. Judging vancomycin on a trough, which is what guidelines asked for until 2020, pushed exposure higher than cure required and raised acute kidney injury without improving outcome, which is why the 2020 consensus guideline abandoned it.

Two things have to be read off every target before it is used. The first is the basis: free drug or total. EUCAST’s rationale documents state their targets as fAUC/MIC and every beta-lactam fT>MIC figure is free drug, while the vancomycin 400 to 600 window and Forrest’s quinolone AUC of 125 are total drug. For ciprofloxacin at a fraction unbound of 70 to 80% that distinction moves a number by a quarter; for a 95%-bound agent it moves it twentyfold. The second is the endpoint and the model the target was derived against. Almost every PK/PD target in use was derived in animal infection models and in-vitro systems — most often the neutropenic murine thigh model — and the human outcome data behind them are thin: the targets are mostly from animal models and in-vitro systems with limited human outcome validation. Susceptibility testing works in doubling dilutions, so a reported MIC of 1 mg/L means growth was inhibited at 1 mg/L and not at 0.5 mg/L — the true value lies somewhere in that bracket, and inter-laboratory reproducibility of one doubling dilution either way is the accepted standard. Any ratio with an MIC in it therefore carries at least a twofold uncertainty, which is larger than the precision in the rest of the calculation. A stasis target is roughly half the exposure of a 2-log10 kill target for the same drug and organism, the neutropenic mouse has no neutrophils to help, and the organism matters — EUCAST’s ciprofloxacin figures for S. pneumoniae are less than half those for S. aureus.

This page also states what it will not do. Probability of target attainment is a Monte Carlo quantity: the proportion of a simulated population reaching a target at each MIC, computed over thousands of draws from a population pharmacokinetic distribution. This site’s calculator engine has no loops, no arrays and no random numbers, so it cannot produce one, and a fitted curve standing in for a simulation would be a confident wrong answer of exactly the kind that is hardest to spot. Read a PTA curve in the paper that drew it, note which target and which population PK model it used, and use these pages for the closed-form arithmetic underneath. This page does not hold a breakpoint table. EUCAST and CLSI publish them, they are revised two or three times a year, and the laboratory that issued your report has already applied its own version. Take the MIC and the interpretive category from the report; this page does the arithmetic that sits on top of them. Susceptible, intermediate and resistant are laboratory interpretations against a versioned breakpoint table, and this page cannot and does not assign one. The regimen is the prescriber’s decision and the interpretive category is the laboratory’s.

Frequently asked questions

Which PK/PD index applies to which antibiotic class?

Beta-lactams are time-dependent and their index is the percentage of the dosing interval with free drug above the MIC. Aminoglycosides are concentration-dependent, classically judged on Cmax/MIC. Vancomycin and the fluoroquinolones are exposure-dependent and judged on AUC/MIC. EUCAST now sets aminoglycoside and quinolone breakpoints against fAUC/MIC for both.

Why do the published targets differ so much for the same class?

Because they are different endpoints in different models against different organisms, and some are free drug while others are total. A stasis target needs roughly half the exposure of a 2-log10 kill target, and EUCAST’s ciprofloxacin figures for S. pneumoniae are less than half those for S. aureus. A target quoted without its endpoint, model, organism and basis is not usable.

How good is the human evidence for these targets?

Berry and Kuti’s 2022 review sets the murine thresholds against the human data: DALI found patients below 50% fT>MIC 32% less likely to have a positive outcome, but EXPAT found neither 100% fT>MIC nor 100% fT>4×MIC significantly associated with 30-day survival, and two phase 3 programmes found no exposure-response relationship at all.

Can you calculate the probability of target attainment?

No, and the reason is worth stating. PTA is the proportion of a simulated population reaching a target at a given MIC, computed by Monte Carlo simulation over thousands of draws from a population pharmacokinetic distribution. This site’s calculator engine has no loops, no arrays and no random numbers. A fitted curve standing in for a simulation would look authoritative and be wrong, so none is offered.

What about agents not on the list?

They are deliberately absent. This page states only targets read in a named source, and daptomycin, linezolid, the polymyxins, fosfomycin and the azoles were not sourced for it. EUCAST publishes a rationale document per agent, each stating the index, the model and the organism, and EUCAST SOP 6.4 makes those documents free to re-use with attribution.

Does this page tell me whether an organism is susceptible?

No. This page cannot say whether an organism is susceptible: that is a laboratory interpretation against a versioned breakpoint table, and the laboratory that issued the report has already made it.

Related calculators

References

Berry AV, Kuti JL. Pharmacodynamic thresholds for beta-lactam antibiotics: a story of mouse versus man. Front Pharmacol. 2022;13:833189. Open access; source for the murine %fT>MIC thresholds and for the human studies set against them.
Bland CM, Pai MP, Lodise TP. Reappraisal of contemporary pharmacokinetic and pharmacodynamic principles for informing aminoglycoside dosing. Pharmacotherapy. 2018;38(12):1229-1238. Source for the Moore 1987 and Kashuba 1999 figures quoted on this page, which it tabulates.
European Committee on Antimicrobial Susceptibility Testing. Amikacin rationale document, version 3.1, September 2024. Source for the fAUC/MIC targets and the fraction unbound; attributes the murine data to the USCAST evaluation (Ambrose et al. 2019).
European Committee on Antimicrobial Susceptibility Testing. Ciprofloxacin rationale document, version 2.0, 1 January 2021. Source for the fAUC/MIC targets, the fraction unbound, and the total-drug AUC:MIC of 125 that it attributes to Forrest et al. 1993.
Rybak MJ, Le J, Lodise TP, et al. Therapeutic monitoring of vancomycin for serious methicillin-resistant Staphylococcus aureus infections: a revised consensus guideline and review of ASHP, IDSA, PIDS and SIDP. Am J Health-Syst Pharm. 2020;77(11):835-864.
Societe Francaise de Pharmacologie et de Therapeutique and Societe Francaise d’Anesthesie et de Reanimation. Recommandations de Pratiques Professionnelles: optimisation du traitement par beta-lactamines chez le patient de soins critiques. 2018; published in English as Guilhaumou R et al, Crit Care. 2019;23:104. Recommendation R2.2.
Dulhunty JM, Brett SJ, De Waele JJ, et al; BLING III Study Investigators and the ANZICS Clinical Trials Group. Continuous vs intermittent beta-lactam antibiotic infusions in critically ill patients with sepsis: the BLING III randomized clinical trial. JAMA. 2024;332(8):629-637.
Abdul-Aziz MH, Hammond NE, Brett SJ, et al. Prolonged vs intermittent infusions of beta-lactam antibiotics in adults with sepsis or septic shock: a systematic review and meta-analysis. JAMA. 2024;332(8):638-649.
Mouton JW, Muller AE, Canton R, Giske CG, Kahlmeter G, Turnidge J. MIC-based dose adjustment: facts and fables. J Antimicrob Chemother. 2018;73(3):564-568. Source for the log2 variability figures, the ISO 20776-2 criterion and the stated conclusion that individual MIC-based dose adjustment is not justified.
European Committee on Antimicrobial Susceptibility Testing. EUCAST SOP 6.4: Operation of websites, 6 May 2025. States that EUCAST retains copyright of its documents and data and that all are freely available for re-use with reference to the website and without resale.
Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing. 35th ed. CLSI supplement M100-Ed35. Wayne, PA: CLSI; January 2025. A commercial standard whose front matter requires express written consent for any reproduction; cited here, never reproduced or paraphrased.

Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/