What's New in Creatine Research: September 2026

August was a good month for research on people who are not twenty-two year old athletes. The headline trial this time ran for twelve weeks in adults aged 45 to 65, most of them women, and it asked a question the creatine literature usually skips: what happens if you take it and do not otherwise change a thing?

It also used a dose twice the size of the one we normally talk about, which is the part that complicates something we published ourselves. That is in here too.

What was the biggest creatine study this month?

A 12-week randomized, double-blind, placebo-controlled trial from the Exercise and Sport Nutrition Lab at Texas A&M University, published in the Journal of the International Society of Sports Nutrition in August 2026, testing 10 grams of creatine monohydrate a day in 64 sedentary adults aged 45 to 65. Seventy-three people started, 64 completed and were analyzed, and 40 of those 64 were women.

Participants chose whether to join a non-exercise track or an exercise and weight-loss diet track. Inside whichever track they picked, they were then randomized, double blind, to creatine monohydrate or a maltodextrin placebo. Body composition was measured by DXA at weeks 0, 6 and 12, alongside fasting bloodwork and a battery of cognitive tests. The trial is registered as ISRCTN83081058, and you can read the record at PubMed 42578920.

What did the trial actually find?

Both tracks reported gains in lean tissue mass and strength, and the abstract describes those results as directions rather than numbers. There are no effect sizes, confidence intervals or p-values printed in the record we can read, so we are not going to hand you any.

The authors' own conclusion keeps its hedge on the surface: creatine "may help middle-aged and older adults maintain muscle mass and strength while favourably affecting selected markers of health and cognitive function." May help. Maintain. Selected markers.

What does this study not show?

It does not show that creatine beats exercise, and it cannot compare its two tracks against each other, because participants chose their own track rather than being randomized into one. Here is the honest list.

• Self-selected tracks. The randomization was creatine versus placebo. Who exercised was a personal choice, so the exercise and non-exercise groups are not a randomized comparison and should not be read as one
• No numbers in the abstract. Every finding is a direction. Without effect sizes we cannot tell you whether "increased lean tissue mass" means half a kilogram or two
• The dose was 10 grams a day. Two servings of 5 grams, twice what most people take and twice what most trials use
• "Selected" markers. Some blood and cognitive measures moved. The abstract does not say which, or how many stayed flat
• The authors call their own cognitive results exploratory. In their words, "Individual cognitive and biomarker findings should be interpreted as exploratory"
• Nine of 73 are missing from the analysis, with no stated reason
• Twelve weeks. Enough to move body composition, not enough to say anything about a year
• Conflicts are real and disclosed. The senior author chairs the scientific advisory board of a company that manufactures creatine monohydrate and co-founded the nonprofit that publishes this journal. That is in the paper, and it belongs in any summary of it

Does this change what we have told you about the 5 gram dose?

It does not overturn it, and it does add a caveat we owe you: the newest trial in our own age group got its results at 10 grams a day, and it cannot tell us whether 5 grams would have done the same.

Our own guide to the 5 gram question says 5 grams a day is the research standard, and that is still true. It is the dose used across most of the literature, and it is the dose in the pooled analyses. But we wrote that guide to answer "how much do I need," and a fair reading of this trial is that we do not know whether the results reported here scale down.

The revised position, stated plainly: 5 grams remains the standard and the sensible default. This trial is not evidence that 5 grams produces what it reported, because it did not test 5 grams. If a future trial runs 5 against 10 head to head in this age group, we will post that here whichever way it lands.

It is worth reading alongside the shelf, too. The meta analysis of 608 women across 7 trials shares this study's senior author and reached a more cautious conclusion, and a two year trial in 237 women found no change in bone density at all. Three papers, one compound, three different tempers.

How do you read a study like this one?

Start with who was randomized to what, because that single fact sets the ceiling on everything a trial can claim. Four habits that travel well:

• Find the randomization, not the conclusion. In this trial people were randomized to creatine or placebo but chose their own exercise track. That means creatine versus placebo is a fair comparison and exercise versus no exercise is not
• Look for the dose before the result. A finding at 10 grams a day is a finding about 10 grams a day. Doses do not quietly transfer
• Treat a direction without a number as a direction without a number. "Improved" with no effect size cannot be compared to anything, including last month's study
• Read the conflict of interest statement. It is usually at the bottom, it is usually short, and it is usually the fastest thing you can do to calibrate what you just read

Common questions

Should I take 10 grams of creatine a day now?

Nothing in this trial establishes that you should. It reported results at 10 grams a day without testing 5, so it tells you what happened at the higher dose and nothing about whether the lower one would match it. Five grams a day is what the broader literature uses and what most people take, and a change to your routine is worth raising with your own clinician rather than with a study summary.

Does creatine do anything if I am not exercising?

This trial is one of the few that even asked. Its non-exercise track reported increases in lean tissue mass, strength and muscular endurance, plus favorable changes in some blood and cognitive markers. Hold that loosely: those participants chose not to exercise rather than being assigned to it, the abstract gives no effect sizes, and it ran twelve weeks at a 10 gram dose.

Is creatine research mostly done on young men?

Historically most of it was, which is exactly why this one is worth the attention. Sixty-four adults with a mean age of 54.5, forty of them women, is much closer to the person reading this than the college-athlete samples that fill the older literature. Our complete guide to creatine for women by life stage covers what is and is not established for this group.

What form of creatine was used?

Creatine monohydrate, given as two 5 gram doses a day against a maltodextrin placebo. Monohydrate is the form with the deepest research base behind it, and it is the form in Vybrance Labs creatine gummies, where each four gummy serving carries 5g of creatine monohydrate alongside 120mcg of vitamin B12 and 500mg of taurine.

What we are watching next

Two entries went onto the research shelf alongside this one, and neither is a clean win for the ingredient it covers. Twenty elite wrestlers took a single dose of taurine at roughly the amount in a normal serving: peak power rose, and average power did not move at all. And 376 adults on long-term acid reflux medication saw their median B12 fall, on numbers sitting right at the edge of significance.

Ahead of us: a large multi-year creatine and ageing trial has published its protocol but no results yet, and a triple-blinded trial of taurine on metabolic health in healthcare workers is registered and running. Both go up here when they report, whichever way they land. The full shelf, with the citation and the limits attached to every entry, lives at Studies.

This article is general education and not medical advice. 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. Reference: PMID 42578920, DOI 10.1080/15502783.2026.2716273.

See you in the wild,

-Lawrence