Oxiracetam

Oxiracetam is Piracetam with a hydroxyl group at the 4-position of the pyrrolidone ring. It was developed in Italy and has been prescribed for cognitive impairment for decades; in South Korea it was widely prescribed and reimbursed, with annual spending reaching 22.8 billion KRW — about €17 million — in 2021 [4]. The 2010 survey of the piracetam family already recorded it as out of clinical use in the West [5].

That single hydroxyl changes more than it looks like it should. It makes the molecule more polar than piracetam, it creates a stereocenter — so oxiracetam exists as two enantiomers, and the (S) form is the more active one [6] — and, compared like for like in animals, it makes the cholinergic effects both stronger and longer-lasting [7].

What the research found. The compound's story is unusually clean for a nootropic, because someone finally ran the big trial. The early trials were small, positive and old: 40 to 65 patients each, published between 1987 and 1992, reporting benefit in mild dementia [1][2][3]. A contemporaneous 24-patient American trial in Alzheimer's disease found no improvement in the treated group or in any individual patient [8]. Then in 2026 a 500-patient randomized trial of 800 mg twice daily for 36 weeks in post-stroke cognitive impairment reported no difference from placebo on either co-primary endpoint — and South Korea's regulator had already recommended suspending prescriptions in January 2023 [4].

The preclinical literature is in better shape than the clinical one, and it points at glutamate rather than at membranes.

The standard account is that oxiracetam stimulates phospholipid biosynthesis, stabilizes neuronal membranes and promotes cholinergic transmission [4] — essentially the membrane-fluidity story told about Piracetam, with a cholinergic addendum.

Those are proposed systems-level mechanisms, not a demonstrated binding profile. Reviews of the class describe oxiracetam's mechanism as unaccepted, and the modern clinical paper repeats the phospholipid and cholinergic proposals without identifying a selective receptor target [4][9]. Reversal of scopolamine amnesia supports a cholinergic downstream effect; it does not show that oxiracetam binds a muscarinic receptor [10].

But there is real receptor pharmacology here, and it is glutamatergic. Micromolar oxiracetam enhanced AMPA-stimulated calcium influx in cultured cerebellar granule cells and raised the maximal density of [3H]AMPA binding sites in rat cortical synaptic membranes. The specificity is the interesting part: oxiracetam changed neither the kainate nor the NMDA calcium response, and did not touch phosphoinositide hydrolysis driven by quisqualate or by a metabotropic agonist. It is a positive modulator of AMPA-sensitive receptors and nothing else in that panel [11][12].

A 2026 study worked out what that modulation does to the channel. In APP/PS1 transgenic mice, oxiracetam increased the frequency and amplitude of spontaneous, AMPA-receptor-dependent evoked and miniature excitatory postsynaptic currents in hippocampal neurons. Kinetically, it slowed the desensitization of both the GluA1 and the GluA2 subunit without changing inactivation, and sped the recovery of desensitized GluA2 while leaving GluA1 recovery alone. The same animals showed better behavioral performance, less CA1 neurodegeneration, and restored hippocampal NeuN, GAP43, MAP2 and synaptophysin [13]. The subunit-selective recovery effect is a more concrete mechanism than anything in the phospholipid story.

Transmitter release, from somewhere other than the terminal. In superfused rat hippocampal slices, 0.01–1 µM oxiracetam increased the potassium-evoked overflow of tritiated D-aspartate — a non-metabolized glutamate analog — and, less strongly, of acetylcholine. GABA, norepinephrine and serotonin were unaffected at those concentrations. Two details constrain the interpretation. Higher concentrations (10–100 µM) did nothing, and if anything tended to inhibit release, giving the bell-shaped curve that recurs across this drug family. And the effect disappeared entirely in isolated synaptosomes, so whatever oxiracetam acts on is not inside the releasing nerve ending; it needs the intact circuit of the slice [14].

Protein kinase C. Oxiracetam given in vivo, or applied to cortical slices at nanomolar concentrations, moved protein kinase C from the soluble to the particulate fraction — the translocation step that activates the enzyme — in rat cortex and hippocampus, with downregulation following a few hours later. In cortex, that translocation was blocked by the NMDA antagonist AP-5 and by scopolamine, but not by the AMPA antagonist CNQX or the metabotropic antagonist L-AP3, which places the PKC effect downstream of NMDA and muscarinic signaling rather than of the AMPA modulation above [15]. The same paper found alpha-GPC does the same thing, and concluded PKC activation may be a shared endpoint across chemically unrelated cognition enhancers.

The cholinergic arm, measured directly. Oxiracetam at 100 and 300 mg/kg increased acetylcholine utilization in rat cortex and hippocampus — the rate at which the transmitter is consumed, measured as the fall in acetylcholine after synthesis is blocked with hemicholinium — while leaving steady-state acetylcholine unchanged. Repeated daily dosing at 100 mg/kg raised hippocampal high-affinity choline uptake by 31%, and a single 300 mg/kg dose still produced a 40% increase three hours later, at a point when the same dose of piracetam had no effect left. Oxiracetam also reversed the active-avoidance deficit caused by blocking acetylcholine synthesis. The authors read this as an action on the septohippocampal cholinergic projection, with a weaker effect on the cortical network [7] — which lines up neatly with where the drug physically ends up.

What is better established is where the drug goes. Radiolabeled oxiracetam given to rats orally or intra-arterially crosses into the brain unmetabolized and concentrates in the septum, then the hippocampus, with less in cortex and striatum. The same distribution appears when it is injected directly into the ventricles, so the regional preference is a property of the molecule rather than of the route. Delivered into the ventricles at brain concentrations matching an active systemic dose, it dose-dependently antagonized scopolamine-induced amnesia [10].

Oxiracetam is barely metabolized in humans: more than 90% of an intravenous dose comes back unchanged in urine [16]. Whatever it does, it does as itself.

Oxiracetam is sold as the racemate, but the two enantiomers have been separated. (S)-oxiracetam, the more active isomer, has been through phase I in Chinese volunteers by both routes.

Orally, (S)-oxiracetam 400–2000 mg was rapidly absorbed with Tmax 0.75–1.00 h and a half-life of 6.1–6.6 h. No chiral inversion occurred; 55% was excreted unchanged in urine and 36% in feces. Exposure rose in proportion to dose from 400 to 1600 mg but flattened between 1600 and 2000 mg. Steady state was reached on day 5. Food left AUC unchanged but pushed Tmax to 3 h [6].

Intravenously, single doses of 2, 4 and 8 g gave Cmax of 111, 231 and 353 µg/mL, with about 60% urinary excretion of unchanged drug and an accumulation index of 1.05 after seven daily 4 g infusions [17].

Oxiracetam is unusually easy to tolerate. A year of 1600 mg a day in elderly demented outpatients produced no serious adverse events and no changes in routine laboratory tests [2]. In the 2026 post-stroke trial, adverse events occurred in 41.0% of the oxiracetam group and 34.9% of the placebo group — a numerical excess that did not reach significance — with serious adverse events causing discontinuation in 1.2% versus 0.4%, and no safety signal attributable to the drug [4].

Both phase I studies of (S)-oxiracetam reported mild-to-moderate, dose-independent adverse events and no serious events at up to 2 g orally or 8 g intravenously [6][17].

The safety profile is not the problem. The efficacy is.