Metabolic #2. Part one argued that what happens to a carbohydrate depends on the body receiving it. This part follows a molecule whose entire job is to buy the cell a few seconds, and part three will ask why energy goes missing while a blood panel reads normal.
Creatine is the most studied supplement on the shelf and still the one people are least able to judge for themselves. The reason isn't marketing. It works on something almost nobody records, and how much it does depends on a starting value almost nobody has measured. I've taken it, dropped it and picked it up again across twelve years of training without ever once establishing whether it did anything for me, which is a common enough story and a slightly embarrassing one for somebody who reads the papers. This page sets out the mechanism, what backs each separate claim made for it, where that backing runs thin, and a test you can run with a notebook.
This is an educational and strategic perspective, not personal medical advice. The views are the author's own and not statements by Atlas Cove Lda.
What the creatine kinase system is for
Skeletal muscle holds very little adenosine triphosphate relative to how fast it can spend it. Cytosolic ATP sits at roughly 3 to 6 mM depending on cell type, while the pool of phosphocreatine alongside it reaches as high as 30 mM (Wallimann et al., 2011). Ten times the reserve, held in a different molecule.
Creatine kinase catalyses one reversible reaction and runs it in opposite directions in different parts of the cell. Out at the myofibril, where ATP gets consumed, the enzyme lifts a phosphate group off phosphocreatine and hands it straight to ADP, regenerating ATP on the spot, while over at the outer mitochondrial membrane, where oxidative phosphorylation is busy making ATP, the very same enzyme runs the whole reaction backwards and recharges creatine instead.
The system therefore does two distinct jobs, and both matter for what supplementation can plausibly change. One is temporal: it defends the ATP concentration across the first seconds of high demand, so the many enzymes that assume a stable ATP level keep working while oxidative supply catches up. The other is spatial: phosphate is ferried between the site of production and the site of use, which is faster than diffusing ATP itself across a crowded cell.
Neither job is fuel supply. Nothing is oxidised, no substrate burns, and the pool is small and finite, so under real load it drains inside tens of seconds, at which point the muscle falls back on glycolysis and oxidative metabolism, both carrying costs the phosphocreatine system never imposed. Ten seconds of grace.
Why the effect appeared in five men out of eight
If phosphocreatine is what empties, ask how fast it comes back, and whether taking creatine speeds that up.
Greenhaff and colleagues supplemented eight men with 20 g daily across five days, then sampled the vastus lateralis at rest and again at 20, 60 and 120 seconds into recovery from one intense isometric contraction driven by electrical stimulation (Greenhaff et al., 1994). Five men raised total muscle creatine by roughly 25 percent. Those same five refilled phosphocreatine about 35 percent faster during the second minute of recovery. The other three managed 5 to 7 percent more muscle creatine and refilled no faster than before.
That split is the single most useful finding in the creatine literature, and it gets dropped almost every time the study is summarised. Nobody found a modest average benefit spread across eight men. Five men clearly gained, three clearly did not, and what sorted them was whether their muscle creatine had moved at all.
Three design features limit how far that number travels. The gain showed up in minute two of recovery and not in minute one, implicating the slower half of the refill. Electrical stimulation drove the contraction instead of the participant, so recruitment was controlled precisely, making this cleaner physiology and weaker sport science. And there was no placebo arm. The supplemented trial always ran second, so nobody can rule out an order effect.
How much room you have, and what determines it
The unit here is millimoles per kilo of dry muscle. Untrained men cluster around 120 mmol/kg dm and the ceiling sits somewhere near 150 to 160, which bounds how far anyone can climb and explains why people arrive with wildly different amounts of room.
Hultman and colleagues put 31 men through four different intake schedules and reported outcomes clean enough to be worth listing in full (Hultman et al., 1996):
Loading, 20 g daily for 6 days. Total creatine rose roughly 20 percent, about +23 mmol/kg dm. Around 17 percent of the ingested dose was retained.
Loading, then 2 g daily. The raised level held flat through day 35.
Loading, then nothing. The level fell steadily and matched baseline again by day 35.
3 g daily, no loading phase. Arrived at the same place by day 28, statistically level with the loading groups.
A loading phase changes the speed and leaves the destination alone. It buys about three weeks. The rest of that list describes a system with a ceiling and a slow leak, which needs no ritual to fill.
