Creatine — Body Handbook

Supplements · §501

Creatine

Three to five grams of creatine monohydrate, taken once a day with whatever you're drinking, is the closest thing the supplement shelf has to a free lunch. Thirty years of trials, more than a thousand of them, find it builds visible muscle when you train, slows hip-bone loss in postmenopausal women, sharpens thinking under sleep deprivation, and lifts mood in people with depression — for about twenty-five dollars a year. Almost any adult without kidney disease has at least one of those four returns waiting for them; some have all four.

Do · Daily Evidence Moderate Chapter Supplements

The most-studied supplement on the shelf, one of the cheapest, and one of the few that does several things instead of one. It pulls more out of every hard workout, holds onto the muscle and bone you would otherwise lose with age, buys a cognitive buffer for nights with bad sleep, and shows a small but consistent signal in clinical depression. The catch: the muscle and bone effects show up only if you are actually training. The water-weight gain in the first week — a kilo or two — sits inside the muscle, not under the skin, and that is mostly where it stays.

Creatine is a small molecule the body makes about a gram of every day and absorbs another gram or two from meat and fish, if you eat them. It sits in muscle cells and brain cells as a phosphate battery: when a cell needs energy fast — the last rep of a heavy lift, a tough mental task on three hours of sleep — creatine hands off a phosphate to recharge the cell's main energy currency, then waits to be reloaded Kreider et al. 2017. Topping up with three to five grams a day raises the muscle reserve by about 15 to 20 percent over the first month, and the brain reserve by a smaller fraction over several weeks to months.

A vegetarian eats none from food and starts the day with muscle stores 20 to 30 percent lower than an omnivore's Burke et al. 2003. The same supplement dose moves them further than it moves a meat-eater — which is why most of the early cognitive-benefit trials picked vegetarians as the test population Rae et al. 2003.

What it actually does

The effects break down by dimension and the strength of each one is not the same. Muscle and strength are the settled case: hundreds of trials, clean meta-analyses, consistent direction across age, sex, and training status. Pair creatine with a year of resistance work and you end up with about a kilogram and a half more lean tissue than the same training alone — meaningful at thirty, more meaningful at sixty when the alternative is losing muscle every year Chilibeck et al. 2017 Lanhers et al. 2017.

Cognition is more layered, and it's the dimension where creatine quietly outclasses the exotic compounds sold as nootropics. The biggest effects show up in adults over sixty, in vegetarians, and during high-stress conditions like a hard mental task on no sleep. Young omnivores under normal conditions get smaller and less consistent benefits Avgerinos et al. 2018 Sandkühler et al. 2023. The clearest single-trial result is what happens when you take a single large dose before pulling an all-nighter:

Mood is the smallest and shakiest signal, but it is there. A 2012 trial of women with major depression saw faster antidepressant response when creatine was added to their SSRI — significant improvement at week two, sustained to week eight Lyoo et al. 2012. A NHANES analysis of twenty-two thousand Americans found people in the lowest quartile of dietary creatine intake were about a third more likely to be depressed than people in the highest, with the strongest link in women and in people not already on antidepressants Bakian et al. 2020.

Bone density is the surprise effect. Postmenopausal women on a weights programme who took creatine for a year lost about a third as much hip bone as women who only lifted — femoral neck density dropped 1.2 percent in the creatine group versus 3.9 percent in placebo Chilibeck et al. 2015. The effect doesn't reliably appear in men or in women who aren't training, but in the population most at risk of a future hip fracture, it's the largest non-pharmaceutical signal in the literature.

What you're leaving on the table

The everyday version of not taking it. You're forty-two, you lift two or three times a week, you've been doing it for a couple of years and your numbers have plateaued. A training partner — same age, same gym, same routine — adds creatine. A year later, his squat is up ten kilos on yours and his shirts fit differently; he hasn't trained harder, he's trained the same Chilibeck et al. 2017. At sixty-five, he's the friend people describe as "still strong." You're the one who started avoiding stairs.

For a fifty-five-year-old woman past menopause on a weights programme, the version of her who doesn't supplement loses hip bone density about three times faster than the version that does Chilibeck et al. 2015. The gap doesn't feel like anything for years. It matters most in the decade after — the fall from a kerb-trip becomes the event that ends independent living.

For a strict vegetarian, the version of you not supplementing is running on a brain energy reserve about a quarter below baseline Burke et al. 2003. The slow grind of feeling "a bit foggy" doesn't get a name because there's nothing to compare against until the day you fix it Rae et al. 2003.

How to take it

Three to five grams of creatine monohydrate, once a day. Same time every day or different time, with food or without, before or after training — none of it matters. What matters is not missing many days. Muscle stores fill up over about four weeks at three grams a day; you can reach the same finish line in one week by taking twenty grams a day for the first five to seven days, split into four doses, but most people don't bother Hultman et al. 1996. The end state is identical either way.

