Creatine and the Brain: A Cheap Supplement With a Modest, Real Signal

Creatine is one of the most studied supplements in the world, it costs a few cents a day, and it turns out to do something measurable for memory. A 2024 meta-analysis of 16 randomized trials found that creatine improved memory with a standardized effect of about 0.31, a small but real bump. That alone would make it mildly interesting. What makes it worth a deep dive for an APOE4 carrier is the bridge: carriers have measurably lower creatine in their brains, and the APOE4 brain has a well-documented energy problem. Creatine is, in theory, exactly the kind of thing that deficit calls for.

Theory is not proof, and the cognitive evidence is genuinely mixed, so this is a “modest, plausible, low-risk experiment” story, not a “proven dementia drug” story. Here is what holds up.

What the evidence actually shows

The best summary is that 2024 meta-analysis: 16 RCTs, about 492 people, ages spanning roughly 21 to 76. It found creatine significantly improved memory (standardized mean difference 0.31, 95% confidence interval 0.18 to 0.44) and significantly improved processing speed and attention time. To translate “0.31”: that is a small effect, the kind you would not notice in yourself day to day but that is real across a population. It is a nudge, not a transformation.

The same meta-analysis is also where the honesty starts. Creatine did not significantly improve overall cognitive function (effect 0.34, but p = 0.22) or executive function (0.32, p = 0.12). So the benefit is specific, it shows up in memory and speed, not in a general “smarter across the board” sense. And contrary to what you might expect, the memory and attention effects were actually clearer in younger adults (18 to 60) than in those over 60, where attention showed no benefit. That cuts against the simple “it helps aging brains most” assumption.

Then there is the counterweight that keeps you grounded. The largest single RCT in healthy adults (Sandkühler 2023, 123 people, 5 grams a day for six weeks) was essentially null. One working-memory measure bordered on significance (p = 0.064) but the effect was tiny, about a fifth of a digit longer span, and reasoning showed nothing. The authors’ own read was that any real effect is likely smaller than the flashy early studies suggested. When the biggest, cleanest trial comes back near zero, you trust the modest meta-analytic signal a little less, not more.

Why it is mechanistically interesting for a carrier

Here is the part that is specific to APOE4, and it is the reason creatine earns a page rather than a footnote. In a proton MRS study (which measures brain chemistry non-invasively) of 22 non-demented older adults, the 8 who carried APOE4 had significantly lower brain creatine than the 14 who did not. In the carriers, creatine levels even tracked with age and cognitive test scores. The authors read it as a sign of higher metabolic strain in the carrier brain. Separately, a large body of work shows the APOE4 brain handles energy poorly, with sluggish glucose metabolism years before any symptoms (the bioenergetics story we cover in brain mitochondria and bioenergetics).

Picture creatine as a rechargeable battery for brain cells. Neurons fire in bursts, and creatine helps them buffer energy when demand spikes, keeping the supply steady through the surge. If the APOE4 brain runs its batteries low, as the MRS data suggest, then topping them up is a plausible thing to try. But plausible is not proven, and that single sentence is the whole tension here. We have a measured deficit and a supplement that fills that exact deficit, and we have no trial showing that filling it protects a carrier’s cognition. The mechanism is a reason to be interested, not a reason to be confident.

The numbers in one place

What	Figure	What it means
Memory benefit (2024 meta-analysis)	SMD ~0.31 (CI 0.18-0.44)	small but statistically real, across 16 RCTs
Processing speed and attention	significant	also improved in the pooled data
Overall cognition / executive function	not significant (p = 0.22 / 0.12)	benefit is specific, not across the board
Largest single healthy-adult RCT	essentially null	123 people, 5 g/day, no clear effect
Brain creatine in APOE4 carriers	significantly lower (MRS)	the deficit creatine is theorized to buffer
Carrier-stratified cognition trial	none	nobody has tested creatine in carriers specifically

What is contested, in plain terms

The contested question is how much to weight a small, somewhat inconsistent cognitive signal. Optimists point to the meta-analysis and the mechanism; skeptics point to the null flagship trial and the fact that the overall-cognition effect was not significant. Both are reading the same evidence honestly. The fair synthesis: creatine probably does a little for memory and speed in some people, more reliably under stress like sleep deprivation, and we have no idea yet whether it does anything for long-term dementia risk in carriers. The APOE4 brain-creatine finding is mechanism, an interesting clue, not evidence of benefit. Keep those two boxes separate.

What to actually do

If you want to try it, this is one of the genuinely low-risk experiments available to a carrier, as long as you set your expectations correctly.

• Creatine monohydrate, about 3 to 5 grams a day. Monohydrate is the cheap, decades-studied form; the fancier versions cost more for no proven advantage. No loading phase is needed for cognition. Take it whenever is convenient.
• Set expectations honestly. The cognitive evidence is modest and mixed, the biggest trial was null, and it is not proven to prevent dementia. Treat any benefit as a small possible bonus, not the main event.
• It is broadly safe, but mind the basics. Decades of use show a strong safety record in healthy people; it draws a little water into muscle (slight weight gain), and anyone with kidney disease should clear it with a doctor first.
• Keep it in its place: a low-cost adjunct on top of the real levers. The bigger, better-proven moves for a carrier’s brain energy are exercise and strength plus aerobic training, which improve brain metabolism directly. Creatine rides alongside those; it does not replace them.

Common questions

Will creatine improve my memory? Maybe a little. The pooled trial data show a small, real memory effect, but the largest single study found essentially nothing, so do not expect a noticeable change in daily life. The clearest benefits tend to show up when the brain is under strain, such as during sleep deprivation.

Does it specifically help APOE4 carriers? Unknown. The interesting fact is that carriers run low on brain creatine and the carrier brain has an energy deficit, which makes creatine a logical candidate. But no trial has tested creatine for cognition in carriers, so the carrier-specific benefit is a hypothesis, not a result.

Is it safe to take long term? For healthy people, the safety record over decades is reassuring at the 3 to 5 gram dose. Expect a small bump in body weight from water in muscle. If you have kidney disease, check with your doctor before starting.

Do I need to “load” with a high dose first? No. Loading speeds up muscle saturation for athletes, but for the cognitive use a steady 3 to 5 grams a day is fine; your brain stores fill more slowly regardless.

Creatine is cheap, broadly safe, and has a small, real memory signal, and the APOE4 brain’s known energy deficit makes it a reasonable low-risk experiment, but it is not proven for dementia prevention and the biggest trial came back null. This is general education, not medical advice.

Sources

• Xu C, et al. Effect of creatine supplementation on cognitive function: a systematic review and meta-analysis. 2024. PMC
• Sandkühler JF, et al. The effects of creatine supplementation on cognitive performance in healthy adults: a randomized controlled trial. 2023. PMC
• Laakso MP, et al. Decreased brain creatine levels in ApoE epsilon4 carriers measured by proton MRS. PubMed
• Creatine supplementation, brain bioenergetics and cognition in aging (context). PMC