Diet sets where you start, because you absorb creatine from meat and fish and also build it yourself out of glycine, arginine and methionine. Burke and colleagues measured vegetarians at roughly 117 mmol/kg dm against 130 in non-vegetarians, and the supplemented vegetarians ended up with a larger rise than any other group in that trial (Burke et al., 2003). Across the full sample, how far someone climbed ran inversely to where they began, at r equals minus 0.77.
Read that alongside Greenhaff's responders and non-responders and the pattern resolves. Someone whose diet already parks them near the ceiling has almost nothing left to buy, someone starting low has a great deal of it, and most arguments about whether creatine works turn out to be arguments about headroom held by two people who have never measured their own. Nobody in that exchange is lying.
The route from a fuller buffer to a bigger muscle
Phosphocreatine reaches a bigger muscle only by an indirect route. Saying so plainly matters, because the explanations that sound more interesting are the ones with less behind them.
A larger buffer lets you do marginally more work at a given intensity before fatigue stops you, that extra work is extra mechanical stimulus, and extra stimulus, given enough protein and enough recovery, turns into extra adaptation. Creatine builds no tissue. It lowers the price of the training that does.
Chilibeck and colleagues pooled 22 trials covering 721 participants, all drawn from studies whose mean age was 50 or above, with resistance training performed two or three times weekly over 7 to 52 weeks (Chilibeck et al., 2017). Adding creatine produced 1.37 kg more lean tissue than training by itself, 95 percent confidence interval 0.97 to 1.76. Standardised mean differences reached 0.35 for chest press strength and 0.24 for leg press strength.
Ten of those 22 individual trials found no creatine effect whatsoever, which is the figure an individual reader should weight most heavily. A pooled estimate is a statement about a population; you are one person whose remaining headroom is unmeasured. This is the same argument as the responder split, arriving from meta-analysis rather than from biopsy.
If you want the training side of this argument in full, it is set out in strength training after 40 and what it actually buys you, which makes the case that resistance work has the widest set of downstream effects per hour invested. Creatine is a small multiplier on that hour and nothing at all without it.
The brain: a smaller rise and a narrower claim
The same enzyme system operates in neurons, which face a comparable problem, since neural activity throws up demand transients faster than oxidative metabolism can answer them, and a brain cell has the same reason as a muscle fibre to keep a phosphate reserve close to where the work happens. Getting creatine across the blood brain barrier is the constraint.
Roschel and colleagues report that brain creatine climbs by roughly 5 to 10 percent when people supplement, about half what muscle manages, and that nobody has yet pinned down a protocol that moves it reliably (Roschel et al., 2021). At least one study measured both tissues and found muscle up, brain flat.
What that rise accomplishes divides along a single line, and the line is metabolic stress. McMorris and colleagues allocated 19 participants to creatine or placebo for seven days, then imposed 24 hours of sleep deprivation with intermittent light exercise (McMorris et al., 2006). Four measures separated the groups: static balance, choice reaction time, mood state and the generation of random movement. The creatine arm lost less ground on each. Verbal recall showed nothing. Neither did spatial recall, plasma catecholamines or cortisol.
Declining less under load is a different result from getting better at rest, and only the first is decently supported. Trials run on rested adults who slept properly come back inconsistent, and the review will not draw a firm conclusion from them.
One inference that looks reasonable does not survive contact with the data. The most frequently cited cognitive trial enrolled vegetarians and vegans exclusively, selected because their baseline is low, and the same review records both a comparable trial in meat eaters that found nothing and evidence that brain creatine concentrations are similar in vegetarians and omnivores regardless of diet. The dietary headroom argument holds for muscle and does not transfer upward.
Creatinine, and the blood result that alarms people
Routine panels read serum creatinine as a proxy for how well the glomeruli filter, and estimated GFR is calculated straight from it. Creatinine is what creatine and phosphocreatine break down into. So the figure in your blood tracks your muscle mass and your dinner alongside your kidneys, which is harmless until somebody reads it as though only the kidneys were involved.
Antonio and colleagues address this directly, noting that creatinine in blood and urine can be pushed up both by supplementation and by a diet heavy in meat, and that increases of that origin are improbable indicators of declining renal function (Antonio et al., 2021). Their collated results are not uniform. Twelve studies saw no rise whatsoever. In eight it rose and stayed within the reference range. In two it crossed the upper limit.