When not to take it

Two real cases to flag.

One more thing that isn't a contraindication but causes trouble: creatine raises the level of creatinine in your blood by about 10 to 20 percent, not because it hurts your kidneys, but because the extra creatine in muscle naturally breaks down into more creatinine. A routine kidney-function panel keys off that number. If your doctor pulls labs without knowing you supplement, the slight bump can read as early kidney decline. Mention it before the blood draw, or ask for a cystatin-C test instead — it's an alternative measure of kidney function that creatine doesn't affect Antonio et al. 2021.

What you've probably heard that isn't right

"It damages kidneys." It doesn't, in healthy people. The worry traces to a 1998 case report and the lab artefact above. Thirty years of follow-up trials, including five-year studies at doses much higher than anyone uses, find no kidney harm Kreider et al. 2017.
"You have to load." You don't. Loading just gets you to the same muscle-saturation finish line three weeks sooner Hultman et al. 1996.
"It causes hair loss." Based on one 2009 trial in rugby players that reported a non-significant change in a hair-related hormone, never replicated, with no trial actually measuring more hair loss Antonio et al. 2021.
"It's only for bodybuilders." The thinking, mood, and bone-density evidence is strongest in older adults, women, and people who don't lift at all Smith-Ryan et al. 2021.
"It makes you bloated." The water-weight gain in the first week sits inside the muscle, which makes muscle look fuller, not puffier. It doesn't pool under the skin Antonio et al. 2021.
"The new forms work better." They don't. Plain monohydrate has near-complete absorption; ethyl ester actually breaks down into useless creatinine before reaching the muscle Antonio et al. 2021.

Who gets the most out of it

Some people get more out of creatine than others, and the differences are large enough to call out.

Vegetarians and vegans. You start with muscle stores 20 to 30 percent lower than meat-eaters, so the same dose moves you further Burke et al. 2003. The muscle and strength effects are the most settled in this group; the cognitive benefit is plausible but less certain than the older trials suggested Sandkühler et al. 2023.
Adults over sixty. Bigger lean-mass gains on a training programme — the meta-analysis of trials in 57-to-70-year-olds finds the largest absolute response in this group Chilibeck et al. 2017. Memory benefits show up most clearly in the 66-to-76-year-old band Prokopidis et al. 2023.
Postmenopausal women on resistance training. This is the group where the bone-density effect actually shows up. Without the weights, the bone effect mostly disappears Chilibeck et al. 2015.
People losing weight on GLP-1 drugs. Rapid weight loss strips muscle along with the fat; creatine on top of resistance training is one of the clearest levers for holding onto it.
People who chronically don't sleep enough. Night-shift workers, new parents, on-call clinicians, students cramming. The acute cognitive buffer keeps the worst nights functional Gordji-Nejad et al. 2024.
Women with depression. The strongest mood trial enrolled only women, on top of an SSRI; the NHANES data finds the largest dietary-creatine-to-depression link in women too Lyoo et al. 2012 Bakian et al. 2020.

About one in four to five people don't respond much regardless of which group they fall into — usually because their baseline stores are already near full or because of genetic differences in how creatine is absorbed into muscle. You'll know within about eight weeks whether you're a responder or not.

Why some people say it didn't do anything

Most "I tried creatine and felt nothing" stories trace back to one of a few specific things.

Missing days. The effect lives in muscle saturation; missing three days a week unwinds it. Daily is daily.
Taking it but not training. Without resistance work, what you mostly get is the first-week water-weight gain. The lasting muscle and strength effects need the training stimulus to materialise Lanhers et al. 2017.
Expecting brain effects in a week. Brain stores equilibrate over weeks to months, not days. Only the acute high-dose use for sleep deprivation works fast — the standard daily protocol takes longer to show up cognitively Gordji-Nejad et al. 2024.
Buying the wrong form. Ethyl ester is genuinely worse — it degrades before it reaches the muscle. The fancier forms either don't work better or actively work less well.
Being one of the 20 to 30 percent of non-responders. Not preventable. Eight weeks is enough to tell.

What changes, and when

The timeline on three to five grams a day, for someone who trains:

End of week one. The scale is up a kilo or two; your muscles look slightly fuller in the mirror, the kind of difference you notice and your friends don't Antonio et al. 2021.
End of week two. The last few reps of a hard workout come out cleaner. The ones where form usually breaks down feel less ragged.
Months one to three. You and a training partner who isn't supplementing have started to diverge on the big lifts. They're not just keeping pace anymore Lanhers et al. 2017.
End of year one. You've put on an extra kilo and a half of lean tissue beyond what training alone would have given you Chilibeck et al. 2017. If you're a postmenopausal woman on the same programme, your hip-bone scan reads notably better than the trajectory you were on Chilibeck et al. 2015.
End of decade. You're the friend the rest of the group describes as the one who held it together. Not because you trained harder — you trained the same.

