Facts About Creatine

Where creatine’s claims sit on the certainty ladder

The tier ladder keeps the parent project’s grammar in a more legible form: bar length encodes evidentiary weight, colour depth encodes certainty, and the bottom rung is deliberately short because it is where arguments live, not where effects are. What is unusual about creatine is that a fifth stratum is needed below the usual contested tier — a debunked shelf, because several of the most widely believed claims about creatine (kidney destruction, dehydration, hair loss, steroid equivalence) have been tested and have failed.

How much does a body actually need?

The gym-culture “requirement” of 5 g/day is a supplementation dose, not a biological need — the first popular belief to challenge. Physiologically, total daily turnover is about 2 g/day in a typical adult, met roughly half-and-half by endogenous synthesis and diet (Brosnan & Brosnan 2016; Brosnan 2011). Creatine is a non-proteinogenic amino-acid derivative built from three amino acids: arginine + glycine (via AGAT in kidney, pancreas and liver, forming guanidinoacetate) then methylated by methionine as S-adenosylmethionine (via GAMT, mainly hepatic). That last step is expensive — creatine synthesis consumes on the order of ~40% of the body’s SAMe methylation budget, which is why the precursor amino acids matter as much as creatine itself, and why low-protein diets quietly strain the system.

Challenge to the popular belief: “everyone needs to supplement” and “nobody needs to supplement” are both wrong. A healthy omnivore synthesises and eats enough to hold the pool. The genuinely exposed groups are vegans/vegetarians (zero dietary supply), older adults with low meat intake, and possibly people on chronic low-protein diets — for whom the precursor amino acids (glycine, arginine, methionine) are also the constraint. NHANES associations linking <1 g/day intake to higher all-cause mortality (HR 0.85 for ≥1 g/d, 95% CI 0.72–1.00, 19.8-yr follow-up) are observational Tier-3 evidence at best: recall-based intake, single 24-h snapshots, and meat intake as an obvious confounder.

The biggest pro-creatine belief, put on trial

Here the strongest independent evidence pushes against the marketing, not for it. The 2025 UNSW randomized trial (Desai, Pandit … Hagstrom, Nutrients) did what almost no prior trial had done: it inserted a 7-day non-exercise wash-in before training began, so the fluid effect of creatine could be separated from actual tissue growth. Sixty-three adults (34 women, 29 men, 31 ± 8 y) took 5 g/day or served as controls, then both groups trained identically for 12 weeks.

Against the UNSW null stands the pooled literature: a 2025 dose–response meta-analysis of 61 trials found creatine increased fat-free mass by +1.39 kg (95% CI 1.07–1.70) and body mass by +0.89 kg, with no effect on fat mass META. But nearly all constituent trials used DXA or similar methods that cannot distinguish intracellular water from contractile tissue — exactly the confound UNSW isolated. At the elite end, an independent Spanish RCT in 23 professional U23 cyclists found 20 g/day for 6 days produced no consistent benefit on recovery, body composition or performance RCT. And in frail older adults, two pooled trials (n=106, ~6 g/day, 14–16 weeks of training) showed training worked but creatine added nothing.

Kidneys: the myth runs both ways — and both extremes fail

Gym forums say creatine destroys kidneys; supplement marketing says it is “proven completely safe.” The 2025–2026 nephrology literature — largely from groups with no supplement-industry ties — supports neither extreme. The mechanistic core is an artifact: supplemental creatine degrades non-enzymatically to creatinine, the very molecule used to estimate kidney function. Serum creatinine rises, creatinine-based eGFR falls, and no injury has occurred.

What survives scrutiny on the cautious side: nobody has run long RCTs in people with pre-existing chronic kidney disease — trials in type 2 diabetics, peripheral artery disease and hemodialysis patients show no injury signal, but they are short and small. Rare case reports (including one acute tubular necrosis) exist but are confounded and number in single digits against tens of thousands of trial participants. The one real-world harm is diagnostic: an artifactually raised creatinine can trigger misdiagnosis, unnecessary work-ups, or wrongly adjusted drug doses. Anyone supplementing should tell their physician; confirmation should use cystatin C or measured GFR, not creatinine-based eGFR.

The long-term record, trial by trial — does chronic intake damage the kidney?

To answer the long-term question directly, here is every substantive long-duration human dataset located in this research pass, with method quality noted. The pattern is consistent: no study using an injury-capable method has detected kidney damage in people with healthy kidneys, across exposures from 12 weeks to 5 years and doses to ~10 g/day — while the honest gaps (pre-existing CKD, decades-long unsupervised use) remain unfilled.

