What's New in Creatine Research: October 2026

This edition covers a quarter instead of a month, so it reaches back across July, August and September. Three papers landed in the same journal within eleven days of each other.

One of them makes our own product look less special than we would like. It is in here anyway.

What was the biggest creatine study this quarter?

A three-level meta-analysis of 63 studies and 1,489 people found that adding creatine to resistance training produced moderate extra gains in strength, a smaller gain in lean mass, and no significant change in fat mass. It published in the Journal of the International Society of Sports Nutrition on 21 September 2026, from a team at Capital University of Physical Education and Sports in Beijing. Read it: PubMed 42765482.

Every one of the 63 trials used creatine monohydrate. Worth holding onto for the third study in this edition.

Those labels are the paper's own thresholds, not ours. It defines small as below 0.20, small to moderate as 0.20 up to 0.50, moderate as 0.50 up to 0.80, and large as 0.80 and above.

Read the last two rows together. Body-fat percentage fell a little while fat mass did not move. That is what happens when the denominator grows: you added lean mass, so fat became a smaller share of a slightly bigger you.

Does the dose or the protocol change the result?

No, not in this analysis: the authors tested intervention duration, weekly frequency, dosing strategy, daily dose and supplementation protocol, and report that intervention-protocol variables did not significantly influence the outcomes.

Those comparisons had real spread in them. Daily dose was split above and below 5g a day, and the supplementation protocols broke down as loading only at 7%, loading plus a long maintenance phase at 54%, and maintenance only at 39%. This is not a case of every trial doing the same thing.

Their discussion offers a reading, and it is worth seeing how firmly they put it. Trials using a maintenance phase tended to report significant upper-body strength improvements while those with only a short loading period did not, and the sentence that follows begins "it stands to reason that" sustained creatine intake through the program is needed to keep muscle creatine elevated. That opener is theirs, and it is doing the same work our hedges do.

The subgroup tables do not settle it either. In the lower body the loading-only group did reach significance, on four effect sizes, and the between-group tests for supplementation protocol were not significant in either limb. So this is a direction the authors find plausible, not a result the data separate.

They are firmer about the dose, writing that "even lower daily doses (3-5 g/day after saturation) are sufficient to elicit strength gains".

Does this change a page we already publish? No, and we went looking. Our page on how much creatine you actually need already says loading is optional, that we are firmly in the steady camp, and that 5g a day lands comfortably in the effective range. We would rather tell you we found nothing to revise than invent a correction so the section has something to announce.

One thing this study cannot tell you is when to take it. Timing was never a moderator. The authors list supplement timing among the variables that were "either unreported or reported inconsistently in many trials", which is why they could not analyze it at all.

Did anything challenge creatine monohydrate this quarter?

Yes, and it deserves to be said plainly: a small trial reported that creatine ethyl ester and creatine hydrochloride produced greater gains in physical performance and peak power than monohydrate over four weeks. Forty collegiate male soccer players, ten per group, four weeks, three sessions a week, published in the same journal on 10 September 2026. Read it: PubMed 42720250.

We sell monohydrate. So read the size of that study honestly, the way you would want us to read a study that flattered us. Ten people per arm is very small. The paper reports p-values for those comparisons without effect sizes in the abstract, so the magnitude of the advantage is not something we can state. Every group improved, including placebo, and all forty were sprint training three times a week throughout.

The forms also came out level on two measures. The paper reports no significant differences among the groups for mean power output or VO2max.

Set that against the first study in this edition: 63 trials, 1,489 people, and creatine monohydrate in every one. One four-week trial with ten people per group does not overturn that, and it is not nothing either. It is a reason to keep reading, which is the position our article on creatine forms should now carry.

What about creatine and the brain?

A six-week trial in 58 healthy, active adults found that most cognitive tests did not move. Participants took guanidinoacetic acid on its own or guanidinoacetic acid with 10g a day of creatine monohydrate, against a placebo. Reaction-time measures showed some significant interaction effects. The Delayed Picture Recognition, Digit Vigilance, Corsi Block and Stroop tests showed what the authors call limited to no effects. Read it: PubMed 42730704.

The authors' own closing line is the one to keep: "Additional research is needed before conclusions can be drawn." Several of their results only approached significance, which they say and we are repeating.

Two things to flag. This trial was not testing creatine by itself, so it cannot tell you what creatine alone does to cognition. And the senior author carries a long disclosure: he reports serving as a paid consultant to the chair of the scientific advisory board of a company that makes both creatine monohydrate and guanidinoacetic acid, states that he has no financial interest in any company that sells dietary supplements, notes a second advisory board seat at a company that sells neither compound, and says the consulting roles are reviewed and approved by his university. The paper states that the study's sponsor does not sell any nutrition or creatine products. We lay it out because you should not have to go looking.

How do you read a study like this?

Start with three things: the sample size, whether a null result was really tested, and who paid. Those three will carry you through most supplement research.

• Find the n before you read the conclusion. Ten people per group and 1,489 people are not the same kind of evidence, even when both papers use the word significant
• Separate "we did not find a difference" from "there is no difference." A study can fail to detect something because it was too small, or because the variable was never recorded well enough to test. Absence of a finding is not a finding
• Read the funding and conflict statements. They are usually at the bottom and usually boring. When they are not boring, they are the most useful paragraph in the paper

An effect size of 0.5, roughly where the strength numbers landed before the bias correction below, is moderate on this paper's own scale. It is a real change you would probably notice over months, not a transformation.