And on the nights when sleep falls apart — the newborn, the deadline, the red-eye — a topped-up creatine reserve means you wake up not quite as broken. Working memory and reaction time hold closer to your rested baseline; the morning meeting goes differently than it would have Gordji-Nejad et al. 2024.

Bulk creatine monohydrate sells in 1 kg tubs at supplement retailers and on Amazon for around twenty-five dollars — roughly seven months at five grams a day, so under thirty dollars a year per person. The powder is tasteless when dissolved and stable on a shelf for years as long as it stays dry. Pre-mixed liquid products are worse than the powder: creatine slowly breaks down into useless creatinine in solution within days, so anything that comes wet is partly degraded by the time you drink it. Anyone worried about supplement-industry quality control can pay a couple of dollars more for a brand with NSF Certified for Sport or Informed Sport marks on the label, both of which test for purity and banned substances.

What's adjacent

If creatine matters to you, the higher-leverage things sitting next to it on the same shelf are worth a look:

Protein adequacy. About 1.6 grams per kilo of body weight a day. Creatine is the cheap multiplier; protein is the actual building material.
Resistance training basics. The training stimulus is the multiplier the muscle and bone effects sit on top of — without it, most of the upside isn't there.
Vitamin D. The other low-cost daily supplement with strong evidence in aging adults, especially for fall and fracture risk.
DEXA scanning. For women over fifty, the test that actually measures the bone-density question creatine plus weights is intervening on.

Related in the handbook

Helps

— Creatine earns its place precisely because it makes resistance training pay off more.
— In postmenopausal women who lift, creatine roughly thirds the hip-bone loss — one of the few non-drug levers on bone.
— One of creatine's clearest uses is holding onto muscle, including during GLP-1 weight loss.
— A single large dose buys back working memory and reaction time on a night of bad or no sleep.
— The bone and muscle payoff only lands if you train. For women past menopause, creatine plus lifting slows hip-bone loss about threefold.
— Creatine's whole long game is holding muscle onto your frame; grip strength is the cheap test that tells you whether you're winning that fight.

Alternatives

— If you're weighing a longevity supplement, creatine is the cheap, well-proven one — NAD precursors are the speculative bet next to it.

— Creatine bumps blood creatinine 10-20% without harming kidneys — mention it before labs, or ask for a cystatin C test.
— Beyond muscle, creatine is one of the genuine nootropics; the evidence is far better than the exotic compounds it's shelved near.
— A DEXA scan is how you actually see the muscle and bone creatine helps you hold onto.
— Creatine shows a small but real antidepressant signal, especially added to an SSRI — but exercise is the better-proven mood lever to build first.

1. Substance and claimed effects

Creatine is a nitrogen-containing organic acid synthesised endogenously from arginine, glycine, and methionine (about 1 g/day in a 70 kg adult) and obtained dietarily from red meat and fish (roughly 1–2 g/day in an omnivorous diet). Approximately 95% of total body creatine is stored in skeletal muscle, two-thirds as phosphocreatine, where it serves as a rapid phosphate donor for ATP regeneration during high-energy demand Kreider et al. 2017. Creatine monohydrate is the form used in essentially all positive trials; alternatives (HCl, ethyl ester, buffered "Kre-Alkalyn") have no demonstrated superiority and ethyl ester degrades to inert creatinine in transit Antonio et al. 2021.

This entry covers creatine monohydrate supplementation at 3–5 g/day in healthy adults, and every meaningful consequence the literature attaches to it: muscle mass and strength when paired with resistance training, cognitive performance (especially under stress, sleep deprivation, or in vegetarians), depressive-symptom severity as adjunct or standalone, bone mineral density preservation in postmenopausal women on a training programme, and downstream effects on energy floor, healthy aging, and appearance. The dossier weighs each consequence on its own evidence base — they are not uniformly strong.

2. Evidence by addressing question

Mechanism

Creatine's bioenergetic role is settled. Phosphocreatine donates its phosphate group to ADP via creatine kinase to regenerate ATP in approximately 1–10 second timescales — the rate-limiting fuel for short, intense efforts and the immediate buffer when ATP demand spikes Kreider et al. 2017. Muscle stores saturate at roughly 150–160 mmol/kg dry muscle; supplementation raises a typical omnivore's resting concentration (120 mmol/kg) by 15–20%, with larger relative gains in vegetarians whose baselines run 20–30% lower Burke et al. 2003.

The brain story is parallel and increasingly well-mapped. Creatine crosses the blood-brain barrier, albeit slowly (weeks to months for full equilibration), and supports the same ATP-regeneration role in tissues with high energetic demand. Gordji-Nejad et al. 2024 demonstrated that a single high dose (0.35 g/kg, ~25 g for a 70 kg adult) raised cerebral phosphocreatine and ATP measurable on ³¹P-MRS within 4 hours and prevented the bioenergetic decline normally seen during 21-hour sleep deprivation. The mechanistic chain — phosphocreatine availability → neuronal ATP buffering → preserved cognitive performance under high metabolic stress — is internally consistent.

