Citicoline is the rare nootropic that has been tested seriously — and that is exactly what makes it awkward. It is a registered prescription medicine across much of Europe, Latin America and Asia, sold in the United States as a dietary supplement under the branded ingredient Cognizin, and it has been through two of the largest, best-run neuroprotection trials ever attempted. Both came back null. What remains is a real but softer body of evidence in post-stroke and vascular cognitive impairment, a handful of small manufacturer-linked trials in healthy people, and a separate and genuinely interesting eye literature. Here is where each of those lines actually falls.
What citicoline actually is — and why it is not just choline
Citicoline is cytidine-5′-diphosphocholine (CDP-choline), a naturally occurring intermediate in the Kennedy pathway — the route every cell in your body uses to build phosphatidylcholine, the dominant phospholipid in cell membranes, including the membranes of neurons. In that pathway choline is phosphorylated to phosphocholine, phosphocholine is joined to CTP to make CDP-choline, and CDP-choline donates its phosphocholine head group to diacylglycerol to produce phosphatidylcholine. Citicoline is that middle step, taken as a pill.
The complication is that the molecule you swallow is not the molecule that reaches your brain. Oral CDP-choline is hydrolysed in the gut wall into its two components, and the cell reassembles it intracellularly afterwards. What that looks like in a human was measured directly: in twelve fasting subjects given 500, 2,000 or 4,000 mg of oral CDP-choline, plasma choline rose in a dose-related way and plasma uridine rose by roughly 70–90% after the 500 mg dose, while cytidine was never reliably detectable at any dose. The authors’ conclusion is the one worth carrying: in humans the circulating substrates are uridine and choline — not cytidine and choline as in rats.[1] A great deal of rodent citicoline literature quietly assumes the rat route.
That is the honest mechanistic difference from the other cholinergics on the shelf. Alpha-GPC and plain choline salts deliver choline and stop there; the pitch for citicoline is that it delivers choline plus a pyrimidine, supplying both halves of the membrane-synthesis reaction rather than one.[2] It is a coherent story. It is also, as the next section shows, a story that did not survive its two hardest tests.
Acute stroke: the whole literature, pooled
ICTUS was run because earlier, smaller trials had looked promising in pooled analysis.[3] That pattern — encouraging small trials, null large trial — is the single most common shape in neuroprotection research, and citicoline is a textbook example of it.
The most independent read of the full record is the 2020 Cochrane review, which identified 10 randomised trials in 4,281 participants. It found little to no difference versus placebo on every outcome it examined: all-cause mortality (17.3% versus 18.5%; RR 0.94, 0.83–1.07), disability or dependence on the Rankin scale (RR 1.11, 0.97–1.26), functional recovery on the Barthel Index (RR 1.03, 0.94–1.13), neurological function on the NIHSS (RR 1.08, 0.96–1.21) and serious cardiovascular adverse events (RR 1.04, 0.84–1.29). All of it was graded low-certainty; the reviewers judged every included trial at high risk of bias, noted that drug companies sponsored six of the ten, and recorded that adverse events were poorly reported, so harms may have been underestimated. A pre-planned trial sequential analysis suggested no further trials are needed for the primary outcomes.[5] An earlier independent meta-analysis of 12 human trials reached the same place, finding no significant difference between citicoline and placebo on neurological, domestic-adaptation or cognitive outcomes.[7]
Read from the other side, in fairness: a 2016 systematic review and formal meta-analysis of randomised, double-blind, placebo-controlled trials did report a benefit for citicoline in acute ischaemic stroke.[6] That paper is worth knowing about and worth discounting for a specific reason — its first author is a medical-department employee of Ferrer, the company that markets citicoline and funded ICTUS, and several co-authors were ICTUS investigators.[3][13] When an independent Cochrane review and an industry meta-analysis of overlapping trials disagree this sharply, the design and the disclosure are the tiebreak, and both point the same way.
Vascular cognitive impairment: where the better case lives
This is the indication that keeps citicoline alive as a serious drug, and it deserves to be reported accurately rather than dismissed by association with the stroke failures.
