Creatine Monohydrate
Metabolic cofactors and energy intermediates. A research review published by South Beach Longevity.
Creatine Monohydrate
The phosphagen that earned its reputation, and the claims that borrowed itCreatine monohydrate is the rare supplement whose core claim — raising intramuscular creatine and improving high-intensity, repeated-bout work in people whose stores are not already high — is supported by decades of human trials. The same reputation has been asked to cover hypertrophy as a primary anabolic, cognition in well-slept omnivores, kidney injury in healthy adults, and commercial salts that outperform the monohydrate. Those extensions are not the same evidence.
Compiled by South Beach Longevity · 20 August 2026 Copyright 2026 Series SBL-41 / SP-CREATINE-MONOHYDRATE · Register A scientific article Sources peer-reviewed human trials, biopsy studies, society positions, meta-analyses, and labelled animal or in-vitro work · verified NCBI records Constraint This document describes published research. It is not medical advice. No human use, dose, route or schedule is recommended anywhere in this document.
How to read this document Every finding is labelled, in the sentence that reports it, by the kind of study that produced it. A vastus-lateralis biopsy is not a one-repetition-maximum. A loading design is not a daily-accumulation design. A vegetarian cognition trial is not a well-slept omnivore. A rise in serum creatinine is not a measured fall in glomerular filtration. Where two results conflict, both are given. Amounts and durations appear only as reported experimental parameters, always with the population attached. Nothing here is a recommendation. Findings are graded in place as established, strongly supported, emerging, plausible, or speculative. Two further labels mark careful absences rather than verdicts: not established, where the evidence is too thin to place a claim on the ladder at all — untested or insufficient, an absence of proof rather than disproof; and not supported, where the weight of evidence leans against a claim but stops short of a formal refutation.
01 What this document is, and four things it is not
This article is a research review of creatine monohydrate: the chemistry of the crystal, the phosphagen that makes the molecule interesting, the human evidence that oral monohydrate raises tissue creatine, the performance tasks that move when those stores rise, and the adjacent claims that have borrowed that reputation. It is written against a market that treats creatine as a general anabolic, a general nootropic, a kidney toxin, or a product category in which the newest salt wins.
Four things follow immediately.
First, this is not a supplementation programme. Harris, Söderlund, and Hultman showed, in a human biopsy study, that oral creatine raised creatine in resting and exercised muscle (Harris, Söderlund, and Hultman, 1992). Hultman, Söderlund, Timmons, Cederblad, and Greenhaff later compared a short high-intake design with a longer daily-accumulation design and found similar muscle increments once enough creatine had been delivered (Hultman et al., 1996). Those papers are established as descriptions of how stores were filled in the men they biopsied. They are not used here as a licence to prescribe a loading week, a maintenance gramme, or a drink schedule for any person.
Second, it is not proof that creatine is a primary anabolic. Resistance-training trials and meta-analyses have reported lean-mass and fibre-size changes that are real in some samples and smaller, and more water-confounded, than the commercial claim (Willoughby and Rosene, 2001; Branch, 2003; Nissen and Sharp, 2003; Chilibeck, Kaviani, Candow, and Zello, 2017). Collapsing a phosphagen effect into a testosterone-like mechanism is the most expensive error in this field after the kidney story.
Third, it is not a nootropic for people who already eat meat and sleep. Rae, Digney, McEwan, and Bates found a cognitive signal in vegetarians (Rae et al., 2003). McMorris and colleagues found signals under sleep deprivation (McMorris et al., 2006; McMorris et al., 2007). Avgerinos, Spyrou, Bougioukas, and Kapogiannis’s systematic review, and Prokopidis and colleagues’ later meta-analysis, do not convert those contexts into a general brain drug for well-slept high-intake omnivores (Avgerinos et al., 2018; Prokopidis et al., 2023). Treating creatine as caffeine with better branding is a category error.
Fourth, it is not medical advice and it is not a demonstration that creatine wrecks healthy kidneys. Serum creatinine can rise because creatine is metabolised to creatinine (Wyss and Kaddurah-Daouk, 2000). Controlled renal-marker studies in healthy adults have not shown a matching fall in true filtration (Poortmans and Francaux, 1999; Poortmans and Francaux, 2000; Gualano et al., 2011). Case reports in multi-supplement users and people with prior kidney disease keep a special-population question alive; they do not overturn the healthy-adult file. This document describes published research. It recommends no product, salt, dose, route, or schedule for any person.
The working thesis is narrower than the aisle. Creatine monohydrate is a well-characterised way to raise tissue creatine. The justified reputation is high-intensity, repeated-bout work in people whose stores can still rise. Safety in healthy kidneys must not be confused with a rise in serum creatinine. Hypertrophy as a primary anabolic, cognition as an omnivore nootropic, kidney-harm folklore in healthy adults, and other salts outperforming monohydrate are overextensions. The rest of the document is that sentence, unpacked.
02 Creatine chemistry and the monohydrate crystal
Creatine is N-methylguanidinoacetic acid. The free base has the molecular formula C4H9N3O2 and a formula weight of 131.13. The monohydrate — the form that built the human literature — is the same zwitterion with one water of crystallisation (C4H9N3O2·H2O; formula weight 149.15). That crystal is about 87.9 percent creatine by mass. A powder gramme of monohydrate is not a gramme of creatine. Commercial salts that advertise a smaller scoop often deliver less creatine once the counter-ion is subtracted. That is arithmetic, not a product ranking.
Wyss and Kaddurah-Daouk’s Physiological Reviews article remains the load-bearing chemistry-and-metabolism paper (Wyss and Kaddurah-Daouk, 2000). Creatine and phosphocreatine undergo non-enzymatic cyclisation to creatinine at roughly 1–2 percent of the body pool per day. In an adult whose total creatine pool is of the order of 120 g, that conversion accounts for a daily loss near 2 g, replaced by diet and by endogenous synthesis. Creatinine is excreted; it is not a second phosphagen. The conversion rate is why a larger creatine pool can raise serum creatinine without a matching change in glomerular filtration. Section 21 returns to that point with a table. Grade for the chemistry and the non-enzymatic fate: established.
The monohydrate is stable as a dry solid. In solution it slowly converts to creatinine; the rate rises with temperature and with time. That is why a dissolved drink is a pharmacokinetic object, not a shelf-stable drug, and why “creatine in a ready-to-drink” is a stability claim that has to be measured, not inferred from the dry crystal (Jäger et al., 2011; Antonio, Candow, Forbes et al., 2021). The 2021 International Society of Sports Nutrition (ISSN) misconceptions paper is a society-adjacent review. Several of its authors disclose ISSN roles, Alzchem scientific-advisory positions, creatine-industry grants, product donations, or speaking fees (Antonio, Candow, Forbes et al., 2021, competing-interest block). That disclosure travels with the citation; it is not moved to a footnote. The biopsy chemistry does not depend on it. The society synthesis does.
Creatine is not an anabolic steroid. The structures do not resemble each other, the legal classes do not overlap, and the mechanisms do not overlap (Antonio, Candow, Forbes et al., 2021). The comparison is folklore. It is recorded here because the folklore is load-bearing in public argument, not because it is a scientific near-miss.
03 The creatine / phosphocreatine system and ATP buffering
Skeletal muscle spends adenosine triphosphate (ATP) faster than oxidative phosphorylation or glycolysis can replace it during the first seconds of maximal contraction. The phosphagen system covers that gap. Creatine kinase (CK) transfers a phosphoryl from phosphocreatine (PCr) to adenosine diphosphate (ADP), regenerating ATP at the site of use. ATP concentration stays nearly constant while PCr falls. Recovery rebuilds PCr from ATP made by glycolysis and by mitochondria. Wallimann and colleagues described this as temporal and spatial buffering, not as a second fuel tank with its own combustion chemistry (Wallimann, Wyss, Brdiczka, Nicolay, and Eppenberger, 1992; Wallimann, Tokarska-Schlattner, and Schlattner, 2011). Grade: established.