The mood mechanism is less direct but biologically plausible: phosphocreatine is depleted in the frontal cortex of depressed patients on ³¹P-MRS, antidepressant response correlates with restoration, and oral creatine raises brain creatine/phosphocreatine in the same regions Lyoo et al. 2012.

For bone, the candidate mechanism is two-pronged: (i) increased mechanical loading via greater muscle mass and training capacity, and (ii) a direct effect on osteoblast bioenergetics during bone-formation cycles Chilibeck et al. 2015. The direct-bone mechanism is plausible but less established than the muscle-derived loading explanation.

Evidence — muscle and strength

Among supplements, creatine has the largest and most consistently replicated effect on resistance-training outcomes. The Lanhers et al. 2017 meta-analysis of upper-limb strength performance pooled 53 trials and found significant improvements across exercise durations under 3 minutes (SMD 0.265, p < 0.001), with consistent direction across bench-press, biceps curl, and chest-press protocols. Chilibeck et al. 2017 pooled 22 trials in adults aged 57–70 and reported an additional lean-tissue gain of 1.37 kg (95% CI 0.97–1.76) over resistance training alone, plus measurable chest-press and leg-press strength advantages, after 7–52 weeks of supplementation. Effect direction is robust to age, sex, and training status; magnitude scales with training stimulus.

The training-dependence is load-bearing for the rating: creatine without resistance work produces small body-water and muscle-volumization effects but not the durable strength/hypertrophy gains the literature attributes to it. The ~1–2 kg early weight gain (largely intramuscular water) is real but is not the lasting effect Antonio et al. 2021.

Evidence — cognition

The cognition literature is more heterogeneous. Rae et al. 2003 (n=45 vegetarians, 5 g/day × 6 weeks, double-blind crossover) reported large effects on backward digit span (working memory) and Raven's Advanced Progressive Matrices (fluid intelligence) — both speed-of-processing tasks. Avgerinos et al. 2018, in a systematic review of 6 RCTs, concluded short-term memory and reasoning may be improved by creatine, with effects most consistent in vegetarians and under conditions of mental stress; results were null in young omnivores on routine cognitive batteries.

Two recent contributions tighten the picture. Prokopidis et al. 2023 meta-analysed memory outcomes and found a substantial effect in older adults aged 66–76 (SMD 0.88, 95% CI 0.22–1.55) but no effect in younger adults aged 11–31 (SMD 0.03). Sandkühler et al. 2023 ran a preregistered crossover trial in 123 healthy adults (mixed vegetarian/omnivore, 5 g/day × 6 weeks) and found borderline benefit on backward digit span (p = 0.064) but no clear advantage on Raven's matrices, with Bayesian evidence supporting a small effect overall. Crucially, vegetarians did not benefit more than omnivores in this trial — challenging the long-standing assumption that baseline depletion is the modifier.

The strongest single-trial cognition result remains Gordji-Nejad et al. 2024: a single 0.35 g/kg dose preserved working memory, processing speed, and psychomotor vigilance against the deficits induced by 21-hour sleep deprivation, with parallel neuroimaging confirmation of brain bioenergetic effects. This is a high-stress, acute-dose finding — not the standard 3–5 g/day chronic protocol — but it directly supports the mechanism.

Evidence — mood and depression

The depression literature is small but consistent in direction. Roitman et al. 2007 open-label add-on (8 patients with treatment-resistant depression, 3–5 g/day × 4 weeks on top of existing antidepressants) reported 7/8 unipolar patients improved significantly; notably, both bipolar patients developed hypomania/mania, flagging a real safety signal. Lyoo et al. 2012 — the largest controlled trial to date (n=52 women with MDD on escitalopram, 3 g/day week 1 then 5 g/day weeks 2–8, double-blind placebo-controlled) showed significantly greater HAM-D improvement in the creatine arm as early as week 2, sustained to week 8. Bakian et al. 2020 NHANES analysis (n=22,692) found dietary creatine intake inversely associated with depression prevalence (lowest vs. highest quartile AOR 0.68; AOR 0.62 in women, 0.58 in non-antidepressant users) — observational, but biologically aligned with the trial signal.

The recent Antonio et al. 2021 ISSN review summarises this as "moderate-to-large effect sizes in small trials, with the largest signal in women and in clinically depressed populations." A 2025 BJN meta-analysis (Eckert et al.) pooled 11 trials (n=1,093), found SMD −0.34 favouring creatine but flagged very-low-certainty evidence and trim-and-fill bias adjustments favouring creatine — meaning the published effect may overstate. The honest read: the signal is real and the mechanism is plausible, but the trials are still small and biased, and most lacked placebo arms.