Long-term verdict, stated plainly: at 3–10 g/day, across every controlled study up to 2 years and observational use to 5 years, chronic intake has not damaged healthy kidneys by any creatinine-independent measure. What long-term intake does do, permanently while you take it, is elevate serum creatinine — a lab artifact that will follow you into every routine blood panel. The remaining unknowns are pre-existing CKD stages 1–5 (zero long-term RCTs — the 2026 meta-analysis authors note this gap explicitly) and unsupervised decades-scale use. If you have one kidney, reduced GFR, diabetes with nephropathy, or take nephrotoxic drugs (NSAIDs chronically, some antibiotics), supplement only with physician monitoring.

“Drink gallons or you’ll dehydrate” is backwards

The osmotic logic is real but points the opposite direction from the myth. Creatine pulls water into muscle cells: loading (20 g/day) measurably raises total and intracellular body water for days, which is most of the famous 1–2 kg scale jump. Over 5–10-week training studies, total body water relative to muscle mass shows no lasting distortion. On dehydration and cramping — a belief once endorsed even by the ACSM in the early 2000s on speculation, not data — the controlled literature finds no causation, and several heat-stress studies found creatine users had lower core temperatures and better fluid retention than placebo; older athlete cohorts reported cramping reduced, not increased. The Brazilian (Gualano-group) 2025 safety review in Frontiers in Nutrition concludes hydration and thermoregulation are unaffected.

Liver: no credible harm signal — and an inverted association

No meta-analysis or long-term trial detects hepatotoxicity at studied doses. Dedicated hepatic-marker meta-analyses (Alizadeh 2022, Sports Medicine – Open; Kashi 2024, Nutrition Research Reviews) find liver enzymes unchanged; a 52-week trial at 5 g/day in older adults kept enzymes stable; adolescent and youth-athlete cohorts through full competitive seasons kept ALT/AST within reference ranges with no hepatotoxicity signal META RCT. The inversion worth reporting: NHANES analyses associate low dietary creatine (<0.95–1 g/day) with higher risk of liver conditions — observational only COHORT, meat-intake-confounded, but it flips the folk narrative. The honest gap: people with existing liver disease are nearly absent from trials; since the liver hosts the GAMT step of synthesis, severe hepatic impairment is a legitimate reason for medical supervision, not a documented creatine harm.

Skin: the weakest, most industry-owned claim in the file

This is where the independent-versus-commercial filter matters most. The frequently cited anti-wrinkle findings come from topical cosmetic formulations studied largely by Beiersdorf (Nivea) scientists: a single-centre controlled study of 43 men applying a cream containing creatine plus guarana and glycerol reported firmness and wrinkle improvements at six weeks (Peirano 2011, J Cosmetic Dermatology) COMMERCIAL; earlier in-house work reported photoprotection and stimulated collagen, glycosaminoglycan and ceramide synthesis (Lenz 2005) COMMERCIAL. Mechanistically plausible — skin’s creatine-kinase energy system genuinely declines with age and oxidative stress — but these are multi-ingredient, industry-run, single-centre studies, never independently replicated at scale.

The strongest genuinely new post-2015 story

The brain burns ~20% of the body’s energy at 2% of its mass, and brain creatine — unlike muscle — responds slowly and modestly to supplementation. The most striking independent finding comes from a German government research institute (Forschungszentrum Jülich): in a randomized double-blind crossover trial, a single oral dose of 0.35 g/kg during 21 hours of sleep deprivation raised brain PCr/Pi, prevented a pH drop, and improved cognitive performance and processing speed — partially reversing fatigue-related deterioration (Gordji-Nejad 2024, Scientific Reports, n=15) RCT INDEPENDENT. A 2026 replication at a lower 0.2 g/kg dose (n=29) still mitigated deterioration in logic, numerical tasks, language-processing speed and psychomotor vigilance — with the authors noting the short-term-memory effect appears dose-dependent.

Women: real physiology, thinner evidence than the headlines imply

The physiology is distinct and understudied: females carry 70–80% lower endogenous creatine stores than males yet ~10% higher resting intramuscular concentrations, and estrogen/progesterone modulate AGAT/GAMT expression and creatine-kinase kinetics across the menstrual cycle, pregnancy and menopause (Smith-Ryan 2021; Ellery 2016). That makes supplementation plausibly more relevant at specific life stages — and also makes early performance studies that ignored cycle phase harder to interpret.

