Creatine: What the Evidence Actually Supports

Creatine is the most researched sports supplement in existence, and one of the few where the marketing has mostly caught up with the science rather than running ahead of it. That makes it unusual in this industry.

We do not sell creatine. EnergyFirst sells whey protein, a greens powder, fish oil, a multivitamin and a pre-workout. Creatine is not in the range and there are no plans to add it. That is worth saying up front, because it means this article has no commercial reason to inflate what creatine does — or to talk you out of it.

What creatine actually does

Creatine combines with phosphate in muscle to form phosphocreatine, which regenerates ATP during very short, very hard efforts — the first five to ten seconds of a sprint or a heavy set. Supplementing raises muscle creatine and phosphocreatine stores by roughly 20 to 40 per cent, moving them from around 60 to 80 per cent of capacity toward a ceiling of about 160 mmol per kilogram of dry muscle (Kreider and colleagues, 2017 International Society of Sports Nutrition position stand).

The important distinction: creatine does not build muscle. It extends the work you can do, and the training you can then perform builds the muscle. Every practical conclusion below follows from that.

How much, and do you need to load?

No, you do not need to load.

The classic loading protocol is 5 g four times daily for five to seven days, followed by 3 to 5 g daily. But Hultman and colleagues showed back in 1996 that 3 g a day for 28 days reaches exactly the same muscle saturation. Loading is a speed choice, not an efficacy choice. It gets you there in a week instead of a month, at the cost of a higher chance of stomach upset and a faster jump on the scale.

More is not better. In a 2025 meta-analysis by Zhang and colleagues covering 14 studies and 523 participants, low-dose groups taking 2.5 to 7.5 g a day showed a larger strength effect than groups taking 19.5 to 23.5 g a day, whose result was not statistically significant. That is almost certainly an artefact of which studies used which doses rather than a real inverse relationship — but it is a clean illustration that the dose-response curve flattens early.

How big is the effect, honestly?

Real, but smaller than the marketing suggests, and it depends heavily on who you are.

The Zhang 2025 meta-analysis found an overall standardised mean difference for strength of 0.43. Split by training status, the picture changes sharply: untrained participants showed 0.88, while trained participants showed 0.32 with a confidence interval that just touched zero. In other words, the less trained you are, the more creatine appears to do — which is true of nearly every training intervention, and rarely mentioned on a tub.

For lean mass, a 2026 analysis in the Journal of the International Society of Sports Nutrition covering seven trials and 608 postmenopausal women found an average gain of 0.37 kg alongside a 7.5 kg improvement in leg-press one-rep max. Bone density did not change. In older adults generally, Sharifian and colleagues (2025) pooled 20 trials and 1,093 participants and found a 2.12 kg improvement in one-rep max.

The trial a creatine seller would leave out

Desai and colleagues published a trial in Nutrients in 2025 that deserves more attention than it received. Sixty-three participants took 5 g a day, with a seven-day wash-in period before twelve weeks of resistance training. During the wash-in — no training at all — the creatine group gained 0.51 kg more lean mass than placebo. Then, across the twelve weeks of actual training, both groups gained roughly 2 kg and there was no difference between them.

The authors interpreted the wash-in gain as water rather than muscle, which is consistent with what creatine does: it is osmotically active inside muscle cells and pulls fluid in with it. The trial had no external funding.

One modest trial does not overturn a large positive literature. But it makes a point that matters for anyone reading short studies: a meaningful share of the early “lean mass” gain attributed to creatine is intracellular fluid registering on a body-composition scan. The honest summary is that the effect is real, small, larger in the untrained, and partly water in the first few weeks.

Which form? Monohydrate, and it is not close

The ISSN position stand puts it plainly: no evidence demonstrates that creatine citrate, serum, ethyl ester, buffered forms or nitrate produce greater creatine retention than monohydrate.