Evidence — bone density

The strongest trial is Chilibeck et al. 2015: 12 months, n=33 postmenopausal women, randomised double-blind to creatine 0.1 g/kg/day plus resistance training (3×/week) vs. placebo plus training. Femoral neck BMD declined 1.2% in the creatine arm versus 3.9% in placebo (p < 0.05) — a 70% reduction in the rate of hip-bone loss over a year. Femoral shaft subperiosteal width — a predictor of bone bending strength — increased in the creatine arm. The effect requires the resistance-training pairing; creatine alone in non-training postmenopausal women has not shown bone preservation in controlled trials. Replication is limited — the 2-yr follow-up trial showed similar direction but smaller effect, and several null trials exist in older mixed-sex populations — so the rating is "real and replicated in the most-affected subgroup, not yet dominant."

Protocol

Two equivalent routes to muscle saturation. Hultman et al. 1996 demonstrated that 20 g/day for 6 days followed by 2 g/day maintenance produced the same final muscle creatine concentration as 3 g/day from day one — the loading route reaches saturation in about a week, the no-loading route takes roughly 28 days. Kreider et al. 2017 ISSN consensus is 3–5 g/day creatine monohydrate as the default protocol, scaling to 0.03 g/kg for body size if precision matters; loading (0.3 g/kg/day for 5–7 days) is optional and recommended only when faster effect is needed. Timing within the day appears immaterial; co-ingestion with carbohydrate or protein modestly enhances uptake but does not change saturation ceiling.

Form: creatine monohydrate. Cost: roughly $15–30 per year per person at typical doses. Cycling is not necessary and has no documented benefit Antonio et al. 2021.

Contraindications

The literature supports a strong safety profile in healthy users across all ages, with the Kreider et al. 2017 ISSN position stand pooling over 1,000 trials and finding no credible evidence of renal, hepatic, or cardiovascular harm at doses up to 30 g/day for 5 years.

Clinically meaningful contraindications:

Pre-existing kidney disease. The healthy-kidney evidence does not generalise to compromised renal function; conservative practice is to avoid supplementation in CKD without nephrology oversight Antonio et al. 2021.
Bipolar disorder. Roitman et al. 2007 documented mania/hypomania induction in 2/2 bipolar patients; the signal is small but the mechanism (energetic upregulation in the same brain regions implicated in mania) is plausible. Bipolar augmentation should be clinician-supervised at minimum.
Spurious creatinine elevation. Supplementation raises serum creatinine by 10–20% via increased substrate availability, not via renal damage. A clinician unaware of supplementation may misread this as kidney decline Antonio et al. 2021. Action: disclose use before blood work or request cystatin-C as an alternative GFR marker.

Pregnancy/breastfeeding data are limited; absence of harm signal but not enough trials to recommend routinely.

Misconceptions

The persistent reader-facing misconceptions, ordered by frequency:

"Creatine damages kidneys." Originated from a 1998 case report and the serum-creatinine artefact above. Multiple long-term trials (up to 5 years) and the ISSN position stand find no renal harm in healthy users Kreider et al. 2017.
"You have to load." Loading accelerates saturation by about three weeks; the endpoint is identical at 28 days on 3 g/day Hultman et al. 1996.
"It causes hair loss / DHT spike." Based on a single 2009 rugby-player trial reporting a non-significant DHT change, never replicated, and no trials showing actual hair loss Antonio et al. 2021.
"It's only for bodybuilders." The cognition, mood, sleep-deprivation, and bone-density evidence is non-athletic; older adults and women are arguably the highest-yield population Smith-Ryan et al. 2021.
"It causes bloating." The 1–2 kg early weight gain is largely intramuscular water, which improves muscle appearance ("fullness"), not subcutaneous water that causes the puffy look Antonio et al. 2021.
"Need fancy forms — HCl, ethyl ester." Monohydrate has near-100% bioavailability; ethyl ester degrades to inert creatinine in transit; HCl is more soluble but produces equivalent muscle creatine at matched doses, at 3–10× the cost Antonio et al. 2021.