Greenhaff, Casey, and colleagues connected that biochemistry to human muscle. In biopsy and 31-phosphorus magnetic-resonance work, creatine supplementation that raised total creatine also raised PCr availability and the rate of PCr resynthesis after intense contraction, and preserved intramuscular ATP better across repeated bouts (Greenhaff et al., 1993; Greenhaff et al., 1994; Casey, Constantin-Teodosiu, Howell, Hultman, and Greenhaff, 1996). Those are human mechanistic studies. They are strongly supported for the tasks they measured — brief, repeated, creatine-limited work. They are not a licence to treat a 10 km run as a phosphagen event.
The implication is task-shaped. A single maximal effort that lasts a fraction of a second is not creatine-limited in the same way as five more efforts with incomplete recovery. A continuous effort that is oxidative from the first minute is limited by oxygen delivery, glycogen, and economy, not by the size of the PCr pool. The performance literature in Part Three is readable only after that split. The market flattened it. The muscle did not.
04 Creatine kinase isoforms
CK is not one enzyme in one place. Cytosolic muscle-type CK (CK-M; the MM homodimer in skeletal muscle, MB in heart) sits near ATPases. Cytosolic brain-type CK (CK-B; BB in brain) serves neural tissue. Mitochondrial sarcomeric CK (sMtCK) and ubiquitous mitochondrial CK (uMtCK) sit in the intermembrane space and feed high-energy phosphoryl toward the cytosol as PCr. Wallimann’s compartmentation argument is that the isoforms are a shuttle, not a bag of soluble activity (Wallimann, Wyss, Brdiczka, Nicolay, and Eppenberger, 1992; Wallimann, Tokarska-Schlattner, and Schlattner, 2011). Grade: established as organisation; strongly supported as the reason a rise in the total creatine pool can change local ATP buffering without changing resting ATP.
Serum CK after hard training is a different object. It is a leak marker, not a measure of the shuttle. Exercise-related CK elevations are not evidence that oral creatine damaged muscle, and they are not evidence that it protected it (Baird, Graham, Baker, and Bickerstaff, 2012). Using one abbreviation for both facts is how laboratory folklore is born.
Isoform geography also explains why brain creatine is not a junior copy of muscle creatine. The brain expresses CK-B and uMtCK, makes some creatine locally, and sits behind a blood–brain barrier that the muscle does not have. Section 07 treats that compartment on its own terms. A muscle biopsy increment does not license a cognitive claim.
05 Endogenous biosynthesis: AGAT and GAMT
Humans make creatine in two enzymatic steps. L-Arginine:glycine amidinotransferase (AGAT; gene GATM) transfers the amidino group of arginine to glycine, producing guanidinoacetate and ornithine. Guanidinoacetate N-methyltransferase (GAMT; gene GAMT) then methylates guanidinoacetate with S-adenosylmethionine (SAM), producing creatine and S-adenosylhomocysteine. The classic inter-organ sketch puts most AGAT in kidney and most GAMT in liver; circulating creatine is then taken up by muscle and brain (Wyss and Kaddurah-Daouk, 2000; Brosnan, da Silva, and Brosnan, 2011; Brosnan and Brosnan, 2016). Brain also expresses the pathway locally. Tissue weights of flux are not uniform. Grade: established.
The methylation step is not free. Creatine synthesis is a substantial consumer of labile methyl groups. Dietary creatine spares that cost. Brosnan and colleagues treated that as chemistry and as a reason vegetarians run a different biosynthetic budget, not as a diet prescription (Brosnan, da Silva, and Brosnan, 2011; Brosnan and Brosnan, 2016). The urea-cycle connection is real — arginine is an AGAT substrate — and it is not a claim that creatine is a urea-cycle drug.
The same two enzymes, when lost, produce creatine-deficiency syndromes. Item, Stöckler-Ipsiroglu, Stromberger and colleagues identified AGAT deficiency (Item et al., 2001). Stöckler, Holzbach, Hanefeld and colleagues identified GAMT deficiency (Stöckler et al., 1994; Stöckler, Isbrandt, Hanefeld, Schmidt, and von Figura, 1996). Both are treatable in a way that creatine-transporter deficiency is not, because the tissue can still take up circulating creatine. Section 20 is the clinical home of that distinction. It is the strongest disease-use literature creatine has. It is not an argument that creatine “treats the brain” in people whose pathway is intact.
06 Transport (SLC6A8) and skeletal-muscle uptake
Muscle does not make enough creatine to fill itself. It concentrates circulating creatine through the sodium- and chloride-dependent creatine transporter SLC6A8 (CT1). Guimbal and Kilimann cloned the transporter from muscle (Guimbal and Kilimann, 1993). Nash, Giros and colleagues characterised the human gene and its expression (Nash et al., 1994). The protein is a solute carrier, not a channel. Uptake is saturable, Na+/Cl−-coupled, and down-regulated when intracellular creatine is already high (Wyss and Kaddurah-Daouk, 2000; Christie, 2007). That last property is why people who start with high stores gain less, and why “more powder” is not a linear loading function. Grade: established.
SLC6A8 defects cause X-linked creatine-transporter deficiency (Salomons et al., 2001). Oral creatine cannot bypass a transporter that does not work. That is the clinical converse of the muscle-loading literature: the same protein that makes Harris’s biopsy increment possible makes Salomons’s patients treatment-resistant. Section 20 returns to the syndrome. The physiology belongs here.
Insulin and exercise can increase net uptake in experimental settings, which is why some trials dissolved creatine in a carbohydrate drink (Green, Hultman, Macdonald, Sewell, and Greenhaff, 1996; Steenge, Simpson, and Greenhaff, 2000). Those are human uptake studies. They are strongly supported as a description of how those protocols moved the biopsy number. They are not a requirement that creatine be taken in fruit juice, and they are not used here as a timing rule. Antonio and Ciccone’s 2013 pre- versus post-training comparison is a small randomised trial in recreational bodybuilders. It is not load-bearing for any claim in this document (Antonio and Ciccone, 2013). Timing folklore is larger than that paper.
A second, related carrier, SLC16A12 (MCT12), has been discussed as a creatine or guanidinoacetate transporter in kidney and retina. It is not a substitute for SLC6A8 in skeletal muscle. This article treats SLC6A8 as the uptake step that the performance literature actually depends on.
07 Brain creatine as a separate compartment
Brain creatine is not muscle creatine with a hat on. The central nervous system expresses AGAT, GAMT, and SLC6A8, maintains its own PCr buffer, and loads more slowly from oral creatine than skeletal muscle does (Wyss and Kaddurah-Daouk, 2000; Andres, Ducray, Schlattner, Wallimann, and Widmer, 2008; Dolan, Gualano, and Rawson, 2019). Magnetic-resonance spectroscopy can show a brain-creatine rise after oral monohydrate, but the increment is smaller and slower than the vastus-lateralis increment, and it is not uniform across regions. Grade: strongly supported that the compartments are separable; emerging as a quantitative map of which regions move in which adults.
Two consequences follow.
A muscle-loading protocol is not a brain-loading protocol. Trials that report a cognitive change after a week of high oral intake may be seeing a muscle or a fatigue effect, a sleep effect, or a true central change. They have to show the compartment. Most do not.
