What the study is and is not

This is a pilot feasibility trial — not a treatment trial. The 11% brain creatine increase is the real finding. The cognitive improvements are real but uncontrolled.

The CABA trial (Smith et al., 2026, Alzheimer's & Dementia: TRCI) is the first human study to test creatine supplementation in Alzheimer's patients. Its primary goal was to check whether 20g/day was safe and achievable over eight weeks — not to prove creatine treats Alzheimer's. It did both things it set out to do. The brain creatine increase is the most important finding — it confirms the supplement actually reaches the brain in this population. The memory improvements are encouraging but cannot be disentangled from placebo effect without a control group. Both things are true simultaneously.

Why this trial matters — and why it's been misread

When this study landed on May 19, 2026 in Alzheimer's & Dementia: Translational Research & Clinical Interventions, the supplement world processed it as a creatine cure for Alzheimer's. That is not what it said, and the researchers were careful about this. But the gap between what a pilot feasibility study demonstrates and what a headline can carry in a tweet is wide enough to drive a supplement marketing campaign through.

So let's start at the top.

Alzheimer's disease is not simply a memory disease. It is, among other things, an energy metabolism disease. Brain energy production is impaired in Alzheimer's patients at multiple levels — mitochondrial dysfunction, reduced glucose uptake, and a specifically disrupted creatine-phosphocreatine system.1 The brain's fast ATP buffer, which neurons rely on during high-demand firing, is running low in Alzheimer's patients even before severe cognitive symptoms appear. This is not a new observation. The question was always whether you could do something about it.

What nobody had checked, until this year, was whether oral creatine supplementation could actually raise brain creatine in Alzheimer's patients. Supplementation raises brain creatine in healthy people — that was established by Dechent et al. in 1999 using proton MRS imaging. But whether the same transport mechanism works in a brain affected by Alzheimer's pathology was an open question. The SLC6A8 creatine transporter could behave differently. Amyloid plaques might interfere. Nobody knew.

That's the question the CABA trial asked. And the answer came back: yes, it still works.

The trial design

CABA — Creatine to Augment Bioenergetics in Alzheimer's Smith et al., 2026 · University of Kansas Medical Center
Design
Single-arm, open-label pilot
No placebo group. No blinding. Participants knew they were taking creatine.
Participants
20 adults, aged 60–90
Probable Alzheimer's diagnosis based on clinical + laboratory criteria. Most on standard medications (memantine or donepezil).
Dose
20g / day
10g creatine monohydrate taken twice daily × 8 weeks. Four times the standard athletic dose.
Primary outcome
Safety + feasibility
Could patients take this dose safely for 8 weeks? Could they comply with the regimen?
Brain imaging
¹H-MRS at week 0 and week 8
Proton magnetic resonance spectroscopy — direct measurement of brain creatine levels in vivo.
Cognitive testing
NIH Toolbox Cognition Battery
Administered at baseline and week 8. Includes working memory, attention, fluid intelligence, reading.

One thing worth flagging: the creatine monohydrate was donated by Life Extension Inc. The company did not design the trial, and the researchers at the University of Kansas conducted the study independently. But it is a disclosure worth knowing. Funded in-kind by a supplement company, no matter how arm's-length, is a different situation from fully independent funding.

What the study found

11%
Average increase in brain creatine (¹H-MRS)
85%
Of participants showed measurable brain creatine increase
~10%
Improvement in working memory scores
20
Participants enrolled; 19 completed the 8-week protocol
1
Brain creatine went up — and this is the important finding

Average brain creatine measured by proton MRS increased 11% over eight weeks. In 85% of participants, the increase was measurable. This is not a cognitive outcome — it is a pharmacokinetic result. It tells you that creatine supplementation successfully raises brain creatine in Alzheimer's patients, which was not previously established. Without this result, nothing else in the trial would mean much. With it, the hypothesis that oral creatine can affect brain bioenergetics in Alzheimer's is confirmed — at least in terms of target engagement.

2
Working memory improved by about 10%

On the NIH Toolbox working memory test — described by lead researcher Matthew Taylor as the type of memory you use in a card-matching game, holding and manipulating information moment to moment — scores rose roughly 10% from baseline to week 8. This is a meaningful-sized improvement for this population. It is also completely uncontrolled.

3
Fluid intelligence improved; crystallised intelligence held steady

Fluid intelligence scores (reasoning and problem-solving independent of prior knowledge) went up. Crystallised intelligence (accumulated knowledge) did not decline, which is notable in a progressive disease. Reading and attention also showed improvements. The pattern — task-specific memory gains without decline in broader knowledge — is at least consistent with what you would predict from better ATP availability in task-demanding neural circuits.