A marker shifting for reasons unrelated to disease still deserves an explanation rather than a shrug. Should a panel report a raised creatinine or a reduced eGFR, say what you take and say what you eat, so the figure gets read with that in hand instead of disputed later. None of this replaces a doctor looking at your particular kidneys. Diagnosed renal disease is a different situation and this page does not address it.
What the evidence does not show
Creatine is unusually well studied, which makes it unusually easy to see where the support runs out. Sorted by how much weight each claim will bear:
Accelerated phosphocreatine restoration. Good support, narrow scope. Demonstrated directly by biopsy, but in 5 of 8 men, in a single trial without a placebo arm, using electrically evoked contraction. What would strengthen it: a larger placebo-controlled replication reporting responder status.
More lean mass and strength alongside resistance training after 50. Good as a pooled result. 22 trials, 721 people, effect sizes small to moderate. What weakens it: 10 of the 22 individually found nothing, and no trial reports outcomes split by whether muscle creatine actually rose.
Cell swelling as an independent anabolic signal. Weak in humans. Creatine is osmotically active and does pull water inward, but intracellular water counts toward lean tissue on a DXA scan, so any trial reporting extra lean mass has fluid and contractile protein tangled together. Burke put the correlation between those two gains at r equals 0.61, and his two supplemented arms added 2.4 kg and 1.9 kg. Genuine hypertrophy explains that. So does water. What would settle it: a design measuring intracellular water independently.
Cognitive protection under sleep deprivation. Moderate. The effect is a smaller decline rather than a gain, the sample was 19 people in parallel groups, and several cognitive measures showed nothing. What would strengthen it: replication at scale.
Cognitive benefit in rested, healthy adults. Weak and inconsistent. Brain creatine barely moves on standard protocols, and the best-known positive trial used a population chosen for low baseline. What would change it: a dosing protocol shown to raise brain creatine reliably, tested in omnivores.
Renal harm in healthy people. No evidence of it. What the data do show is a measurement artefact rather than an injury. The caveat is real all the same: the studies are in healthy populations, and existing kidney disease is outside what they can speak to.
Two limits run across that whole list. Most of this work was done in men, a genuine gap given who is most exposed to sarcopenia after 50. And almost no trial splits its results by responder status. I trained as an engineer, so the trial I want is obvious to me and apparently to nobody funding this: measure who saturated, then report the outcomes separately. It still hasn't been run.
Testing it on yourself across twelve weeks
Very little in a supplement cupboard is testable by the person paying for it. This one is. The mechanism predicts something specific that you can count.
What it predicts sits in the gap between your first working set and your last, at a load you hold constant. A fuller buffer refills more phosphocreatine between sets, so that gap should shrink. It says nothing whatsoever about your one-rep maximum. Looking for a change there is the usual reason people decide the stuff does nothing.
So the protocol is a notebook. Count the repetitions you complete on one main lift across three or four sets, holding the load and the rest between sets steady. Take that baseline before you start, then repeat it twelve weeks later with both unchanged. My own working sets sit at 8 to 12 reps three times a week, which is exactly where I'd take the reading if I were starting today. If the fall-off has narrowed, you had room and you used it. If it hasn't, and nothing else changed, your headroom was small and this lever belongs to someone else.
One ordering point before any of that. Since everything here reaches you through the volume of training you can absorb, a buffer applied to a stimulus you are not delivering has nothing to buffer. So the strength work comes first, writing down the fall-off comes second, and the supplement question only becomes answerable once both of those exist, which is why it usually stays open for years in people who are perfectly capable of closing it. The same reasoning governs recovery capacity, set out in the nervous system as a performance budget.
Where Atlas Cove fits
Every Atlas Cove week opens with measurement of the quantities that later decisions have to rest on, rather than with a list of things to take. Whether a compound like this offers a given person anything at all depends on their headroom, their diet, their training load and whatever their bloodwork already says. None of that is answered by a category on a shelf. How a week gets built from a person's own numbers is the shape of the argument rather than a protocol decided in advance.
The same logic runs through the rest of this series. What a nutrient does depends on the condition of the person eating it. That argument gets made at greater length in the right question to ask about fructose.
Questions people actually ask
Do I need a loading phase?