A vegetarian or a sleep-deprived subject is not a well-slept omnivore. If the brain pool is already near its practical ceiling, oral creatine has less to do. That is the same ceiling logic as Figure 6, applied to a tissue that is harder to biopsy. Part Four is written under that constraint. Cognition claims that ignore it are spending the muscle literature.
DeVivo, Mercimek-Andrews, and the creatine-deficiency literature are the existence proof that brain creatine matters: when the pool is truly empty, the phenotype is intellectual disability, hypotonia, and seizures, not a missed chess game (Stöckler et al., 1994; Salomons et al., 2001; Mercimek-Andrews and Salomons, 2009). Deficiency is not a model of ordinary supplementation. It is a model of what the pool is for.
08 Dietary sources; vegetarian and omnivorous status
Creatine in food is a meat and fish fact. Skeletal muscle of terrestrial animals and the corresponding tissues of fish contain creatine in the grammes-per-kilogramme range; dairy and eggs contribute little; plants contribute none of nutritional account (Wyss and Kaddurah-Daouk, 2000; Brosnan and Brosnan, 2016). Antonio, Candow, Forbes et al. (2021), citing food-composition summaries, put a 6-ounce serving of meat near 0.7 g of creatine. That figure is a composition estimate, not an intake target. Habitual omnivorous diets in the papers below typically supply one to two grammes per day. Strict vegetarian and vegan diets supply near zero.
Burke, Chilibeck, Parise, Candow, Mahoney, and Tarnopolsky biopsied vegetarians and found lower muscle creatine than omnivores, and a larger rise after oral monohydrate (Burke et al., 2003). Watt, Garnham, and Snow related habitual meat intake to muscle creatine and to the increment after supplementation (Watt, Garnham, and Snow, 2004). Kaviani, Shaw, and Chilibeck’s later systematic review treated the vegetarian advantage as a baseline-store story, not as a moral property of the diet (Kaviani, Shaw, and Chilibeck, 2020). Grade: strongly supported that vegetarian muscle creatine is often lower and that the oral increment is often larger; not established that every vegetarian is a responder or every omnivore is saturated.
The dietary split does three jobs in this document. It predicts who has room under the ceiling (Section 10). It predicts who might show a cognitive signal (Section 17). It forbids treating an omnivorous student sample as a universal brain result. It does not rank diets.
09 Pharmacokinetics of oral monohydrate
Persky and Brazeau reviewed the clinical pharmacology of creatine: oral monohydrate is absorbed from the gut, appears in plasma within an hour, distributes into muscle against a concentration gradient via SLC6A8, and is either stored, converted to creatinine, or excreted when the filtered load exceeds tubular recovery (Persky and Brazeau, 2001; Persky, Brazeau, and Hochhaus, 2003). Harris and colleagues measured drink absorption in humans and showed that dissolved monohydrate raises plasma creatine; the same work sits behind the biopsy-loading papers (Harris, Nevill, Zhang, and Sale, 2002; Harris, Söderlund, and Hultman, 1992). Grade: established for the qualitative kinetic picture; trial-to-trial peak times and bioavailabilities vary with the drink, the meal, and the assay.
Two practical facts follow, still as descriptions of experiments.
Large single oral amounts increase the chance of gastrointestinal overflow. Ostojic and Ahmetovic, and later summaries in the ISSN file, reported diarrhoea and stomach discomfort more often when a serving exceeded about 10 g in a single bolus (Ostojic and Ahmetovic, 2008; Antonio, Candow, Forbes et al., 2021). That is a tolerance observation in the samples studied. It is not a serving instruction.
Urinary creatine rises once muscle uptake saturates. High urine creatine during a loading design is evidence of overflow, not of kidney injury (Hultman et al., 1996; Rawson, Clarkson, Price, and Miles, 2002). Interpreting that overflow as renal “strain” is the same error as interpreting a rise in serum creatinine as a fall in filtration. Section 21 keeps those objects apart.
Plasma creatine after a single oral amount is a poor guide to the muscle pool. The tissue integral over days is the quantity Harris biopsied. Marketing that cites a faster plasma peak for a new salt, without a matching biopsy, is citing the wrong compartment.
10 Muscle saturation, responders, and body-water change
Harris, Söderlund, and Hultman gave oral creatine to healthy men and biopsied vastus lateralis. Total creatine rose in resting muscle and rose further in muscle that had been exercised; some men moved little (Harris, Söderlund, and Hultman, 1992). Hultman, Söderlund, Timmons, Cederblad, and Greenhaff then showed that 20 g per day for 6 days and 3 g per day for 28 days produced similar muscle increments of about 20 percent in the men they studied (Hultman et al., 1996). Greenhaff’s group and Casey’s intramuscular work tied the larger increments to better PCr resynthesis and better maintenance of ATP across repeated contractions (Greenhaff et al., 1994; Casey, Constantin-Teodosiu, Howell, Hultman, and Greenhaff, 1996). Grade: established that oral monohydrate raises muscle creatine in most studied adults; established that the increment is larger when the starting store is lower.
Syrotuik and Bell classified responders and non-responders on biopsy and performance and found that the people who started lower, had more type II fibre, and had lower initial PCr were the people who gained (Syrotuik and Bell, 2004). That paper is a human classification study. It is strongly supported as a description of heterogeneity. It is the existence proof that “creatine works” is false as a universal. Already-saturated athletes are not a marketing problem. They are a ceiling.
Body mass often rises 0.5–2.0 kg in the first week of a high-intake design, and the early rise is largely water (Hultman et al., 1996; Kraemer and Volek, 1999). Creatine is osmotically active; SLC6A8 co-transports sodium; water follows. Ziegenfuss, Lowery, and Lemon measured acute fluid-volume changes over three days (Ziegenfuss, Lowery, and Lemon, 1998). Powers, Arnold, Weltman and colleagues found a rise in total body water with creatine that did not rearrange the intra- versus extracellular ratio in the way a dehydration scare requires (Powers et al., 2003). Longer training studies are mixed: some see no lasting change in total body water relative to muscle mass (Rawson, Stec, Frederickson, and Miles, 2011; Jagim et al., 2012; Spillane et al., 2009); Ribeiro and colleagues, in an eight-week resistance-training trial, saw total body water and intracellular water rise with creatine while the intracellular-water-to-muscle-mass ratio stayed similar to placebo (Ribeiro et al., 2020). Grade: established that short-term body-mass gain is largely water; strongly supported that this is not the same as extracellular dehydration; emerging to mixed as a long-term body-water story once muscle mass is changing.
The intracellular-water rise has been offered as an anabolic signal (Safdar, Yardley, Snow, Melov, and Tarnopolsky, 2008). That is a plausible mechanism paper, not a demonstration that creatine is a primary hypertrophic drug. Section 14 keeps the order: training volume first, water second, myosin heavy-chain stories third.
11 Loading versus daily accumulation as experimental designs (not prescriptions)
The field sometimes talks as if there were one correct way to take creatine. There is not. There are two experimental designs that fill the same muscle pool on different calendars.
Hultman et al. (1996) is the comparison that matters. In healthy men, a short high-intake design (20 g per day for 6 days) and a longer daily-accumulation design (3 g per day for 28 days) reached similar muscle creatine. Harris, Söderlund, and Hultman (1992) is the demonstration that oral creatine can move the biopsy at all. Later society papers restated both designs as options, not as a hierarchy of virtue (Buford et al., 2007; Kreider et al., 2017; Antonio, Candow, Forbes et al., 2021). Grade: established as a description of two ways the men in those studies reached a similar ceiling.