The women’s trial base, paper by paper

Because the headlines outrun the trials here, the substantive female-specific studies located in this pass are tabulated below. Two things stand out. First, the two largest and longest women’s trials ever run — both independent — were bone trials, and both were essentially null on bone density. Second, the most consistent positive female-specific signal is not muscle at all: it is depression augmentation and brain energetics, from an independent psychiatry group (Univ. of Utah / Seoul) with no supplement-industry stake.

Reading that table honestly: for bone, the two independent 2-year RCTs (437 women combined) say creatine does not raise bone mineral density — at most it preserves some geometry of the femoral neck when combined with training (Chilibeck) and does nothing without training (São Paulo). For muscle, the effect in women is real but small (+0.37 kg pooled lean mass) and appears to require training. For mood and brain, women are where creatine’s most interesting clinical signal lives — plausibly because of the 70–80% lower baseline stores — but the total randomized evidence is a few hundred participants, with the adolescent trial null on symptoms. For pregnancy, human work has only just reached pharmacokinetics; every efficacy claim you may read is extrapolated from spiny-mouse and guinea-pig models.

Challenge to the popular beliefs: “creatine makes women bulky” fails twice — women’s hormonal milieu limits hypertrophy, and the UNSW sex-disaggregated data showed women’s early scale gain was the fluid component with no added muscle growth thereafter. But the mirror-image marketing claim — creatine as a proven menopause therapy — also overreaches: the peri-menopause trials are tiny (n=15–36), weeks long, and partly non-randomized. Real signal, Tier 3, worth watching, not yet worth prescribing.

Hair, DHT and testosterone: one study built the myth, one finally tested it

The hair-loss belief traces to exactly one trial: 2009, college rugby players, three weeks, reporting a +56% DHT rise after loading. It was never replicated — twelve subsequent studies on testosterone and its metabolites found no significant hormonal increases. In 2025 the first direct test arrived: a 12-week RCT in 45 resistance-trained men (5 g/day vs maltodextrin placebo) measuring hormones and actual hair-follicle outcomes found no group differences in DHT, DHT-to-testosterone ratio, or any hair-growth parameter — the authors call it strong evidence against the claim (Lak 2025, JISSN).

Taking creatine with other nutrients, supplements — and water

The direct answer to “will taking it with other macro/micronutrients or supplements damage me or be unhealthy?” is: no harmful nutrient–creatine interaction has been documented in the controlled literature. The known interactions all concern uptake and effectiveness, not safety. The one practical safety issue is not an interaction at all — it is product purity: multi-ingredient pre-workouts with unverified formulations are the setting where most real-world adverse-event reports arise, which is why plain third-party-tested monohydrate is the defensible choice.

How much is too much? Less alarming — and less magical — than believed

There is no documented lethal or organ-toxic oral dose of creatine monohydrate in humans; the practical ceiling is gastrointestinal. Above muscle saturation (reached by ~3–5 g/day within 3–4 weeks, or 20 g/day for 5–7 days), excess creatine is simply excreted unchanged in urine — expensive urine, not toxicity. The largest safety dataset ever compiled pooled 685 clinical trials: 12,839 participants on creatine (average ~12.5 g/day, durations to 14 years) versus 13,452 on placebo.

The misconceptions table — claim, verdict, evidence

The precursor system and the co-supplement neighbours

Glycine, arginine and methionine are the three amino acids the body spends to make creatine, and NHANES III (n=29,945) shows precursor availability declines with age — lowest in adults ≥65, precisely the group with the lowest meat intake and the most muscle to lose. Supplemental creatine measurably spares this pathway: it down-regulates AGAT, freeing arginine and glycine for other roles and relieving the ~40% SAMe methylation burden — a mechanistic argument for supplementation in low-protein contexts that is stronger than most marketing claims, though still Tier 3.

The commonly bundled neighbours, briefly and honestly: guanidinoacetate (GAA) — creatine’s direct precursor, effective at raising creatine in animal feed science, but raises homocysteine and is not an approved human supplement. β-alanine — a genuinely evidence-backed buffer (carnosine) for 1–4-minute efforts; a 2025 systematic review of 7 RCTs (n=263) found creatine + β-alanine co-supplementation adds little beyond each alone, with inconsistent synergy. Taurine and carnitine — fellow “carninutrients” absent from plants; neither substitutes for creatine’s phosphagen role. The pattern across all of them: the omnivorous diet quietly supplies a package of nitrogen compounds that vegan diets must synthesise or supplement, and creatine is simply the best-studied member.