The most-cited study randomised 347 patients six weeks after a first-ever ischaemic stroke to citicoline 1 g/day for 12 months or to usual treatment. At 6 and 12 months the citicoline group did better on attention-executive function (OR 2.379, 95% CI 1.269–4.462, p = 0.007 at 12 months) and temporal orientation (OR 2.155, 1.017–4.566, p = 0.045). Functional outcome favoured citicoline numerically (modified Rankin ≤2 in 57.3% versus 48.7%) but not significantly (p = 0.186), and the authors were explicit that large confirmatory trials were needed.[8] A two-year follow-up of the same cohort reported better cognitive status and quality of life on citicoline.[9]
The design caveat is not a technicality. Both reports are open-label — the control arm received no placebo and everyone knew who was being treated. For neuropsychological testing, where effort, expectation and examiner knowledge all move the score, unblinded allocation is the weakest available design. The IDEALE study has the same problem in a milder population: 349 elderly Italians with mild vascular cognitive impairment, 265 given citicoline 500 mg twice daily, 84 untreated controls, open and multicentre. MMSE held steady in the treated group over nine months and diverged significantly from controls — but ADL and IADL scores did not differ, meaning the day-to-day functional benefit was not demonstrated.[10] The CITIRIVAD study, often cited alongside them, is a retrospective case-control review of 174 outpatients on rivastigmine with or without citicoline — useful as a signal, not as evidence of effect.[11]
Older dementia-adjacent data are similarly mixed. A Cochrane review of CDP-choline in cognitive and behavioural disturbances in the elderly found no significant benefit on attention, significant short-to-medium-term effects on memory and behaviour, stronger effects on global clinical impression, and good tolerability — while noting the trials were heterogeneous in dose and criteria and mostly ran only 20–30 days.[12] A frequently referenced review of citicoline in cognitive impairment describes it, accurately, as “an old drug with new perspectives” — and with doubts.[13]
Healthy cognition: small, short, and mostly paid for
Almost everyone buying citicoline in the US is a healthy adult wanting focus. The trials that support that use are the weakest in this monograph, and the funding matters.
The adolescent-male trial gave 75 healthy boys 250 or 500 mg of Cognizin citicoline or placebo for 28 days and reported improved attention (p = 0.02) and psychomotor speed (p = 0.03).[14] The older-adult trial gave 100 men and women aged 50–85 with age-associated memory impairment 500 mg/day of Cognizin or placebo for 12 weeks and reported significantly greater improvement in episodic memory and composite memory.[15] Both are properly randomised and placebo-controlled, which is more than most shelf nootropics can claim. Both also carry the same two problems. Author affiliations include Kyowa Hakko Bio and Kyowa Hakko USA — the manufacturer of Cognizin in both papers.[14][15] And in the older-adult trial, the memory results that get quoted are labelled in the paper itself as secondary outcomes; the abstract reports no significant finding on the primary.[15]
Searching from the opposite side for a clean, independent, adequately powered trial of citicoline alone for cognition in healthy adults turns up near-misses rather than evidence. The most-quoted “citicoline improves concentration and working memory” study tested a citicoline-caffeine beverage — caffeine has well-established acute effects on exactly those measures, so the design cannot separate the two.[16] That is a confounded trial, not a citicoline trial. The honest summary for healthy users is that the effect, if real, is small, short-term, and has never been shown by anyone without a commercial stake in the answer.
Evidence by indication, graded
| Use | Best evidence | Grade |
|---|---|---|
| Acute ischaemic stroke | ICTUS: 2,298 patients, stopped for futility (OR 1.03). Cochrane: 10 trials, 4,281 patients, no difference on any outcome | Negative — tested properly and failed |
| Traumatic brain injury | COBRIT: 1,213 patients, phase 3, no difference at 90 or 180 days | Negative — tested properly and failed |
| Post-stroke / vascular cognitive impairment | 347-patient open-label randomised study: attention-executive and temporal orientation favoured citicoline at 12 months; IDEALE (n = 349) open-label | Weak-positive — unblinded designs, subjective endpoints |
| Cognitive/behavioural disturbance in the elderly | Cochrane: benefit on memory and behaviour, none on attention; trials mostly 20–30 days | Low certainty — short and heterogeneous |
| Memory in healthy older adults | 100-person 12-week RCT: episodic and composite memory improved — as secondary outcomes; manufacturer-affiliated authors | Very weak — industry-linked, secondary endpoints |
| Attention in healthy young people | 75 adolescent males, 28 days, manufacturer-affiliated authors | Very weak — one small industry trial |
| Glaucoma / retinal neuroprotection | Randomised placebo-controlled cross-over trial (vision-related quality of life, p = 0.041); electrophysiology pilots | Preliminary but real — a separate literature |
| Longevity or healthspan | No trial has ever tested it | Untested |
The eye literature, briefly
Citicoline in glaucoma is a genuinely separate research programme, built on the idea that glaucoma is partly a neurodegeneration of the retinal ganglion cells and optic pathway rather than purely a pressure problem. The strongest entry is an international, multicentre, randomised, double-masked, placebo-controlled cross-over trial of citicoline oral solution 500 mg/day in open-angle glaucoma, in which the pre-specified primary outcome — change in the Visual Function Questionnaire-25 composite score — reached significance at p = 0.0413, with a larger effect in patients whose vision-related quality of life was worse at baseline.[18] That is a real randomised result on a pre-specified primary endpoint, which is more than the cognitive literature offers. It is also a quality-of-life questionnaire, not visual-field preservation, and every listed author disclosed consultancy for the company marketing the product.[18]
Below it the designs weaken quickly. The often-cited progression study followed 41 patients with progressing glaucoma on citicoline oral solution for two years and reported the visual-field decline rate slowing from −1.1 to −0.15 dB/year — but the comparison is against each patient’s own prior trajectory, with no concurrent placebo group.[17] And the liposomal eye-drop electrophysiology work is an explicitly labelled pilot in 12 patients, single-arm, funded by the manufacturers, reporting changes in pattern electroretinogram amplitude and visual-evoked-potential timing rather than in what people can see.[19] Interesting; not yet a reason to buy citicoline for your eyes.