This article uses those numbers only as reported experimental parameters, always with the population and the duration attached. It does not recommend either design for any person. A researcher choosing a protocol is not a reader being told what to swallow.
Two further design facts, still not prescriptions.
Splitting a large daily amount into smaller servings was used in the biopsy papers and in later trials, in part because single boluses above about 10 g were more often associated with diarrhoea in the samples that reported it (Ostojic and Ahmetovic, 2008). That is a methods note.
Cessation lets the pool fall. Hultman et al. (1996) and later wash-out observations put the return toward baseline on the order of weeks, consistent with the 1–2 percent daily conversion to creatinine (Wyss and Kaddurah-Daouk, 2000). Creatine is not a depot steroid. It is a pool with a leak.
The word “loading” in a shop is a sales verb. In this document it is a study arm.
12 Strength and maximal force
The strength literature is large, mixed, and better than the folklore that every lift goes up.
Branch’s meta-analysis of creatine supplementation found a performance effect that concentrated in anaerobic and resistance tasks, with a mean effect that was real and not enormous (Branch, 2003). Lanhers, Pereira, Naughton and colleagues later split the strength question by limb: a 2015 systematic review and meta-analysis on lower-limb strength and a 2017 paper on upper-limb strength both found a favourable effect of creatine versus placebo, larger in some multi-week training designs than in acute-only designs (Lanhers et al., 2015; Lanhers et al., 2017). Rawson and Volek’s narrative review of creatine and resistance-exercise performance reached the same practical shape: more repetitions, more work in a set, more force where the task was creatine-limited (Rawson and Volek, 2003). Grade: strongly supported that creatine monohydrate, after stores rise, improves strength and repeated-force outcomes in the populations those papers pooled; not established as a universal one-repetition-maximum drug for already-saturated lifters.
Individual trials that still earn their citation include Earnest, Snell, Rodriguez, Almada, and Mitchell on strength and body composition in resistance-trained men (Earnest et al., 1995); Becque, Lochmann, and Melrose on a six-week arm-training design (Becque, Lochmann, and Melrose, 2000); Vandenberghe, Goris, Van Hecke, Van Leemputte, Vangerven, and Hespel on women training for ten weeks (Vandenberghe et al., 1997); and Volek, Duncan, Mazzetti and colleagues on fibre and performance responses (Volek et al., 1999). Kreider, Ferreira, Wilson and colleagues reported in-season gains in college football players on a creatine-containing supplement (Kreider et al., 1998). Those are randomised or controlled training studies. They are strongly supported inside their samples. They are not a claim that every athletic department result was clean of other ingredients.
ISSN position stands collected this file and graded creatine as effective for high-intensity exercise (Buford et al., 2007; Kreider et al., 2017). Wax, Kerksick, Jagim, Mayo, Lyons, and Kreider’s 2021 nutrients review is the same literature in a later frame (Wax et al., 2021). Cooper, Naclerio, Allgrove, and Jimenez’s 2012 update sits between the two ISSN papers (Cooper et al., 2012). These are society and narrative syntheses.
Conflicts of interest, kept in place Several authors of Buford et al. (2007), Kreider et al. (2017), Antonio, Candow, Forbes et al. (2021), and Wax et al. (2021) disclose consulting, speaking, grant, or industry relationships with companies that manufacture or market creatine or other sport-nutrition products. The disclosure is part of the citation. The Harris–Hultman biopsy record and the Lanhers meta-analyses do not depend on it. The society sentence “creatine is safe and effective” does. This article uses the society papers as maps of a literature, not as a second data set.
A one-repetition-maximum that does not repeat is a poor creatine test. A set that does is a better one. That is physiology, not a coaching slogan.
13 Power and repeated-sprint work
If creatine has a home, this is it.
Casey, Constantin-Teodosiu, Howell, Hultman, and Greenhaff showed that the men whose intramuscular ATP and PCr were better preserved across repeated bouts were the men who had raised muscle creatine (Casey et al., 1996). Greenhaff’s earlier papers tied PCr resynthesis to the same pool (Greenhaff et al., 1993; Greenhaff et al., 1994). Kreider et al. (1998), Earnest et al. (1995), and a long tail of repeated-sprint and isokinetic studies sit on that mechanism. Branch’s meta-analysis found the effect where the task was anaerobic and intermittent (Branch, 2003). Grade: established to strongly supported for brief, repeated, high-intensity work in people whose stores rose; not established for a single unrepeated peak-power test in a saturated athlete.
The warrant is the one in the argument map. A rise in the relevant tissue store plus a matching task physiology licenses a performance claim for that task class. Adjacent outcomes do not inherit it. A Wingate that is repeated with short rest is creatine-shaped. A three-hour ride is not, except insofar as body mass, glycogen, or sprint finishes intervene (Section 15).
Non-responders remain. Syrotuik and Bell (2004) did not vanish because a meta-analysis was later. A trial that does not verify the muscle increment is testing an intention to supplement, not a rise in PCr. That is why this document keeps saying “in people whose stores can still rise.” The clause is not a hedge. It is the independent variable.
14 Resistance training and hypertrophy
This is the first overextension that has some truth in it.
Willoughby and Rosene reported myosin heavy-chain expression changes with creatine plus resistance training (Willoughby and Rosene, 2001). Volek et al. (1999) reported fibre-area changes. Becque, Lochmann, and Melrose (2000) and Earnest et al. (1995) reported lean-mass and strength changes. Nissen and Sharp’s meta-analysis of dietary supplements and lean mass put creatine among the few that moved the number (Nissen and Sharp, 2003). Chilibeck, Kaviani, Candow, and Zello’s meta-analysis in older adults found additional lean tissue and strength when creatine was added to resistance training (Chilibeck et al., 2017). Delpino, Figueiredo, Forbes, Candow, and McCarthy later examined age, sex, and training status as modifiers (Delpino et al., 2022). Grade: strongly supported that creatine plus resistance training can increase lean mass more than training alone in some samples; not established that creatine is a primary anabolic; emerging as to how much of the lean-mass increment is dry contractile protein rather than water.
Forbes, Candow, Krentz, Roberts, and Chilibeck’s systematic review and meta-analysis on fat mass in older adults found a small favourable effect on percent fat with creatine plus training, without a matching large loss of absolute fat (Forbes et al., 2019). That paper is a reason to stop saying creatine “makes you fat.” It is not a fat-loss drug. Antonio, Candow, Forbes et al. (2021) collected the same nulls on fat-mass gain. The early scale rise is water (Section 10).
Two limits have to stay in the same paragraph as the positives.
First, creatine without a training stimulus is a poor hypertrophy agent. Gualano and colleagues, and Lobo, de Oliveira Sales, and the two-year postmenopausal studies, found that creatine alone did not produce the lean-mass change that creatine plus training did (Gualano et al., 2014; Lobo et al., 2015; Sales et al., 2020). The mechanism that is easiest to defend is still the phosphagen one: more work in the session, more total tension over weeks. Safdar et al. (2008) and cell-swelling arguments are plausible adjuncts. They are not a substitute for the training file.
Second, dual-energy X-ray absorptiometry and bioimpedance call water-rich tissue “lean.” A week-one increase is not a new myofibril. Trials that last months and measure fibre area, as Volek et al. (1999) and Willoughby and Rosene (2001) did, are the ones that can talk about hypertrophy without blushing. Even those effects are smaller than the poster. The popular claim treated a phosphagen as if it were a steroid. The trials did not.
