Alpha-GPC (choline alfoscerate) and citicoline (CDP-choline) both supply choline for acetylcholine and cell membranes. Each has positive randomized trials in Alzheimer's disease or mild cognitive impairment, and a small 2025 meta-analysis favoured alpha-GPC over citicoline in dementia. Citicoline failed in the two largest trials of either compound, in acute stroke and traumatic brain injury. Evidence in healthy people is thin, and EFSA found a citicoline memory claim unsubstantiated. A Korean cohort of 12 million adults linked alpha-GPC prescriptions to higher stroke risk, which remains unexplained.
Summary drafted from this entry’s sources; editorial review pending.
01Key questions
Does alpha-GPC improve memory?
In patients, yes, modestly: a 261-patient Alzheimer's trial and a 100-patient mild cognitive impairment trial both found better ADAS-Cog scores than placebo. In healthy young men, two small short studies found mixed results.561011
Does citicoline work?
A Cochrane review found some short-term benefit for memory and behaviour in older people with chronic brain disorders. But the ICTUS stroke trial (2,298 patients) and the COBRIT brain-injury trial (1,213 patients) were both null, and in 2024 EFSA found a citicoline memory claim unsubstantiated.2349
Is alpha-GPC better than citicoline?
In dementia, a 2025 meta-analysis of three trials with 358 patients found alpha-GPC better on a geriatric clinical scale but not on memory tests. There is no head-to-head comparison in healthy people.7
Does alpha-GPC raise stroke risk?
Possibly. Among 12 million South Koreans, alpha-GPC users had a 43% higher 10-year stroke risk after matching, and in mice alpha-GPC promoted atherosclerosis via the gut metabolite TMAO. A later Korean study in mild cognitive impairment found lower dementia conversion and no stroke excess. Both are observational.11213
What would change this rating?
A long, independently funded trial tracking both cognition and cardiovascular events. That would settle whether the dementia benefit is real and whether the stroke signal is.
Alpha-GPC (choline alfoscerate) and citicoline (CDP-choline) are the two choline donors nootropic users argue about. Both are prescription drugs in parts of Europe and Asia and supplements in the US. Both have positive randomized trials in dementia and mild cognitive impairment, and in a small head-to-head analysis alpha-GPC came out ahead. Citicoline failed in two very large trials, in acute stroke (2,298 patients) and traumatic brain injury (1,213 patients). Evidence in healthy people is thin: one industry-sponsored citicoline trial in older adults with memory complaints and two small, short alpha-GPC studies. And alpha-GPC carries a population-level stroke signal from 12 million South Koreans that nobody has yet explained away.
02What they are
| Alpha-GPC | Citicoline | |
|---|---|---|
| Chemistry | L-alpha-glycerylphosphorylcholine, a phospholipid breakdown product | Cytidine 5'-diphosphocholine, the intermediate the body uses to build phosphatidylcholine |
| What it delivers | Choline for acetylcholine and membrane synthesis | Choline plus cytidine (converted to uridine), for acetylcholine and membrane phospholipids |
| Drug status | A prescription drug in South Korea, given to over 100,000 adults aged 50 and over in 2006 to 2008 alone (PMID 34817582) | Approved for acute stroke in some countries and for brain injury in 59; often prescribed for cognitive decline in Europe (PMID 15846601; PMID 22691567; PMID 23168823) |
| United States | Dietary supplement | Dietary supplement |
Swipe table →
Both are popular as add-ons to racetams and Noopept, on the theory that those drugs raise acetylcholine demand. That pairing has never been tested in a trial.
03Dementia and mild cognitive impairment: the positive trials
- Alpha-GPC in Alzheimer's disease: 261 patients with mild to moderate Alzheimer's, randomized to alpha-GPC or placebo for 180 days. ADAS-Cog improved by 3.2 points on alpha-GPC and worsened by 2.9 on placebo, with MMSE, global and behavioural scores also favouring the drug (De Jesus Moreno Moreno, 2003, PMID 12637119).
- Alpha-GPC in mild cognitive impairment: 100 adults in a Korean multicentre trial, 12 weeks, double-blind. Among the 74 who were analysed, ADAS-Cog fell 2.34 points on alpha-GPC and 0.97 on placebo, a 1.4-point difference (p=0.048), with no difference in adverse events (Jeon et al., 2024, PMID 39300341).
- Citicoline in chronic brain disorders of ageing: a Cochrane review found some evidence of benefit on memory and behaviour in the short to medium term, none on attention, and good tolerability. The trials were short and heterogeneous (Fioravanti and Yanagi, 2005, PMID 15846601).
