AICAR has more human dose data than almost anything else in this catalogue — one phase 3 trial, over a thousand people dosed, posted results. It is worth reading precisely, because the headline number usually quoted about it is roughly double the truth, and the outcome is usually not quoted at all.
It is not a peptide
AICAR — acadesine, AICA riboside — is a nucleoside analogue and an activator of AMP-activated protein kinase. It is recorded here for the same reason as SLU-PP-332 and 5-amino-1MQ: it is sold in the same market, beside the same compounds, to the same buyers.
The largest trial, and the number that gets misquoted
NCT00872001 — RED-CABG. Sponsor Merck Sharp & Dohme. Phase 3, high-risk patients undergoing coronary artery bypass graft surgery. April 2009 to October 2010. Terminated.
| Acadesine | Placebo (normal saline) | |
|---|---|---|
| Started | 1,536 | 1,544 |
| Completed | 1,400 | 1,413 |
| Did not complete | 136 | 131 |
Enrolled: 3,080. Dosed with acadesine: 1,536.
This site's own keyword map carried the figure "over 3,100 people dosed". That counts the randomised total, and about half of it received saline. The correction matters beyond one row: in a placebo-controlled trial, enrolment and exposure are different numbers, and the registry publishes both.
The dose
| Amount | 42 mg/kg |
|---|---|
| Dilution | Into 500 mL normal saline |
| Route | Intravenous infusion |
| Duration | ≈ 7 hours |
| Rate | 0.1 mg/kg/min (1.2 mL/min of the 500 mL bag) |
| Start | Within ≈ 30 min before induction of anaesthesia |
| Plus | 5 μg/mL added to the cardioplegia solution |
For an 80 kg adult that is 3,360 mg in a single administration, dripped in over seven hours, in an operating theatre, with a second dose route through the solution used to arrest the heart.
The comparison worth making is with the unit the compound is sold in. A typical research-market vial or capsule is 50 mg. The trial dose is more than sixty times that, and it is not given as a bolus — the seven-hour infusion at a fixed rate per kilogram per minute is part of the specification, not a convenience of the surgical setting.
The result
Posted results, primary composite — all-cause death, non-fatal stroke, or need for mechanical support for severe left ventricular dysfunction, through post-operative day 28:
| Outcome | Acadesine | Placebo |
|---|---|---|
| Primary composite (all-cause death version) | 4.9% | 4.9% |
| Composite with cardiovascular death | 4.8% | 4.7% |
| Component | 1.7% | 1.6% |
| Component | 1.7% | 1.7% |
| Component | 2.2% | 2.3% |
Identical on the primary endpoint, and within a rounding place of each other on every other posted measure. The trial was terminated.
This is a rare thing in this catalogue and it should be said plainly: the compound was tested properly, at scale, against placebo, with results posted — and it did not separate from saline. Most compounds here have no such test. AICAR's problem is not an absence of evidence.
The only attempt to go higher stopped for toxicity
NCT01813838 — GFM-Acadesine. Groupe Francophone des Myélodysplasies. Phase 1/2 in high-risk myelodysplastic syndromes, acute myeloid leukaemia with 20–30% marrow blasts, and chronic myelomonocytic leukaemia type 2 not responding to azacitidine or decitabine.
The design was a classic escalation: three patients at 140 mg/kg/day, then three more at 210, then 315, with expansion cohorts on any toxicity signal.
It is recorded as terminated with 5 participants actually enrolled, and the registry's why-stopped field reads:
Renal toxicity
That is the entire upper end of this compound's human dose range, and it is a stopping point rather than a ceiling. Nothing in the record describes a dose above 42 mg/kg that was given to a meaningful number of people without a problem.
What the rest of the registry holds
| Study | n | Setting |
|---|---|---|
| NCT00559624 | 40 (estimated) | B-cell chronic lymphocytic leukaemia, open-label dose escalation, 4-hour IV infusions |
| NCT00004314 | 2 | Phase 2 pilot, AICA-ribosiduria |
| NCT06845501 | 10 (recruiting) | Purine supplementation in AICA-ribosiduria |
A search for acadesine returns 4 records; a search for AICAR returns 6, and most of those use the term for a laboratory reagent or a metabolic pathway rather than an administered drug — sunitinib and atrial trabeculae contractility, pemetrexed in lung cancer, inflammation and obesity.
Every registered human administration found here is intravenous. No registered trial gave the compound by subcutaneous injection, and none gave it by mouth at a dose intended to produce a systemic effect. The oral and injectable products sold in this market have no human route precedent at all.
Banned, and null, at the same time
AICAR is on the World Anti-Doping Agency Prohibited List as a metabolic modulator, on the strength of preclinical work suggesting that AMP-activated protein kinase activation could produce endurance adaptations without training.
Both facts are accurate and they are not in tension: a compound can be prohibited on the basis of a plausible mechanism in animals while its one large human trial, in a different indication, returns a result identical to saline. A ban is a statement about intent and plausibility. A phase 3 result is a measurement.
Method
The ClinicalTrials.gov v2 API was queried on 21 September 2026 for AICAR and acadesine, and each administering record was opened individually — including its results section, its enrolment type and its why-stopped field — rather than read from a result list. Participant-flow figures are taken from the posted results of NCT00872001 and are quoted as started and completed per arm, not as enrolment. Dose text is quoted from the intervention descriptions of the records themselves. Related records: the dosage chart grades every compound here by the tier of source behind its figure, how to read a dosing claim sets out why an infusion dose and a vial are not the same kind of number, and research use only explained covers what the label on a research-market product does and does not assert. Nothing on this page is a dosing instruction.