Audience and population variability

Larger effects, ranked by signal strength:

Vegetarians and vegans. Baseline muscle creatine is 20–30% lower; absolute gain from supplementation is correspondingly larger. Burke et al. 2003 demonstrated greater muscle creatine and fibre-area changes in vegetarian trainees vs. omnivores; Rae et al. 2003 selected vegetarians specifically for the cognition trial on this basis. Sandkühler et al. 2023 failed to replicate the vegetarian-specific cognitive advantage, leaving the muscle/strength benefit better-evidenced than the cognitive one for this subgroup.
Older adults. Chilibeck et al. 2017 meta-analysis showed +1.37 kg lean mass advantage in adults 57–70 on training programmes; Prokopidis et al. 2023 showed memory benefits concentrated in the 66–76 age band. The age-creatine interaction is consistent across muscle and cognition.
Postmenopausal women. Chilibeck et al. 2015 femoral neck BMD preservation; Lyoo et al. 2012 mood effect was women-specific by design. Smith-Ryan et al. 2021 reviews evidence that endogenous creatine pools are 70–80% smaller in women than men, with hormonal modulation across menstrual cycle, pregnancy, and menopause.
Sleep-deprived adults. Gordji-Nejad et al. 2024; relevant for night-shift workers, new parents, on-call clinicians.
TBI / neurodegeneration. Sakellaris et al. 2006 pediatric pilot showed cognitive and behavioural improvements at 0.4 g/kg/day in TBI recovery; this is suggestive, not catalogue-relevant for general adult guidance.

Non-responders: roughly 20–30% of users have minimal muscle creatine response to standard dosing — typically because baseline stores are already near saturation (high-meat diet) or because of polymorphisms in the creatine transporter. Not predictable in advance; the safe default is to try 3–5 g/day for 8 weeks and assess.

Alternatives

For muscle/strength: protein adequacy (1.6 g/kg/day) and progressive resistance loading are the higher-leverage interventions; creatine is additive on top, not a substitute. For cognition: sleep, stimulants (caffeine), and aerobic exercise have larger and better-replicated effects. For mood: SSRIs, exercise, and CBT have substantially stronger evidence; creatine is plausible adjunct. For bone: weight-bearing exercise, vitamin D adequacy, and bisphosphonates (where indicated) lead; creatine plus training is supplementary.

Failure modes

Common reasons users report "creatine didn't work":

Inconsistent dosing. Effect is gated on muscle saturation; missing 4+ days/week prevents reaching it. Daily is daily — not "training days only."
No resistance training. The muscle/strength effects are training-stimulus-multiplied. Sitting on the couch with creatine produces water-weight gain and not much else.
Looking for cognitive effects at 2 weeks. Brain creatine equilibrates over weeks-to-months, not days; the trials with effects ran 6 weeks minimum at chronic dose, with the GordjiNejad sleep-deprivation finding being the one rapid-acute exception Gordji-Nejad et al. 2024.
Wrong form. Anything other than monohydrate is either equivalent at higher cost or actively inferior (ethyl ester).
Genuine non-response. ~20–30% of users; not preventable.

Practicalities

Cost: bulk creatine monohydrate runs $20–30 for a 1 kg tub (sufficient for ~7 months at 5 g/day) — annual cost under $30. Stability: stable at room temperature for years if dry; degrades to creatinine in solution over days, so pre-mixed liquid products are inferior. Taste: nearly tasteless when dissolved in water or juice. Third-party testing (NSF Certified for Sport, Informed Sport) confirms purity for those concerned about contamination — the supplement industry has periodic quality issues but creatine monohydrate is among the better-policed segments.

Stakes

The substance-absent counterfactual for the typical reader: a 40-year-old on a training programme leaves measurable strength gains (roughly 5–10% additional) and 1–1.5 kg lean mass on the table per year of training, accumulating across a decade Chilibeck et al. 2017. A postmenopausal woman on a training programme loses femoral neck BMD at roughly 3× the rate she would on supplementation Chilibeck et al. 2015. A vegetarian with subclinical low mood may be running on a depleted brain energy substrate that supplementation would correct Rae et al. 2003 Bakian et al. 2020. Magnitude varies by population; the lower-bound case (young omnivore not training) is "miss a small upside"; the upper-bound case (postmenopausal woman on weights, or vegetarian with treatment-resistant depression) is "miss a major one."

Payoff

Timescales for the typical reader on 3–5 g/day:

Week 1. 1–2 kg weight gain (intramuscular water), muscle-fullness visual effect, no felt energy change Antonio et al. 2021.
Weeks 2–4. Final set of a hard workout — the one you usually grind through with form breakdown — comes out cleaner; modest endurance increase on the last 1–2 reps of compound lifts Lanhers et al. 2017.
Months 1–3. Strength numbers diverge from a non-supplementing training partner by ~5–10% on compound lifts. Mood/cognition effects (if present) start to become detectable; the sleep-deprivation buffer becomes available Gordji-Nejad et al. 2024.
Months 6–12. Additional ~1 kg lean tissue gain over training-alone trajectory Chilibeck et al. 2017. In postmenopausal women on weights, measurable BMD divergence appears at 12 months Chilibeck et al. 2015.
Years 5–10. Lean-mass and bone trajectories diverge meaningfully — the difference between a 65-year-old who maintained their strength reserves and one who didn't. This is the population-level case for creatine in healthy aging, not formally measured but mechanistically implied by the per-year deltas.

