Alpha-GPC

What alpha-GPC is

Choline is the building block of acetylcholine, one of the most important messenger substances in the brain, responsible among other things for attention and memory formation. The idea on the shelf is correspondingly simple: more building block, more messenger, better thinking.

Alpha-GPC is a form that reaches the brain well and releases choline there directly. It sits alongside the second common choline source, citicoline, which additionally supplies a building block for cell membranes. The two are constantly confused, even though their data could hardly be more different. In Germany they are mostly found on the food supplement shelf.

The Korean analysis

South Korea has a national health insurance system that records practically all prescriptions. A research group used it to analyze 12,008,977 people aged 50 and over, all without previous stroke, without Alzheimer’s disease and without cerebrovascular disease. The observation period was 9 years.

Those who were prescribed alpha-GPC had a 46 percent higher risk of stroke. Because the number of cases is huge, the confidence interval is extremely narrow: 43 to 48 percent. In addition, there is a dose-response relationship — those who took it for longer than 12 months had a higher risk than those who took it for less than 2 months. Such a relationship is one of the classic arguments for causality.

Why this is still not proof

It is an observation, not allocation by lot. The people who received alpha-GPC were 68 years old on average, the non-users just under 62 — a difference of almost 7 years — and they had more pre-existing conditions. So they were at greater risk anyway.

What matters is why someone in Korea is prescribed alpha-GPC: because they complain of memory problems. A common early cause of these is incipient vascular calcification in the brain, that is, exactly the condition that later leads to stroke. The drug would then not cause the stroke but indicate that someone was already on the way there. The authors themselves name this bias from the prescribing indication in their limitations.

What remains is a signal. Not proof, but not nothing either. A second Korean analysis from 2025 with 508,107 people with mild cognitive impairment found less frequent progression to dementia among alpha-GPC users and no increased stroke risk, and in people without progression to dementia even a lower one. This, too, is observation, with its own biases. The two large data sets thus point in different directions, and the question of blame remains open.

What the comparative studies actually tested

In 2025, a review was published comparing 3 randomized trials with 358 dementia patients. In it, alpha-GPC performs better than citicoline: in the overall clinical assessment, in cognitive function and in apathy. Most sales texts latch onto this paper.

Three points of context belong to it. First: 358 people in 3 mostly open, that is, unblinded, and also old studies. Second: exclusively dementia patients with vascular damage, no healthy people. Third, and this weighs most heavily: in all 3 studies, 1,000 milligrams daily were given as an injection into the muscle, for 90 days. These studies say nothing about what a capsule does in a healthy 40-year-old. In memory and verbal fluency — the tests that come closest to measuring what buyers hope for — there was no difference anyway.

What the studies show

Korean cohort analysis, 12,008,977 insured people

All insured people aged 50 and over without previous stroke, without Alzheimer’s disease and without cerebrovascular disease were analyzed over 9 years. Those who were prescribed alpha-GPC had a 46 percent higher risk of stroke, confidence interval 43 to 48 percent. The users were 68 years old on average, the non-users just under 62, and they had more pre-existing conditions.

Dose-response relationship in the same cohort

Those who took alpha-GPC for longer than 12 months had a higher risk than those who took it for less than 2 months. That more substance goes along with more events lifts the finding above a mere correlation, without ruling out the bias from the prescribing indication.

Comparison with citicoline in dementia, 2025

A review of 3 randomized trials with a total of 358 dementia patients. Alpha-GPC performed better in overall clinical assessment, cognitive function and apathy; in memory and verbal fluency there was no difference. In all 3 studies, 1,000 milligrams daily were given intramuscularly, over 90 days; the studies were mostly unblinded and old.

What the studies do not support

For healthy people who want to think more sharply, there are no robust studies — neither for alpha-GPC nor for citicoline. All serious data come from sick people, mostly with dementia, mild cognitive impairment or after a stroke. The comparative studies with citicoline gave an injection into the muscle over 90 days; two placebo-controlled trials tested capsules, but likewise in people with dementia or mild cognitive impairment. A dementia patient with vascular damage is not a healthy 40-year-old.

The stroke signal also remains unresolved. The Korean analysis shows that users suffered a stroke more often, not that alpha-GPC triggered it. Conversely, there is no randomized trial of comparable size that could give the all-clear; the second Korean cohort from 2025 found no increased risk, and in people without progression to dementia even a lower one, but it, too, is an observation. The contrast with citicoline is revealing: there, in the ICTUS trial, 2,298 patients with moderate to severe stroke were studied placebo-controlled in 59 centers, with an odds ratio of 1.03 — that is, without benefit, but also without differences in side effects. With citicoline, we know that nothing happens; with alpha-GPC, we do not know.

Status, approval and legal

In Germany, alpha-GPC is sold as a food supplement. In the EU it is considered a novel food that requires authorization; in May 2026, EFSA saw no safety concerns at up to 203.7 mg per day, but authorization by the European Commission had not yet been granted at the time of writing. In parts of Europe it is listed as a medicinal product, and in South Korea it is prescribed by physicians — which is also where the large analysis comes from. There is no approval anywhere as a nootropic for healthy people, and the studies used in advertising were conducted in people with dementia or mild cognitive impairment, in part with an injection into the muscle.

Safety

The safety situation is the clearest difference between the two choline sources, and it does not come out in favor of alpha-GPC. For citicoline there are data from 2,298 patients in a placebo-controlled trial in which nothing noteworthy happened. For alpha-GPC there are 12,008,977 records with a warning signal that nobody has resolved. This particularly concerns the group the signal comes from: people over 50 with vascular risk. Anyone who has high blood pressure, has smoked or has atrial fibrillation should discuss this with a physician and not decide in the drugstore.

Sources

• Lee et al., JAMA Netw Open 2021 (Korean cohort analysis on alpha-GPC and stroke risk)
• Sagaro and Amenta, Front Neurol 2025 (alpha-GPC compared with citicoline in dementia)
• Dávalos et al., Lancet 2012 (ICTUS trial on citicoline in stroke)
• Traini et al., Curr Alzheimer Res 2013 (alpha-GPC and cholinergic transmission in the brain)
• Parnetti et al., J Neurol Sci 2007 (cholinergic precursors in cognitive impairment of vascular origin)