- Head to head: pooling three randomized trials (358 patients with dementia), alpha-GPC beat citicoline on a geriatric clinical scale covering cognition, mood, apathy and social function, but not on memory or word fluency tests (Sagaro and Amenta, 2025, PMID 41426989). Amenta has published widely on alpha-GPC, and three small trials are a thin base.
04The big failures
- Acute stroke (ICTUS): 2,298 patients across Spain, Portugal and Germany, citicoline or placebo for six weeks. Stopped for futility; no difference in recovery at 90 days was detected (odds ratio 1.03, 95% CI 0.86 to 1.25). The trial was funded by the manufacturer (Dávalos et al., 2012, PMID 22691567).
- Traumatic brain injury (COBRIT): 1,213 patients at eight US trauma centres, citicoline or placebo for 90 days. No difference in functional or cognitive status at 90 or 180 days (Zafonte et al., 2012, PMID 23168823).
These are the two largest choline-donor trials ever run, and both were null. They do not rule out a small benefit in slow cognitive decline, but they end the idea that citicoline is a general brain-repair drug.
05Healthy people
In 100 healthy adults aged 50 to 85 with age-associated memory complaints, 12 weeks of a branded citicoline improved episodic memory and a composite memory score more than placebo. These were secondary outcomes (Nakazaki et al., 2021, PMID 33978188). In 2024 the European Food Safety Authority found a citicoline memory claim unsubstantiated: only that one trial in healthy people was positive, another at a higher dose was not, and dementia trials did not support it (PMID 38966137).
For alpha-GPC the healthy-person data are short and small. In a crossover study of 20 resistance-trained young men, a single dose produced larger gains in Stroop test scores than placebo across a session that included a bout of exercise, but no difference on the Flanker or N-back tests; the product was supplied by its manufacturer (Kerksick, 2024, PMID 39683633). In 48 college-age men given alpha-GPC, caffeine or placebo for seven days, psychomotor vigilance did not differ between groups (Marcus et al., 2017, PMID 29042830). No trial has tested longer-term use for cognition in healthy people.
06The alpha-GPC stroke signal
- The finding: among 12 million South Koreans aged 50 and over, the 108,877 who were prescribed alpha-GPC had a 43% higher 10-year risk of stroke after matching on known risk factors (hazard ratio 1.43, 95% CI 1.41 to 1.46), and risk rose with cumulative use (Lee et al., 2021, PMID 34817582).
- A mechanism: in atherosclerosis-prone mice, alpha-GPC was converted by gut bacteria into TMAO, shifted the microbiome and accelerated plaque (Wang et al., 2021, PMID 34948275).
- The counter-finding: among 508,107 Koreans newly diagnosed with mild cognitive impairment, alpha-GPC users were less likely to progress to Alzheimer's (hazard ratio 0.90) or vascular dementia (0.83), and stroke risk was lower in those who did not progress (Kim et al., 2025, PMID 40155153).
Both Korean studies are observational, and both are vulnerable to the same problem: doctors prescribe alpha-GPC to people they think are at risk of brain decline, and that group differs from everyone else. The mouse data make the stroke signal plausible rather than proven. If you have vascular risk factors, that unresolved question matters more than a few points on a memory scale.
07Claims audit
| Claim | Verdict | Why |
|---|---|---|
| "Improves memory in cognitive decline" | Supported, modestly | Positive randomized trials for both in dementia and MCI; short and mostly small |
| "Repairs the brain after stroke or injury" | Refuted for citicoline | ICTUS and COBRIT, together over 3,500 patients, were null |
| "Boosts memory in healthy people" | Weak | One positive industry-sponsored citicoline trial on secondary outcomes, which EFSA found insufficient; two small, short alpha-GPC studies with mixed results |
| "Alpha-GPC is the better choline" | Plausible in dementia | One meta-analysis of three trials; no head-to-head comparison in healthy people |
| "Needed with racetams" | Untested | Based on mechanism and user reports only |
| "Completely safe" | Unsettled for alpha-GPC | Well tolerated in trials, but a large stroke association and a TMAO mechanism in mice |
Swipe table →
08What would change the rating
A long placebo-controlled trial with stroke and cardiovascular events tracked would settle the alpha-GPC safety question. A replicated, independently funded trial in healthy adults would settle the nootropic question. For the other cholinergic route into cognition, see nicotine; for the wider supplement picture, natural nootropics.