Common questions

Does creatine work without resistance training?

This meta-analysis cannot answer that, because every trial in it compared resistance training plus creatine against resistance training alone. The comparison was never creatine against sitting still. Our shelf has an entry on creatine with and without the workout in 64 adults that looks directly at that question.

Does age change how much creatine helps?

Both the youngest and the oldest groups gained strength significantly, and the one group that did not reach significance was the middle. Upper-body strength was g = 0.62 in adults 18 to 44, g = 0.52 in adults 60 and over, and g = -0.10 in adults 45 to 59, which did not reach significance. That middle estimate rests on only four effect sizes and 108 people, though the contrast between the middle band and both other groups did reach significance. Across all the studies the meta-regression does show effect size falling as average age rises, which is the version the abstract leads with. The authors applied no correction for testing many moderators at once, call the age finding hypothesis-generating, and note it comes from study-level average ages rather than individual participants. For what the research says specifically about older adults, our shelf covers eleven trials in adults over 60.

How much creatine is in a serving of Vybrance?

A serving is four gummies and carries 5g of creatine monohydrate, 120mcg of vitamin B12 as methylcobalamin, and 500mg of taurine. A bag holds 120 gummies, which is 30 servings. You can read more about the formula on the page for Vybrance Labs creatine gummies.

Will creatine make me gain weight?

This analysis found a very small increase in body mass, g = 0.12, and the authors attribute it primarily to gains in lean body mass. Body-fat percentage went slightly down in the same pool. We go through what the scale is actually showing in a longer piece on bloating and weight.

Does the B12 in a creatine supplement give me energy?

Not if you are not deficient. Research does not support the idea that taking B12 raises energy, focus or mood in someone whose levels are already adequate. What B12 does is act as a cofactor in energy metabolism, and it is essential to red blood cell formation. Who actually tends to run low is the more useful question, and we cover it in vegan or vegetarian? you may be running low on creatine and B12.

What are the honest limits of this quarter's research?

The biggest one is that correcting for publication bias moves two of these numbers, and it moves them in opposite directions. The rest follow.

• Publication bias, and we are giving you both corrections rather than the convenient one. The funnel plot and Egger's test flagged asymmetry in the strength results, with smaller null or negative studies likely underrepresented. Statistically filling in the missing studies dropped upper-body strength from g = 0.54 to about g = 0.36, still significant. The authors' own verdict on that is that "although publication bias may be present, its impact on conclusions appears limited". The same method run on fat mass went the other way, turning the null result in our table into g = -0.16, p = 0.007. We are reporting that second one too, and in the same breath reporting why it is the weaker of the two: for fat mass the funnel was symmetric and Egger's test found no bias at all, so nothing prompted that imputation, and the authors say that because the method "relies on symmetry assumptions" and the filled studies are hypothetical, "this result should be interpreted with caution." The correction that cuts against us followed detected bias. The one that flatters us did not
• The big one is a meta-analysis, not a trial. It pools 63 studies and inherits every weakness in them. The authors close on the heterogeneous and incompletely reported nutritional intake and training backgrounds across the included trials
• Its search ended on 21 April 2025. Anything published after that date is not in it, including the other two studies in this edition
• The moderator findings are exploratory. The authors applied no formal adjustment for testing multiple moderators, and describe the age result as hypothesis-generating
• The forms trial is forty men. Collegiate soccer players, ten per group, four weeks. Nothing about women, nothing about anyone who is not a trained athlete, and no effect sizes reported in the abstract
• The cognition trial did not isolate creatine. Its arms were placebo, guanidinoacetic acid alone, and guanidinoacetic acid with creatine. There is no creatine-only arm, so creatine's own contribution cannot be read from it
• The meta-analysis measured bodies, not feelings. Strength, lean mass and body fat are its outcomes. The cognition trial did assess mood, stress and sleep quality and reported scattered effects on them, several of which only approached significance
• All three came out of the same journal issue. Every paper here appears in volume 23, supplement 1 of the Journal of the International Society of Sports Nutrition. Three findings from one issue of one journal are not three independent lines of evidence, and we would rather say that than let the count do work it has not earned
• Funding, stated in full. The meta-analysis reports no conflict of interest and was supported by a grant from the authors' own institution, the Science and Technology Strengthening Project of Capital University of Physical Education and Sports. The cognition trial carries the long disclosure described above. The forms trial we have read only at the abstract level

What are we watching next?

Two threads: whether the forms comparison gets run again at a usable size, and whether the age signal survives a study built to test it. At ten per arm the forms question is open rather than answered, and nothing in that trial speaks to women. The age pattern came out of a subgroup with four effect sizes in the middle band, which is exactly the kind of finding that evaporates or sharpens when somebody designs around it.

Everything we verify entry by entry goes on the studies shelf, including the ones that find nothing.

See you in the wild,

-Lawrence

These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. Nothing on this page is medical advice. Talk to your doctor before starting any supplement. References: PMID 42765482, PMID 42720250, PMID 42730704.