What the evidence says about creatine and thinking

Creatine is not exotic. It is found mainly in red meat and seafood, your body also makes it, and about 95 per cent of it sits in skeletal muscle, with small amounts in the brain [3]. Its role in how the brain handles energy is the reason given for testing it on memory [1] — and a mechanism is where a question starts, never where it ends.

The question a supplement page should be asked is narrow: in people like me, eating what I eat, does taking this change how I perform on a task that resembles something I care about? Read that way, the position stand on creatine is mostly about exercise and sport [3], and the cognitive trials are a small set: one review found ten trials of memory in healthy people that met its criteria, and pooled eight of them [1].

Prokopidis and colleagues (2023) pooled eight randomised trials of creatine and memory in healthy people and reported a small overall benefit, a standardised mean difference of 0.29. Its own subgroup analysis placed that effect in older adults, aged 66 to 76 (0.88), and found none in people aged 11 to 31 (0.03) [1]. A letter to the journal has since argued that the analysis double-counted participants and that this produced false-positive findings [4].

Xu and colleagues (2024) pooled 16 randomised trials — 492 adults aged 20.8 to 76.4, healthy people and patients with specific diseases together — and reported an effect on memory of 0.31, rating their certainty for it moderate, with low-certainty signals on attention time and processing-speed time and no significant change in overall cognitive function or executive function [2]. They did not split memory by age. On attention time, which they did split that way, three trials of people over 60 showed no significant effect and five trials of adults aged 18 to 60 showed one; split by health instead, the same table found none in healthy participants (five trials) and an effect in patients (three) [2]. The paper carries a published corrigendum and a published commentary [2]. A pooled estimate inherits the weaknesses of what it pools; it does not launder them.

Put side by side, the two do not tell one story. One finds the memory effect in older adults and nothing in the young; the other finds its only age difference, on a different outcome, in adults under 60 and nothing over it. Prokopidis also excluded vegetarians and vegans from its pool [1], so its result says nothing about people who eat no meat. The honest summary is that a memory effect appears in pooled trials and that who it belongs to is not settled.

Every cognitive number on this page comes from a test task: recall a list, respond to a cue, hold a sequence. Those tasks are chosen because they are stable and repeatable, not because they resemble a Tuesday afternoon. The outcomes the two meta-analyses report are memory, attention, processing speed, executive function and overall cognitive function [1][2]; neither reports whether anyone wrote a better report, studied more effectively or felt clearer.

So the correct summary of a positive result is: on this task, in this group, over these weeks, the supplemented group scored better than the placebo group. Anything beyond that sentence is added by the person quoting it.

The International Society of Sports Nutrition's 2017 position stand reviewed the safety literature and found no evidence of harm from creatine monohydrate at recommended doses in healthy people. The only side effect it found consistently reported is weight gain [3].

Two flags stay attached. Anyone with kidney disease should raise it with a clinician before starting, and this site is written for adults only. Nothing on this page is medical advice, and nothing here diagnoses, treats or prevents any condition.

Fuel opens on diet, not on supplements, and the order is deliberate: sleep and recovery, then nutrition adequacy, then physical activity, then learning, training and environment, and only then targeted supplementation, where justified. Nothing in the seven-day protocol the Snapshot produces requires a purchase, a device, an app or a supplement, and that stays true.

If you came here from a video promising a step change, the honest summary is shorter than the video: a memory effect in some pooled trials, disagreement about whom it belongs to, and nothing measured outside a test task.

A grade belongs to one claim, in one population, for one preparation — never to a compound, a technique or an article. The weakest link is the part of the chain between the evidence and the thing you would actually do that is doing the least work.