15 Endurance contexts and recovery
Creatine is not an endurance supplement in the VO2max sense. Continuous efforts that are oxidative from the first minutes do not fail for lack of PCr. Body-mass gain can be a cost in weight-bearing endurance. Society reviews that are otherwise friendly to creatine say this plainly (Kreider et al., 2017; Wax et al., 2021; Cooper et al., 2012). Grade: established that creatine is not a general endurance ergogenic; emerging for selected endurance settings that still contain repeated surges, finishes, or glycogen problems.
Two adjacent literatures are sometimes spent as if they were that missing endurance effect.
Creatine plus carbohydrate has been reported to raise muscle glycogen more than carbohydrate alone (Green, Hultman, Macdonald, Sewell, and Greenhaff, 1996; Nelson, Arnall, Kokkonen, Day, and Evans, 2001; van Loon et al., 2004). Those are human muscle-glycogen studies. They are strongly supported as a storage observation in the protocols used. They are not a race-day drink plan, and they are not a demonstration that creatine without carbohydrate is a glycogen drug.
Recovery and soreness are mixed. Some trials report faster return of force or lower soreness ratings after damaging exercise; others do not. Twycross-Lewis, Kilduff, Wang, and Pitsiladis reviewed thermoregulation and performance and did not find a closed story (Twycross-Lewis et al., 2016). This document treats recovery as emerging and task-specific, not as a second established indication.
A creatine-plus-HMB trial in elite male endurance athletes is in the local full-text file (Fernández-Landa et al., 2020). It is a combination study. It does not isolate monohydrate as an endurance agent, and it is not used here as one.
16 Women, adolescents, and response heterogeneity
Vandenberghe et al. (1997) is still the cleanest long training trial in women: ten weeks of resistance training with creatine raised muscle PCr, strength, and fat-free mass versus placebo. de Guingand, Palmer, Snow, Davies-Tuck, and Ellery’s systematic review and meta-analysis of adverse outcomes in females did not find a safety signal that would set women apart from the healthy-adult file (de Guingand et al., 2020). Ellery, Walker, and Dickinson reviewed creatine across the reproductive cycle as a physiological essay, not as a dosing manual (Ellery, Walker, and Dickinson, 2016). Grade: strongly supported that women are not non-responders as a class; not established that every female trial matches the male mean increment, because baseline stores, diet, and fat-free mass differ.
Adolescents are a thinner file. Jagim, Stecker, Harty, Erickson, and Kerksick reviewed safety in active adolescents and did not find a pattern of laboratory harm in the trials they collected; they also noted that most performance studies did not run a clinical panel (Jagim et al., 2018). Unnithan, Veehof, Vella, and Kern asked whether there was a physiologic basis for use in children and adolescents (Unnithan et al., 2001). Clinical paediatric trials — Tarnopolsky et al. (2004) in Duchenne muscular dystrophy; Hayashi et al. (2014) in childhood-onset lupus; Sakellaris et al. (2006) in traumatic brain injury — are disease studies, not sports studies. Grade: emerging for adolescent performance; strongly supported that the short clinical trials that measured laboratories did not show a renal or hepatic signal in those disease samples. This is not a recommendation that adolescents supplement.
Heterogeneity is the rule, not the exception. Syrotuik and Bell (2004), Harris, Söderlund, and Hultman (1992), Burke et al. (2003), and Rawson, Clarkson, Price, and Miles (2002) on young versus older muscle PCr all say the same thing in different dialects: the increment follows the vacancy. Age, sex, diet, fibre type, and current store are predictors. They are not moral categories.
| Population / setting | Outcome usually tested | What the file supports | Grade | Limits |
|---|---|---|---|---|
| Healthy men, biopsy | Muscle total Cr / PCr | Rise in most, larger from a low baseline | established | Non-responders exist (Harris 1992; Hultman 1996; Syrotuik and Bell 2004) |
| Resistance-trained adults | Strength, repeated force | Favourable pooled effect | strongly supported | Already-high stores flatten the increment (Branch 2003; Lanhers 2015, 2017) |
| Repeated-sprint / interval | Power across bouts | Best match to PCr physiology | established / strongly supported | Single unrepeated peaks are weaker tests (Casey 1996; Greenhaff 1994) |
| Women in training trials | Strength, fat-free mass | Respond when stores rise | strongly supported | Fewer long RCTs than in men (Vandenberghe 1997; de Guingand 2020) |
| Vegetarians | Muscle Cr increment; some cognition tests | Larger muscle increment; cognition not general | strongly supported (muscle) | Not every vegetarian is depleted (Burke 2003; Rae 2003) |
| Older adults + training | Lean mass, strength | Additional lean mass and strength in metas | strongly supported | Creatine alone is much weaker (Chilibeck 2017; Gualano 2014) |
| Adolescents / youth sport | Performance, safety labs | Thin performance file; no clear lab harm in small clinical samples | emerging | Not a sports recommendation (Jagim 2018; Tarnopolsky 2004) |
| Well-slept omnivores, cognition | Memory / executive tests | Inconsistent; often null | not established | Do not spend vegetarian or sleep-loss trials here (Avgerinos 2018; Prokopidis 2023) |
17 Cognition and sleep deprivation
The cognitive literature is real and smaller than the poster.
Rae, Digney, McEwan, and Bates randomised vegetarians to creatine and found improvements on backward digit span and a Raven-style task (Rae et al., 2003). That is a randomised trial in a low-habitual-intake sample. McMorris, Harris, Howard and colleagues, and a related sleep-deprivation experiment, reported better performance on selected executive or psychomotor tests after creatine when sleep had been lost (McMorris et al., 2006; McMorris et al., 2007). Those are sleep-deprivation trials. They are strongly supported as evidence that a stressed, possibly depleted brain can show a creatine signal. They are not a well-slept omnivore result.
Avgerinos, Spyrou, Bougioukas, and Kapogiannis’s systematic review concluded that the evidence was promising and thin, and that the positive trials clustered in vegetarians or in stressed conditions (Avgerinos et al., 2018). Prokopidis, Giannos, Triantafyllidis, Kechagias, Forbes, and Candow’s meta-analysis found a favourable effect on memory in a pooled sample, with age and other moderators discussed (Prokopidis et al., 2023). Dolan, Gualano, and Rawson’s narrative review of brain creatine, cognitive processing, and traumatic brain injury is the cautious synthesis this document prefers: the compartment is real; the loading is slower; the cognition claim is conditional (Dolan, Gualano, and Rawson, 2019). Grade: strongly supported in depletion, low habitual intake, or sleep loss; emerging as a pooled memory effect that may not survive restriction to well-slept high-intake omnivores; not established as a nootropic for that last group.
A recent meta-analysis can be read as broader than its ingredients. That is objection O3 in the argument map. The honest reply is to keep the ingredients visible. If the pooled effect is being carried by vegetarians, older adults, and sleep-loss trials, it is not a licence to sell creatine as coffee. Recency is not automatic generality — the Tesamorelin lesson, applied here without the peptide.
Cook, Crewther, Kilduff, Drawer, and Gaviglio’s sleep-deprivation trial with caffeine or creatine is a skill-execution study, not a memory battery (Cook et al., 2011). It does not rescue a general nootropic claim. It does illustrate the actual experimental setting: a brain that is short of something.
18 Depression and neurology, only where the human record warrants
Creatine is not an antidepressant with a sports-nutrition hobby.
Roitman, Green, Osher, Karni, and Levine reported an open-label add-on signal in a small depressed sample (Roitman et al., 2007). Lyoo, Yoon, Kim and colleagues ran a randomised trial of creatine add-on to a selective serotonin-reuptake inhibitor in women with major depressive disorder and reported a faster response (Lyoo et al., 2012). Those are small psychiatric trials. They are emerging as a possible adjunct signal. They are not a stand-alone treatment, not a mechanism proof, and not a reason to treat the sports literature as a mood literature.