09Where to get it
Supplement brands fund much of the citicoline research, and the alpha-GPC safety question is unresolved. We don't link to sellers, affiliate or otherwise. How we make money
What studies and practitioners report
Amounts as reported in the cited sources: trial doses with the population studied, and practitioner-reported ranges marked as such. Not medical advice.
- Trial 1,200 mg/day alpha-GPC · 400 mg three times a day for 180 days 261 patients with mild to moderate Alzheimer's disease PMID 12637119
- Trial 600 mg/day alpha-GPC · 12 weeks 100 adults with mild cognitive impairment (74 analysed) PMID 39300341
- Trial 500 mg/day citicoline · 12 weeks 100 healthy adults aged 50 to 85 with age-associated memory impairment PMID 33978188
- Trial 2,000 mg/day citicoline · 90 days; no benefit 1,213 patients with traumatic brain injury (COBRIT) PMID 23168823
Further reading: detailed protocols
- NIH Office of Dietary Supplements: choline fact sheet ods.od.nih.gov
- ClinicalTrials.gov: ICTUS citicoline stroke trial clinicaltrials.gov
- ClinicalTrials.gov: COBRIT citicoline brain injury trial clinicaltrials.gov
- Wikipedia: Choline alfoscerate en.wikipedia.org
10Connections
11Sources
- Lee G, Choi S, Chang J et al. Association of L-α glycerylphosphorylcholine with subsequent stroke risk after 10 years. JAMA Netw Open. 2021;4(11):e2136008. PMID 34817582PubMed 34817582
- Fioravanti M, Yanagi M. Cytidinediphosphocholine (CDP-choline) for cognitive and behavioural disturbances associated with chronic cerebral disorders in the elderly. Cochrane Database Syst Rev. 2005;(2):CD000269. PMID 15846601PubMed 15846601
- Dávalos A, Alvarez-Sabín J, Castillo J et al. Citicoline in the treatment of acute ischaemic stroke (ICTUS trial). Lancet. 2012;380(9839):349-57. PMID 22691567PubMed 22691567
- Zafonte RD, Bagiella E, Ansel BM et al. Effect of citicoline on functional and cognitive status among patients with traumatic brain injury (COBRIT). JAMA. 2012;308(19):1993-2000. PMID 23168823PubMed 23168823
- De Jesus Moreno Moreno M. Cognitive improvement in mild to moderate Alzheimer's dementia after treatment with the acetylcholine precursor choline alfoscerate. Clin Ther. 2003;25(1):178-93. PMID 12637119PubMed 12637119
- Jeon J, Lee SY, Lee S et al. Efficacy and safety of choline alphoscerate for amnestic mild cognitive impairment: a randomized double-blind placebo-controlled trial. BMC Geriatr. 2024;24(1):774. PMID 39300341PubMed 39300341
- Sagaro GG, Amenta F. Comparison of the effects of choline alphoscerate and citicoline in patients with dementia disorders: a systematic review and meta-analysis. Front Neurol. 2025;16:1649661. PMID 41426989PubMed 41426989
- Nakazaki E, Mah E, Sanoshy K et al. Citicoline and memory function in healthy older adults: a randomized, double-blind, placebo-controlled clinical trial. J Nutr. 2021;151(8):2153-2160. PMID 33978188PubMed 33978188
- EFSA Panel on Nutrition. 'Citicoline' and support of the memory function: evaluation of a health claim. EFSA J. 2024;22(7):e8861. PMID 38966137PubMed 38966137
- Kerksick CM. Acute alpha-glycerylphosphorylcholine supplementation enhances cognitive performance in healthy men. Nutrients. 2024;16(23):4240. PMID 39683633PubMed 39683633
- Marcus L, Soileau J, Judge LW, Bellar D. Evaluation of the effects of two doses of alpha glycerylphosphorylcholine on physical and psychomotor performance. J Int Soc Sports Nutr. 2017;14:39. PMID 29042830PubMed 29042830
- Wang Z, Hazen J, Jia X et al. The nutritional supplement L-alpha glycerylphosphorylcholine promotes atherosclerosis. Int J Mol Sci. 2021;22(24):13477. PMID 34948275PubMed 34948275
- Kim HK, Park S, Kim SW et al. Association between L-α glycerylphosphorylcholine use and delayed dementia conversion: a nationwide longitudinal study in South Korea. J Prev Alzheimers Dis. 2025;12(4):100059. PMID 40155153PubMed 40155153
12Entry history
Latest content update.
Heretic informs and never prescribes. Where amounts appear, they describe what published studies used. Talk to a qualified clinician before acting on anything here.