No. Hultman's data show 3 g daily with no loading reaching the same muscle concentration by day 28 that six days of 20 g reaches in under a week. Loading buys you roughly three weeks and nothing else. If you are starting twelve weeks out from something that matters, the distinction is irrelevant.
Does creatine damage your kidneys?
The available evidence in healthy people shows no renal damage. What it does show is that supplementation raises serum creatinine, which is the marker used to estimate kidney function, so a panel can look worse without anything being wrong. Tell your doctor you are taking it before the blood is drawn. If you have diagnosed kidney disease, this is a conversation for your clinician rather than a question a website should answer.
How long before I know whether it works for me?
Muscle saturates inside a week on a loading protocol and inside four on a low dose. Saturation is not a noticeable effect, though. Twelve weeks of consistent training with the set-to-set fall-off written down is the shortest honest test, since the benefit arrives through accumulated volume rather than directly.
Is it worth taking if I eat a lot of meat?
Possibly not, and that is a real answer rather than a hedge. Meat eaters start nearer the ceiling, and the size of any increase runs inversely to where you begin. Non-response clusters among people who had little room to start with.
Does it do anything for the brain?
Under metabolic stress, probably. Sleep deprivation is the condition with the most support, and the finding there is a smaller decline rather than an improvement. In rested, well-slept adults the evidence is inconsistent, brain creatine rises only about half as much as muscle creatine, and no dosing protocol has been shown to move it reliably. Treat cognitive claims made for it as considerably weaker than the muscular ones.
Sources
Antonio, J., Candow, D. G., Forbes, S. C., Gualano, B., Jagim, A. R., Kreider, R. B., Rawson, E. S., Smith-Ryan, A. E., VanDusseldorp, T. A., Willoughby, D. S., & Ziegenfuss, T. N. (2021). Common questions and misconceptions about creatine supplementation: what does the scientific evidence really show? Journal of the International Society of Sports Nutrition, 18(1), 13. DOI: 10.1186/s12970-021-00412-w
Burke, D. G., Chilibeck, P. D., Parise, G., Candow, D. G., Mahoney, D., & Tarnopolsky, M. (2003). Effect of creatine and weight training on muscle creatine and performance in vegetarians. Medicine & Science in Sports & Exercise, 35(11), 1946-1955. DOI: 10.1249/01.MSS.0000093614.17517.79
Chilibeck, P. D., Kaviani, M., Candow, D. G., & Zello, G. A. (2017). Effect of creatine supplementation during resistance training on lean tissue mass and muscular strength in older adults: a meta-analysis. Open Access Journal of Sports Medicine, 8, 213-226. DOI: 10.2147/OAJSM.S123529
Greenhaff, P. L., Bodin, K., Söderlund, K., & Hultman, E. (1994). Effect of oral creatine supplementation on skeletal muscle phosphocreatine resynthesis. American Journal of Physiology, 266(5 Pt 1), E725-E730. DOI: 10.1152/ajpendo.1994.266.5.E725
Hultman, E., Söderlund, K., Timmons, J. A., Cederblad, G., & Greenhaff, P. L. (1996). Muscle creatine loading in men. Journal of Applied Physiology, 81(1), 232-237. DOI: 10.1152/jappl.1996.81.1.232
McMorris, T., Harris, R. C., Swain, J., Corbett, J., Collard, K., Dyson, R. J., Dye, L., Hodgson, C., & Draper, N. (2006). Effect of creatine supplementation and sleep deprivation, with mild exercise, on cognitive and psychomotor performance, mood state, and plasma concentrations of catecholamines and cortisol. Psychopharmacology, 185(1), 93-103. DOI: 10.1007/s00213-005-0269-z
Roschel, H., Gualano, B., Ostojic, S. M., & Rawson, E. S. (2021). Creatine supplementation and brain health. Nutrients, 13(2), 586. DOI: 10.3390/nu13020586
Wallimann, T., Tokarska-Schlattner, M., & Schlattner, U. (2011). The creatine kinase system and pleiotropic effects of creatine. Amino Acids, 40(5), 1271-1296. DOI: 10.1007/s00726-011-0877-3
This is an educational and strategic perspective, not personal medical advice. The views are the author's own and not statements by Atlas Cove Lda.
Tom Wuerden · Co-Founder
Engineer turned Ironman