The best head-to-head remains Jagim and colleagues (2012), who compared monohydrate against a buffered product at both its label dose and a monohydrate-equivalent dose in 36 resistance-trained men over 28 days. At the manufacturer’s recommended dose, the buffered form raised muscle creatine roughly five times less than monohydrate. There were no differences in fat-free mass, strength or anaerobic capacity between forms — and, notably, no evidence of fewer side effects, which is the usual justification for the premium. That study was funded by a monohydrate supplier, which is worth knowing; it is also still the best-designed comparison available.

Monohydrate is also the cheapest form and carries essentially all of the safety data. The exotic forms charge more for claims that have not been demonstrated.

The brain research: promising, small, and heavily oversold

This is where creatine marketing has got ahead of the evidence in the past two years.

Luo and colleagues (2024) pooled 16 trials covering 492 people in total. Memory improved with a standardised effect of 0.31, graded moderate certainty. But overall cognition showed 0.34 with a confidence interval crossing zero and a p-value of 0.22 — not significant. Executive function was also not significant.

The age pattern is the interesting part. Prokopidis and colleagues (2023) found a memory effect of 0.88 in adults aged 66 to 76, and 0.03 — essentially nothing — in young adults. A 2024 study by Gordji-Nejad and colleagues found meaningful cognitive improvements during 21 hours of sleep deprivation, but with 15 participants and a single 24.5 g dose that nobody takes daily.

On mood, a 2025 meta-analysis in the British Journal of Nutrition covering 11 trials and 1,093 people reported a small effect that its own authors described as falling below clinically meaningful thresholds, with bias favouring treatment detected and a GRADE certainty rating of very low. Their words: the true effect may be trivial or null.

The pattern across the whole cognitive literature — fewer than 500 participants in total — is that effects cluster in older adults, in people under acute stress, and probably in those with low baseline stores. In well-fed young adults the memory effect is close to zero. Promising, small, mostly in the depleted, and dramatically over-marketed is the accurate summary.

Safety, and the two questions everyone asks

Your bloodwork will look odd

Creatine supplementation raises serum creatinine — the marker clinicians use to estimate kidney function — without changing actual filtration. A 2025 meta-analysis by Kabiri Naeini and colleagues found a small but statistically significant rise in serum creatinine alongside no significant difference in glomerular filtration rate, and interpreted it as metabolic turnover rather than renal impairment.

The practical consequence is worth more than the physiology: a routine blood panel can look alarming in a creatine user. Tell your doctor you take it. Cystatin C is an alternative marker that supplementation does not affect. Long-term data beyond a year or two is sparse, and creatine has not been studied in people with existing kidney disease — that is a question for your doctor, not a blog.

Hair loss

The entire claim rests on one study. Van der Merwe and colleagues (2009) gave 20 college rugby players a loading dose and found DHT rose 56 per cent, with testosterone unchanged. It has never been replicated in the seventeen years since. More to the point, it never measured hair — not density, not follicle counts, not shedding. Both DHT values stayed inside the normal reference range.

In 2025, Lak and colleagues finally ran the direct test: 12 weeks, 5 g a day, with trichogram and photographic assessment of hair density and thickness alongside hormone panels. No significant differences on any hair or hormonal outcome.

That trial has real limitations, and the American Hair Loss Association listed them: 38 completers, no genetic screening for male-pattern baldness predisposition, twelve weeks is short relative to hair cycling, and scalp DHT activity — where the process actually happens — was not measured. So: the claim was never well supported, the one study that looked at hair found nothing, and “definitively disproven” overstates it. If you have a strong family history of male-pattern baldness, the question is not fully closed.

Water retention

Expect roughly one to two pounds in the first week, faster if you load. It is largely intracellular — inside the muscle — rather than the subcutaneous puffiness people picture. Skipping the loading phase blunts the jump.

So should you take it?

If you lift and you want a small, well-evidenced edge on high-intensity work, creatine monohydrate at 3 to 5 g a day is one of the few supplements that earns its place. It is cheap, it is well studied, and the safety record over decades is good.

Set your expectations at the size of the actual effect. And note the pattern that runs through this whole literature and through pre-workout supplements generally: the training is the intervention. Creatine amplifies it. Nothing in the research supports creatine as a substitute for the work, and the same is true of the protein you eat — supplements make good training work better, and do very little on their own.