Final tier placements

Summary: the popular beliefs versus this evidence base

Does long-term intake damage the kidneys?

Not in healthy kidneys, on every creatinine-independent measurement ever taken. Five years of observational use (Poortmans), two years of continuous placebo-controlled use (Bender), tracer-measured GFR trials including on high-protein diets, in diabetics, and in postmenopausal women — all clean. What long-term intake does do is permanently raise serum creatinine while you take it, which can be misread as kidney damage in routine blood work; the harm there is misdiagnosis, not nephrotoxicity. The honest boundary: no long-term RCTs exist in people who already have chronic kidney disease, so pre-existing CKD, a single kidney, or nephrotoxic medication use means physician supervision, not self-experimentation. The popular belief “creatine destroys kidneys” is debunked for healthy users; the marketing belief “proven safe for everyone forever” overreaches into the untested CKD gap.

Is it good for women’s health?

Selectively — and less broadly than the current wave of marketing implies. Best-supported: strength and performance gains alongside resistance training, and the mood/brain-energetics signal, where the independent Lyoo RCT (52 women, creatine + SSRI) showed one of the largest augmentation effects in the file and imaging trials show the mechanism engaging. Genuinely disappointing: bone — the two independent 2-year RCTs totalling 437 postmenopausal women found no bone-mineral-density benefit, and the São Paulo trial found no lean-mass or function benefit without training either. Genuinely unknown: perimenopause (first tiny trials only appeared in 2025) and pregnancy (human research reached pharmacokinetics in 2025; all efficacy claims are still animal-model extrapolation — do not supplement in pregnancy outside a trial). Female physiology (70–80% lower stores) makes women plausible high-responders, and female-specific safety data are clean (951 women, no serious adverse events), but “every woman should take creatine” is a slogan, not a finding.

What are the side effects — men and women, short and long term?

Across 685 trials and ~26,000 participants, side-effect rates are statistically indistinguishable from placebo (13.7% vs 13.2% of studies; 4.60% vs 4.21% of participants). The reproducible effects: 1–2 kg of intracellular water weight in the first weeks (larger and quicker in women per the UNSW wash-in data — fluid, not fat or muscle), and dose-dependent GI upset when large boluses (≥10–20 g at once) are taken, solved by splitting doses. Sex-specific fears tested and failed: hair loss/DHT in men (one unreplicated n=20 study vs a direct 2025 follicle RCT showing nothing), “bulkiness” in women (hormonal ceiling plus fluid misread). Long-term (to 2–5 years at 3–10 g/day): no organ-damage signal in kidney, liver, blood, or cardiometabolic panels; the permanent lab artifact on creatinine is the only lasting change. The genuine long-term unknown is unsupervised decades-scale megadosing, which no one has studied.

Will taking it with other nutrients or supplements be damaging or unhealthy?

No documented harmful nutrient interaction exists. Carbohydrate and protein co-ingestion help (insulin-mediated uptake); the high-protein-plus-creatine kidney scare failed its direct tracer-GFR test; β-alanine coexists neutrally; micronutrients show no adverse interplay. The only substantive debate is caffeine — a possible blunting of creatine’s performance effect under chronic high-dose co-ingestion, unresolved since 1996 and irrelevant to health; normal coffee drinking is fine. The one real-world hazard is not an interaction: it is unverified multi-ingredient products, where most adverse-event reports originate. Plain, third-party-tested creatine monohydrate avoids it.

How much water should be taken with it?

Normal, thirst-guided intake. The dehydration/cramping belief is not just unsupported — controlled studies point the other way (better fluid retention, lower core temperature in heat, fewer cramps in athlete cohorts). Dissolve doses in roughly a glass (250–500 mL) to protect the gut, expect the scale to rise 1–2 kg from water stored inside muscle cells, and ignore the gallon rule.

The reference base at a glance — all 73 sources, tabulated

How to read this table: "—" = not stated in the citation or in the document text · "n/a" = not applicable (reviews, meta-analyses, consensus documents and dataset analyses have no single intervention period) · a "~" before a country = inferred from the lead author's affiliation, not stated in the source · author counts reflect names listed in the citation, and "+" means "et al." (more authors than listed) · calendar dates of trial conduct are almost never stated in citations, so the from–to column is filled only where a source states it. Consistent with the standing caveat, every field requires verification against the primary source before clinical or published use.