Neurological claims outside creatine-deficiency syndromes are thinner still. Traumatic-brain-injury work in children (Sakellaris et al., 2006) was an open-label randomised pilot, not a definitive neuroprotection trial. Animal and in-vitro papers on Huntington disease, Parkinson disease, and amyotrophic lateral sclerosis have not produced a matching human standard of care. Balestrino and Adriano’s review of creatine beyond sport is a map of hopes and small files, not a new indication (Balestrino and Adriano, 2019). Grade: emerging for small depression add-on trials; plausible as a bioenergetic hypothesis in selected neurological diseases; not established as treatment for any of them in this document.
The strong clinical story is still creatine-deficiency syndrome (Section 20). Spending that story on ordinary low mood is how a rare-disease literature gets used as advertising.
19 Older adults, sarcopenia, and bone
Older muscle has less PCr, loses type II fibre, and responds to resistance training. Creatine sits on that intersection without becoming a sarcopenia drug of its own.
Brose, Parise, and Tarnopolsky randomised older adults to creatine plus training and found strength and lean-mass changes (Brose, Parise, and Tarnopolsky, 2003). Devries and Phillips’s systematic review treated creatine as an adjunct to resistance training in aging muscle (Devries and Phillips, 2014). Chilibeck, Kaviani, Candow, and Zello’s meta-analysis is the lean-mass and strength statement this document uses: creatine during resistance training increased lean tissue mass and strength in older adults versus training plus placebo (Chilibeck et al., 2017). Candow, Forbes, Chilibeck and colleagues have reviewed the same file for falls, inflammation, and therapeutic variables (Candow et al., 2019a; Candow et al., 2019b). Gualano, Rawson, Candow, and Chilibeck’s aging review is the cautious companion (Gualano, Rawson, Candow, and Chilibeck, 2016). Grade: strongly supported that creatine plus resistance training can add lean mass and strength in older samples; not established that creatine without training prevents sarcopenia.
The alone-arm matters. Gualano et al. (2014) in older women, Lobo et al. (2015) at one year, and Sales, de Oliveira Sales, and the two-year postmenopausal trial (Sales et al., 2020) found that creatine without a training programme did not deliver the musculoskeletal change the combination did. Short-term scale gains in older adults can still be water (Rawson, Clarkson, Price, and Miles, 2002). Treating a seven-day fat-free-mass rise as hypertrophy is the same error as in the young.
Bone is mixed. Chilibeck, Candow, and colleagues have reported, in training-plus-creatine designs, some favourable changes in bone mineral content or a slower loss at the hip; the two-year creatine-alone trial in postmenopausal women did not improve bone mass or geometry (Chilibeck et al., 2015; Sales et al., 2020; Candow et al., 2019a). Grade: emerging for bone as an adjunct to training; not established as an osteoporosis intervention.
Forbes et al. (2019) on fat mass in older adults, already cited, is the reason this section does not treat creatine as a fattening agent. The aging claim that survives is modest and conjunctive: training first, creatine as a store that can make the training slightly more productive in people whose stores can still rise. The claim that does not survive is creatine as a longevity molecule. That paper has not been written.
20 Creatine-deficiency syndromes
Three inborn errors empty the pool.
AGAT (GATM) deficiency blocks the first biosynthetic step (Item et al., 2001). GAMT deficiency blocks the second and accumulates guanidinoacetate, which is itself neurotoxic (Stöckler et al., 1994; Stöckler, Isbrandt, Hanefeld, Schmidt, and von Figura, 1996). SLC6A8 deficiency blocks uptake into brain and muscle and is X-linked (Salomons et al., 2001; van de Kamp, Betsalel, Mercimek-Mahmutoglu et al., 2013). All three present with developmental delay, intellectual disability, and a low creatine peak on brain magnetic-resonance spectroscopy. GAMT adds a movement and epilepsy burden from guanidinoacetate. The transporter defect adds the cruel fact that the brain cannot be refilled from the blood.
Oral creatine monohydrate has been used as metabolic treatment in AGAT and GAMT deficiency, with biochemical correction that is often clearer than cognitive rescue if treatment starts late (Stöckler et al., 1996; Mercimek-Andrews and Salomons, 2009). Transporter deficiency does not respond in the same way. That is clinical genetics, not a sports footnote. Grade: established as disease entities and as a demonstration that the creatine pool is not optional in the developing brain; strongly supported that AGAT and GAMT defects are the setting in which oral creatine is a disease-modifying metabolic therapy; established that SLC6A8 deficiency is a different problem.
This is the clinical literature that warrants the word “treatment.” Depression add-on trials do not. Sports cognition trials do not. A article that spent more pages on a mood stack than on AGAT would have its priorities backwards.
21 Kidney-function claims and creatinine interpretation
The kidney story is the most durable piece of creatine folklore, and it is mostly a measurement error.
Creatine and phosphocreatine convert non-enzymatically to creatinine at about 1–2 percent of the pool per day (Wyss and Kaddurah-Daouk, 2000). Serum creatinine is a filtration marker only if the generation rate is stable. Increase the precursor pool, or eat a large meat load, and generation rises. Estimated glomerular filtration rate (eGFR) equations that are built on serum creatinine will then look worse without a nephron having failed. Cystatin C is generated by all nucleated cells and is not a creatine metabolite. Measured GFR — inulin, iohexol, 51-chromium-EDTA — is the reference. Pline and Smith reviewed creatine and the kidney with that distinction in view (Pline and Smith, 2005). Persky and Rawson’s safety chapter tabulated serum creatinine across supplementation studies: many showed no rise, some showed a rise that stayed inside reference intervals, few showed a rise above them (Persky and Rawson, 2007). Grade: established that creatinine is a metabolite of creatine; established that a creatinine-only eGFR can move after oral creatine without a matching filtration diagnosis.
Poortmans, Auquier, Renaut and colleagues measured renal responses to short-term creatine in men and did not find a filtration injury (Poortmans et al., 1997). Poortmans and Francaux followed longer use and wrote the adverse-effect reviews that still organise this file (Poortmans and Francaux, 1999; Poortmans and Francaux, 2000). Gualano, de Salles Painelli, Roschel and colleagues ran a randomised, double-blind trial in people with type 2 diabetes — a higher-risk kidney population than college football — and did not find impaired kidney function on the markers they used (Gualano et al., 2011). de Souza e Silva and colleagues’ later systematic review and meta-analysis of renal function after creatine did not produce a healthy-adult harm signal (de Souza e Silva et al., 2019). Grade: strongly supported that recommended-range experimental exposures have not reduced measured or carefully interpreted kidney function in healthy adults; not established as a licence for use in established kidney disease.
The 1998 Lancet case report that launched the folklore was a young man with long-standing focal segmental glomerulosclerosis and cyclosporine exposure whose serum creatinine rose after oral creatine (Pritchard and Kalra, 1998). Greenhaff, and Poortmans and Francaux, wrote letters pointing at the metabolite problem and at the already-published negative renal studies (Greenhaff, 1998; Poortmans and Francaux, 1998). Gualano, Roschel, Lancha, Brightbill, and Rawson later reviewed further case reports and found them confounded by pre-existing disease, other supplements, extreme amounts, or anabolic steroids (Gualano et al., 2012). Case reports are not a reason to ignore a signal in a special population. They are a reason not to let a confounded letter overturn a controlled file.