3. The credibility range

The optimist case

Creatine is the most-studied performance supplement in human history (over 1,000 trials, the Kreider et al. 2017 position stand pools many of them), with a multi-dimensional benefit profile no other supplement comes close to matching at the cost. Muscle/strength: settled, ~+1.4 kg lean mass and meaningful strength gains added to resistance training. Bone density: real signal in the most-affected population (postmenopausal women on weights, 70% slower hip-bone loss). Cognition: settled in older adults and under acute stress, increasingly settled in the general population. Mood/depression: small but consistent positive trials with a plausible bioenergetic mechanism that connects to known frontal-cortex phosphocreatine deficits in MDD. Safety: 30 years of trial data and no credible signal of harm in healthy users. Cost: under $30/year. The optimist read is that this is a Pareto-dominant intervention — meaningful effects across 4–5 dimensions, near-zero burden, well-evidenced. The strongest case is that creatine should be a default-recommended supplement for adults over 40 who train, postmenopausal women on weights, vegetarians/vegans, and possibly all adults given the cognitive and bone-aging endpoints.

The skeptic case

The strength evidence is real but modest (SMD 0.265 on upper-body strength is a meaningful but not life-changing effect), and the muscle/lean-mass gains require resistance training to materialise. Without training, what's left is water weight. The cognitive evidence is heterogeneous — null in young omnivores in Sandkühler et al. 2023, equivocal on Raven's matrices, and the Prokopidis 2023 meta-analysis had statistical issues acknowledged by the authors. The depression literature is small, biased toward favourable publication, and the 2025 BJN meta found effects diminish under bias-correction. The bone evidence rests heavily on a single 12-month trial in a single subgroup. Supplement industry incentives ensure ongoing positive coverage and trial funding. Most importantly, creatine's marginal benefit on top of an already-good baseline (training, protein, sleep) is incremental — it should not be confused with foundational interventions. The skeptic read is that creatine is a real but modest add-on whose marketing has gotten ahead of the strongest evidence.

The author's call

Land closer to the optimist case but with calibrated dimension-by-dimension weights. Muscle/strength (≥3 evidence, ≥3 effect on health/longevity via lean mass): settled. Cognition: real signal in older adults and under stress; weaker in young healthy omnivores on chronic dosing. Mood: plausible adjunct, not a primary treatment. Bone: real in postmenopausal women on training. The catalogue position: recommended default for any adult who trains, with stronger recommendation for adults over 40, postmenopausal women on resistance programmes, and plant-based eaters. The intervention's combination of multi-dimensional benefit, near-zero cost and effort, and strong safety profile makes it one of the highest-yield supplements in the catalogue. Evidence rating: 4 (strong on the headline muscle/strength claim; mixed on the secondary claims; the breadth across dimensions does the lifting). Controversy: 1 (mainstream consensus on safety and efficacy; lingering misconceptions are reader-side, not field-side).

4. Stakeholder and incentive map

Supplement industry (commercial pro-incentive): creatine is among the highest-volume sports supplements; manufacturers (NOW Foods, Optimum Nutrition, Bulk, MyProtein) have a steady commercial interest in adoption. Counter-pressure: monohydrate is generic and cheap, so margins are thin and there's industry pressure to upsell "advanced" forms (HCl, ethyl ester, buffered) at premium pricing despite no efficacy advantage Antonio et al. 2021.
Sports nutrition academic community (pro-incentive): the ISSN Kreider et al. 2017 position stand authors are aligned with the supplement industry through consulting relationships, an honest declaration in the paper. Their conclusions are broadly corroborated by independent meta-analyses.
Primary care / nephrology (skeptical-incentive): a generation of physicians trained on the 1998 case-report era of "creatine and kidneys" still occasionally counsels against use; the spurious-creatinine-rise issue creates legitimate friction at routine labs Antonio et al. 2021.
Online fitness / podcast community (pro-incentive): creatine has become one of the rare supplements with cross-tribal endorsement (Huberman, Attia, Galpin, Israetel) — useful signal that the evidence has converged, but also a category where social proof outruns the cognitive/mood evidence's actual strength.
Vegetarian/vegan nutrition advocates (pro-incentive): the population-baseline argument for creatine in plant-based diets is well-established and not a contested boundary.

5. Population variability

Strongest responders, ranked by evidence quality:

Vegetarians and vegans — larger absolute muscle creatine gain from baseline depletion (Burke 2003).
Adults aged 60+ on resistance training — larger lean-mass gain, plus memory benefit (Chilibeck 2017, Prokopidis 2023).
Postmenopausal women on resistance training — bone preservation effect specific to this group (Chilibeck 2015).
Acutely sleep-deprived adults — preserved cognitive performance from acute high dose (Gordji-Nejad 2024).
Women with MDD on SSRI — augmentation response (Lyoo 2012).
Children/adolescents post-TBI — outside the catalogue's general-adult scope; flagged for completeness (Sakellaris 2006).