Dose, tolerability and cost
Supplement trials cluster at 250–500 mg/day, and both Cognizin trials used doses in that band.[14][15] Clinical use runs far higher: the vascular cognitive-impairment studies used 1,000 mg/day,[8][10] ICTUS used 2,000 mg/day,[3] COBRIT used 2,000 mg/day for 90 days,[4] and the Cochrane trials ranged from 500 to 2,000 mg/day.[5] A practical implication people miss: a 250 mg capsule is a fraction of what the serious trials used, so a consumer dose is not “the studied dose” simply because the molecule is the same.
On tolerability the record is unusually reassuring. Cochrane found no difference in serious cardiovascular adverse events versus placebo across three trials,[5] ICTUS reported no significant differences in safety variables or adverse-event rates in 2,298 patients,[3] the 12-month post-stroke study saw adverse events in 4 of 172 treated patients without discontinuation,[8] IDEALE recorded none,[10] and the elderly Cochrane review described the drug as well tolerated.[12] Thousands of patients have taken it for weeks to months at doses far above supplement levels without a safety signal emerging. That is a meaningful distinction from alpha-GPC, whose thinner human record carries an unresolved observational stroke-risk association.
On price, citicoline is a mid-tier supplement rather than a cheap one: a branded Cognizin product at 250–500 mg/day typically runs meaningfully more per month than bulk choline bitartrate, and the premium buys the branded ingredient and the two trials attached to it. Whether that is worth paying depends entirely on how much weight you put on two small manufacturer-linked studies.
The regulatory split, stated plainly
Citicoline occupies two categories at once, and the mismatch drives most of the confusion around it. In much of Europe, Latin America and Asia it is a prescription medicine with approved neurological indications — the COBRIT investigators recorded that it was approved for use in traumatic brain injury in 59 countries at the time they ran the trial,[4] and the Cochrane reviewers noted it is a frequently prescribed drug for cognitive impairment in several European countries.[12] Ferrer’s pharmacological and clinical review, the standard reference work on the compound, is written from that drug-development perspective throughout.[2]
In the United States it is none of those things. It is a dietary supplement, most often sold as the branded ingredient Cognizin made by Kyowa Hakko,[14] not FDA-approved for any indication, and not permitted to be marketed as a treatment for stroke, brain injury, dementia or glaucoma. Two consequences follow. A US buyer gets no regulatory guarantee that the capsule contains what the label says or that any medical claim about it has been reviewed. And — the point that is almost always inverted in marketing — being an approved drug abroad is not evidence that it works. Approvals in several of those markets predate ICTUS and COBRIT entirely.
The honest bottom line
Citicoline is a well-characterised molecule with a coherent mechanism, an unusually clean safety record at high doses, and a legitimate place in prescribing practice in dozens of countries. It is also a compound whose two most rigorous tests — 3,511 patients between them, in the two indications it was most expected to work in — both returned nothing, and whose independent pooled analyses in stroke find no benefit at low certainty.[3][4][5][7]
What survives is narrow and worth stating precisely. There is a plausible, unconfirmed signal in post-stroke and mild vascular cognitive impairment, held back by open-label designs that would not be accepted for a new drug today.[8][10] There is a preliminary randomised result in glaucoma quality of life that deserves replication by someone without a commercial interest.[18] And there is a very thin case for healthy cognition, resting on two small trials whose author lists include the ingredient manufacturer.[14][15] If you are a healthy adult buying citicoline for focus, you are buying the weakest part of the file, and the strongest part of the file is the part that failed. Anyone comparing options in this category should read our alpha-GPC and lion’s mane monographs alongside this one — the pattern of small positive trials and absent large ones repeats across the whole aisle.
This article is research information, not medical advice. In the United States citicoline is sold as a dietary supplement and is not FDA-approved to treat, prevent or manage stroke, traumatic brain injury, dementia, cognitive decline or glaucoma; where it is a prescription medicine abroad, that status predates the large null trials described above. Nothing here should be used to delay emergency treatment for stroke or head injury, to substitute for intraocular-pressure-lowering therapy in glaucoma, or to replace evaluation of new memory problems. If you take medication affecting acetylcholine, are being treated for glaucoma or cerebrovascular disease, or are pregnant or breastfeeding, talk to a licensed clinician before starting citicoline.