Antonio, Candow, Forbes et al. (2021) restated the healthy-adult renal conclusion. The author conflicts disclosed in Section 12 remain in force for that restatement. The Poortmans and Gualano trials do not need it.
| Marker | What it measures | Moves after oral creatine? | Use in this literature |
|---|---|---|---|
| Serum creatinine | A creatine metabolite plus a filtration residue | Often yes, because generation rose | Cannot diagnose a fall in GFR by itself (Wyss and Kaddurah-Daouk, 2000; Pline and Smith, 2005) |
| Creatinine-based eGFR | An equation that assumes stable creatinine generation | Can look worse when generation rises | Mis-classifies loaded subjects unless the assumption is checked |
| Cystatin C / cystatin-based eGFR | A non-creatine filtration marker | Should not rise merely because the creatine pool rose | The practical second marker when creatinine moved |
| Measured GFR (inulin, iohexol, radio-tracer) | Filtration | The quantity the folklore claims fell | Sparse in sports trials; where used, not the Pritchard story |
| Urinary creatine | Overflow once muscle uptake saturates | Yes, during high-intake designs | Evidence of overflow, not of nephron injury (Hultman 1996; Rawson 2002) |
| Urinary albumin, measured clearance | Kidney damage / true filtration | The outcomes that would change the verdict | Required before a healthy-adult harm claim is paid |
This table is an interpretation key. It is not a clinical protocol and it is not a recommendation that any person start, continue, or stop creatine.
22 Dehydration, cramping, GI effects, long-term use, contaminants
The dehydration-and-cramp story began as a theory: creatine pulls water into muscle, so the rest of the body must run dry, especially in heat. Terjung, Clarkson, Eichner and colleagues, writing an American College of Sports Medicine roundtable with limited data, advised caution in people who were cutting weight or exercising in the heat (Terjung et al., 2000). Dalbo, Roberts, Stout, and Kerksick later reviewed the claim and found it wanting (Dalbo et al., 2008). Greenwood, Kreider, Melton and colleagues monitored NCAA football players through a hot season and did not find more cramping, heat illness, or dehydration in those who chose creatine; they reported fewer of some of those events (Greenwood et al., 2003). Self-report surveys that found perceived cramping did not control other supplements or the amount ingested (Greenwood, Farris, Kreider, Greenwood, and Byars, 2000). Chang, Wu, Yang and colleagues, in a small haemodialysis series, reported fewer cramps after creatine — a clinical observation, not a sports trial (Chang et al., 2002). Grade: strongly supported that experimental and in-season monitoring files do not validate creatine as a dehydrator or cramp-inducer in healthy athletes; not established that creatine is a cramp medicine.
Gastrointestinal effects are the adverse event this document does not talk away. Ostojic and Ahmetovic (2008) and the ISSN summaries report diarrhoea and stomach discomfort, more often with large single boluses (Antonio, Candow, Forbes et al., 2021). That is a real, usually transient, tolerance finding. It is not kidney failure and it is not proof of a bad product, though a bad product can add to it.
Hair loss is a single-trial story. van der Merwe, Brooks, and Myburgh gave college rugby players 25 g per day for 7 days then 5 g per day for 14 days and reported a rise in dihydrotestosterone that stayed inside clinical limits, from a lower baseline in the creatine arm, without a rise in total testosterone (van der Merwe, Brooks, and Myburgh, 2009). Hair was not measured. The finding has not become a replicated endocrine literature (Antonio, Candow, Forbes et al., 2021). Grade: not established. One small hormone trial is not a dermatology indication and it is not a demonstration of safety either. It is a loose end.
Long-term healthy-adult use, on the time scale of months to a few years in training studies and in the postmenopausal trials, has not produced a systematic organ-toxicity pattern in the papers already cited (Poortmans and Francaux, 2000; Lobo et al., 2015; Sales et al., 2020; Kreider et al., 2017). That is strongly supported as a negative in healthy kidneys and livers at the experimental exposures used. It is not a lifetime surveillance system.
Contaminants are a product-quality problem, not a creatine-molecule problem. Sport supplements have been found, in other surveys, to contain undeclared stimulants or steroids. A creatine label does not certify the bottle. This document can grade the molecule. It cannot grade an untested powder. Third-party testing is a procurement fact, not a physiology finding, and it is not prescribed here.
23 Formulation comparison: monohydrate as the reference standard
The reference standard is creatine monohydrate because that is the form Harris biopsied, Hultman compared, Greenhaff and Casey mechanised, and the meta-analyses pooled. Other forms are judged by the creatine they deliver to tissue, not by the novelty of the salt.
Spillane, Schoch, Cooke and colleagues compared creatine ethyl ester with monohydrate in a resistance-training trial and found that ethyl ester did not raise muscle creatine better; serum creatinine rose more, consistent with easier conversion of the ester to creatinine (Spillane et al., 2009). Jagim, Oliver, Sanchez and colleagues compared a buffered creatine with monohydrate and found no advantage in muscle creatine, body composition, or training adaptations (Jagim et al., 2012). Herda, Beck, Ryan and colleagues tested a polyethylene-glycosylated creatine and did not displace monohydrate as a loading agent (Herda et al., 2009). Jäger, Purpura, Shao, Inoue, and Kreider reviewed forms and stability and kept monohydrate as the standard (Jäger et al., 2011). Grade: established that monohydrate loads muscle in humans; strongly supported that ethyl ester is not superior and can raise creatinine; not established that any alternative salt, on a like-for-like creatine dose with a biopsy or a matching performance test, has displaced monohydrate.
A different salt can still win on taste, solubility, or gastrointestinal comfort in a given person. That is a palatability observation. It is not a tissue-loading victory (argument-map objection O5). Powder grammes are not creatine grammes. Hydrochloride, nitrate, magnesium, and pyruvate salts carry less creatine per gramme of powder than the monohydrate’s 87.9 percent. A trial that does not match the creatine dose is not a head-to-head.
Liquid stability is a separate claim. Monohydrate in a warm drink becomes creatinine over time (Wyss and Kaddurah-Daouk, 2000; Jäger et al., 2011). A ready-to-drink that does not show remaining creatine at the end of shelf life is not “more advanced.” It is empty.
| Form | Creatine delivery claim | Human loading evidence | Performance / body-composition vs MH | Grade vs monohydrate |
|---|---|---|---|---|
| Creatine monohydrate | ~87.9% creatine by mass; reference | Harris 1992; Hultman 1996; decades of biopsies | The pooled strength and repeated-bout file | Reference standard |
| Creatine ethyl ester | Better uptake / less conversion | Spillane 2009: not superior; creatinine rose | No advantage in that training trial | Inferior or not superior |
| Buffered creatine (“Kre-Alkalyn” class) | Less conversion in the gut | Jagim 2012: no greater muscle Cr | No greater training adaptation in that trial | Not superior |
| PEG-creatine | Enhanced absorption | Herda 2009: did not displace MH | Not a replacement loading literature | Not superior |
| HCl / nitrate / Mg / pyruvate salts | Smaller dose, better solubility | Must match creatine dose, not powder dose | No like-for-like biopsy literature that wins | Unproven superiority |
| Liquid / RTD creatine | Convenience | Stability to creatinine is the unshown step | Not a loading data set | speculative until remaining Cr is measured |
This table is a comparison of evidence classes. It is not a purchasing recommendation.
24 Claim/evidence ledger and the red-team question
The red-team question that organises this article is the one the argument map named:
Where is creatine’s unusually strong reputation scientifically justified, and where has that reputation encouraged overextension into weaker claims?