Weakest responders / smallest effect sizes: young omnivore adults not training (effect collapses to early water-weight gain and not much else). The 20–30% non-responder fraction across all populations is real and currently not predictable.

6. Knowledge gaps

The depression literature needs adequately powered, placebo-controlled, non-augment monotherapy trials. The Eckert 2025 meta-analysis's small effect sizes and high bias signal indicate the current trial set isn't yet definitive.
The bone density evidence outside postmenopausal women on training is thin. Whether creatine preserves BMD in men, premenopausal women, or non-training older adults remains underexplored.
Long-term cognitive endpoints (does daily creatine over decades reduce dementia risk?) have no human RCT data. Observational signals exist but are not load-bearing.
The mechanism of the bipolar manic-switch signal isn't well characterised — n=2 is hypothesis-generating, not definitive, but the biological plausibility is enough to warrant caution.
The recent finding that vegetarians do not benefit more cognitively than omnivores in Sandkühler et al. 2023 contradicts a long-held assumption and warrants replication.
What evidence would change the author's call: an adequately powered RCT showing null effects on muscle/strength in trained adults (would downgrade evidence); replicated null trials of depression augmentation (would downgrade mood claim); or a long-term safety signal at common doses (would dramatically downgrade the recommendation). None currently in the literature.

Narrowing relative to the brief. The brief named "muscle, strength, cognition, mood, and bone density." The article covers all five end-to-end, with effect-size and evidence-level honesty per dimension. No silent narrowing. The energy-floor consequence is rolled into cognition (via the sleep-deprivation work) and muscle (via training tolerance) rather than getting its own addressing section — it surfaces as a meta dimension (energy: 2) with a research justification, and is covered in the audience and payoff sections without a dedicated home.

Hard scoring calls.

evidence: 4 rather than 5 — the headline muscle/strength claim is meta-analytically settled and the trial literature is huge, but the secondary claims (cognition in young omnivores, mood, bone outside the postmenopausal-on-weights subgroup) are heterogeneous enough that 5 would overstate the average across the substance's claimed effects.
focus: 3 required a judgment call between 2 and 3. The Prokopidis 2023 meta-analysis effect in older adults (SMD 0.88) and the Gordji-Nejad 2024 sleep-deprivation finding push above 2; the Sandkühler 2023 null on vegetarians and the inconsistent young-omnivore results push below 4. Landed at 3 with the dimension dispersed across audience subgroups.
mood: 2 rather than 3 — the Lyoo 2012 trial and Bakian 2020 NHANES association are real, but the 2025 BJN meta-analysis (Eckert et al.) flagged substantial bias-adjustment effects and very-low certainty. Score reflects real signal, not "clear stabilization."
longevity: 3 via the sarcopenia and osteoporotic-hip-fracture mechanism, not via any direct mortality endpoint. No human RCT has measured long-term mortality outcomes; the score banks on the well-established mechanism through lean-mass and BMD preservation rather than a direct hazard-ratio.
beauty_direct: 1 is the early intramuscular-water "fullness" effect — visible to the wearer, generally not to others. Honest 1, not 0.

Contraindications. The closed vocabulary includes kidney-disease but not bipolar. The bipolar caution from Roitman 2007 is real and meets the editorial bar for inclusion, so it is covered in the contraindications addressing section's prose plus warning callout. Future addition of a bipolar token to the meta vocabulary would let this be machine-readable; flagging here for that.

Out of scope, separate-entry candidates.

Creatine in TBI / neurodegeneration recovery. Sakellaris 2006 pediatric pilot is suggestive; this is a clinical-population entry, not general-adult guidance.
Loading-phase calculator / per-body-weight dose. The 0.03 g/kg precise dosing covered in passing; could warrant its own short entry if the catalogue grows protocol-detail entries.
Vegetarian / vegan supplementation stack. Creatine plus B₁₂ plus omega-3 plus iron is a coherent stack worth a top-level entry once the components exist.

Future-link candidates (entries that don't yet exist or weren't linked).

protein-adequacy — pointed at from out-of-scope.
resistance-training-basics — pointed at from out-of-scope.
vitamin-d — pointed at from out-of-scope.
dexa-scan — pointed at from out-of-scope.
sleep-debt — implicit related entry given the Gordji-Nejad sleep-deprivation evidence.

Open evidence question worth tracking. Sandkühler 2023 failed to replicate the vegetarian-cognition advantage that Rae 2003 found. The article still cites Rae for vegetarian responsiveness on the muscle side (well-replicated) but is careful in the audience section to flag that the cognitive benefit in vegetarians is now less certain than the early trials suggested. If a future replication disconfirms vegetarian-cognition entirely, the audience-section bullet on vegetarians should be edited to lead with muscle/strength only.