The justified core is narrow and strong. Oral monohydrate raises muscle creatine in most studied adults, more so when baseline stores are low (Harris, Söderlund, and Hultman, 1992; Hultman et al., 1996; Syrotuik and Bell, 2004; Burke et al., 2003). The phosphagen buffers ATP in brief, repeated contraction (Wallimann et al., 1992; Greenhaff et al., 1994; Casey et al., 1996). Human performance trials cluster positive where the task is creatine-limited (Branch, 2003; Lanhers et al., 2015; Lanhers et al., 2017; Rawson and Volek, 2003). Healthy-adult kidney markers, properly read, do not show a filtration injury (Poortmans and Francaux, 2000; Gualano et al., 2011). Alternative forms have not beaten monohydrate on tissue loading (Spillane et al., 2009; Jagim et al., 2012; Jäger et al., 2011).
The overextensions are also specific. Hypertrophy is real as a smaller, training-dependent, partly water-confounded increment, not as a primary anabolic (Willoughby and Rosene, 2001; Nissen and Sharp, 2003; Chilibeck et al., 2017; Gualano et al., 2014). Cognition is real in depletion, sleep loss, or low habitual intake, and is not a nootropic for well-slept omnivores (Rae et al., 2003; McMorris et al., 2006; Avgerinos et al., 2018; Prokopidis et al., 2023). Kidney harm in healthy adults is a creatinine story wearing a GFR costume (Wyss and Kaddurah-Daouk, 2000; Pritchard and Kalra, 1998; Pline and Smith, 2005). Other salts outperform monohydrate only in advertisements.
| Claim | Grade | Best evidence class | Limit |
|---|---|---|---|
| Oral monohydrate raises muscle creatine in most adults | established | Human vastus-lateralis biopsy (Harris 1992; Hultman 1996) | Non-responders; already-high stores (Syrotuik and Bell 2004) |
| Increment is larger when baseline is lower, including many vegetarians | strongly supported | Biopsy + diet studies (Burke 2003; Watt 2004; Kaviani 2020) | Not every vegetarian is a responder |
| Improves high-intensity, repeated-bout work after stores rise | established / strongly supported | Mechanistic biopsy + RCTs + metas (Casey 1996; Branch 2003; Lanhers 2015/2017) | Task must be creatine-limited |
| Improves maximal strength in training designs | strongly supported | RCTs and strength metas (Vandenberghe 1997; Lanhers 2015/2017; Rawson and Volek 2003) | Weaker for a single unrepeated 1RM in saturated athletes |
| Primary anabolic / steroid-like hypertrophy | not established (overextension) | Lean-mass RCTs and metas are smaller and water-confounded (Nissen and Sharp 2003; Chilibeck 2017) | Training-dependent; creatine alone is weak (Gualano 2014; Sales 2020) |
| Additional lean mass with resistance training | strongly supported, smaller than the poster | Training RCTs, older-adult meta (Willoughby 2001; Chilibeck 2017; Delpino 2022) | DXA “lean” includes water |
| General endurance / VO2max agent | not established | Society reviews and physiology (Kreider 2017; Wax 2021) | Mass cost; not a PCr-limited task |
| Glycogen storage with carbohydrate | strongly supported as a storage finding | Human glycogen studies (Green 1996; Nelson 2001; van Loon 2004) | Not an isolated-creatine endurance effect |
| Nootropic for well-slept high-intake omnivores | not established (overextension) | Positives cluster in vegetarians / sleep loss (Rae 2003; McMorris 2006; Avgerinos 2018) | Prokopidis 2023 must not be over-generalised |
| Cognition under sleep loss or low habitual intake | strongly supported / emerging | RCTs and SR/MA in those contexts | Compartment rarely verified |
| Antidepressant or general neurological treatment | emerging at best; CDS excepted | Small add-on trials (Roitman 2007; Lyoo 2012) | CDS is the strong clinical file (Section 20) |
| Older adults: lean mass and strength with training | strongly supported | Meta-analysis (Chilibeck 2017); RCTs (Brose 2003) | Not a stand-alone sarcopenia therapy |
| Bone protection | emerging / mixed | Training adjunct signals vs two-year alone null (Chilibeck 2015; Sales 2020) | Not an osteoporosis indication |
| Kidney injury in healthy adults at research exposures | not established (folklore) | Controlled renal studies (Poortmans 1997/2000; Gualano 2011; de Souza e Silva 2019) | Does not license use in established CKD |
| Serum creatinine can rise without a GFR fall | established | Chemistry + marker logic (Wyss 2000; Pline and Smith 2005) | Still check a second marker if the number matters |
| Dehydration and cramping as a creatine effect | not established | In-season monitoring and reviews (Greenwood 2003; Dalbo 2008) | Early ACSM caution was data-thin (Terjung 2000) |
| Hair loss / DHT-driven baldness | not established | One small rugby trial, hair not measured (van der Merwe 2009) | Unreplicated endocrine loose end |
| Other salts outperform monohydrate on tissue loading | not established (overextension) | Head-to-heads (Spillane 2009; Jagim 2012; Herda 2009; Jäger 2011) | Palatability is not a biopsy |
| Timing (pre vs post workout) is load-bearing | not established | Antonio and Ciccone 2013 is a small trial | Explicitly not used as a pillar |
| AGAT/GAMT deficiency responds to oral creatine | strongly supported | Clinical genetics (Item 2001; Stöckler 1994/1996) | Transporter deficiency does not |
The conclusion does not exceed the premises if hypertrophy and cognition are graded as smaller, more conditional, and more often surrogate. It would exceed them if it were read as “creatine never helps those outcomes.” The ledger is the difference.
What would change the verdict is also specific. A large, preregistered literature showing no muscle-creatine rise after monohydrate in typical adults would break the foundation. A body of high-quality trials in which verified loading left high-intensity repeated-bout performance reliably null would break the performance claim. Head-to-head trials in which another form saturated muscle creatine better than an equal creatine dose of monohydrate, with matching performance, would retire the reference standard. A well-controlled demonstration that creatine, not creatinine artefact, reduces GFR in healthy adults at common research exposures would reopen the kidney question. None of those falsifiers is in hand.
The reputation is earned. It was then asked to cover a shop.
25 Standing constraint
Standing constraint This document describes published research. It is not medical advice. No human use, dose, route or schedule is recommended anywhere in this document. Amounts and durations appear only as reported experimental parameters, always with the population attached. Creatine-deficiency syndromes are clinical genetics, not a sports indication. A rise in serum creatinine is not, by itself, a diagnosis of reduced glomerular filtration, and a healthy-adult safety file is not a licence for use in established kidney disease. Nothing in this article is a recommendation to start, continue, or stop any product.
References
References are compiled from NCBI records. In-text citations are author–year. The numbered list is generated at build from the verified store.
Evidence handling
Every finding is labelled, in the sentence that reports it, by the kind of study that produced it: randomised trial, biopsy study, magnetic-resonance study, meta-analysis, systematic review, society position, case report, open-label pilot, animal study, or in-vitro work. Animal and in-vitro results are never phrased as human outcomes. Society positions and ISSN reviews are maps of a literature; where their authors disclose consulting or industry relationships, that disclosure stays in the body (Section 12). Conflicting results are printed as conflict, not averaged. A newer meta-analysis does not automatically generalise a vegetarian or sleep-loss trial to well-slept omnivores. A case report in a multi-supplement user does not overturn a controlled renal-marker study in healthy adults, and those studies do not license use in established kidney disease. Amounts are experimental parameters. Adjacent outcomes — hypertrophy, mood, estimated GFR, alternative salts — do not inherit the phosphagen performance file. They require their own trials, and they have them. This article grades those files in place.
Continue reading