4
Brain creatine increase correlated with reading improvement

The researchers found a statistically significant correlation between the magnitude of the brain creatine increase and reading performance gains. This dose-response relationship — more creatine in the brain, more reading improvement — is the most mechanistically interesting secondary finding. It is not proof of causation, but it is the kind of signal that makes a placebo explanation harder to sustain for the reading endpoint specifically.

5
20g/day was safe and tolerable

No serious adverse events. Mild gastrointestinal complaints in a few participants early on — which is expected at this dose and typically resolves. All 19 completers got through eight weeks without discontinuing for safety reasons. For an elderly population (60–90 years old) already on Alzheimer's medications, this is genuinely useful information for trial design going forward.

"Seeing it change in Alzheimer's patients was really exciting. That 11% is a significant increase."

Matthew Taylor, Ph.D. — Lead researcher, University of Kansas Medical Center

The limitations — and why they matter a lot here

This is where it gets harder to write honestly without sounding like you are dismissing something genuinely promising. The CABA trial is important preliminary work. It is also a study with structural weaknesses that make its cognitive findings essentially uninterpretable as evidence that creatine improves Alzheimer's cognition. Both are true.

Fatal flaw #1

No control group

Every participant received creatine. There was no placebo arm. Without a control group, you cannot know how much of the cognitive improvement was due to the intervention versus natural variation, practice effects on the cognitive tests, increased attention from study participation, or placebo response. Alzheimer's trials are particularly vulnerable to placebo effects in subjective and semi-objective measures like cognitive batteries.

Fatal flaw #2

No blinding

Participants knew they were taking creatine. Assessors knew participants were taking creatine. Open-label design inflates cognitive test performance through expectation alone — a well-documented effect in neurological trials. The test administrators may unconsciously provide different support or encouragement to participants they know are being treated.

Sample size

20 participants

This is a pilot. It was designed to be small. But n=20 means any result is highly sensitive to individual outliers, and the statistical power is too low to detect modest effects reliably or to estimate effect size with any precision. The cognitive improvements could be driven by two or three individuals responding well.

Duration

Eight weeks only

Alzheimer's is a progressive disease measured in years. Eight weeks is a very short window to detect changes against a background of gradual decline — and too short to assess whether any improvement is sustained. The real clinical question is whether brain creatine elevation translates into slower progression over 12–24 months.

Practice effects

Test-retest inflation

The NIH Toolbox Cognition Battery was administered twice — baseline and week 8. Cognitive tests show practice effects: people perform better the second time simply from familiarity with the format. Without a placebo group taking the same tests twice, it is impossible to separate genuine cognitive improvement from test familiarity in this cohort.

Industry funding

Creatine donated by Life Extension Inc.

The researchers appear to have conducted the study independently, and Life Extension Inc. does not appear to have had input into design or analysis. But in-kind supplement donation from a commercial supplier is a conflict worth disclosing prominently, and it adds reason to await replication by fully independent groups.

What the researchers themselves said

The CABA paper describes itself explicitly as a feasibility and preliminary evidence trial. The authors call for a randomised, placebo-controlled trial as the necessary next step. The Brain Docs (neurologists Drs. Ayesha and Dean Sherzai) noted publicly that the study "wasn't designed to test whether creatine treats Alzheimer's, slows disease progression, or meaningfully improves quality of life." The researchers agree with the critics. The social media version of this study is not the study.

Why creatine might actually help in Alzheimer's — the biological case

The honest answer is that the mechanism is plausible but not proven in humans. Here is what the preclinical evidence established before this trial.

Alzheimer's disease pathology includes amyloid-beta plaques and tau tangles, which are the most visible biomarkers. But running underneath both is a metabolic crisis. Glucose uptake in Alzheimer's brains drops significantly — PET imaging shows hypometabolism in the temporal and parietal lobes years before symptoms appear. As mitochondrial function degrades, neurons become increasingly reliant on the phosphocreatine-creatine kinase buffer system to handle ATP demand spikes. And creatine kinase activity is specifically impaired in Alzheimer's brain tissue, measured post-mortem.2

In mouse models of Alzheimer's (the 3xTg and 5xFAD lines), creatine supplementation improved spatial memory, reduced amyloid-beta accumulation, and decreased phosphorylated tau levels.3 There is also evidence that creatine protects neurons against glutamate toxicity — relevant because excitotoxicity from excess glutamate is part of the neuronal death cascade in Alzheimer's. Rat hippocampal neurons exposed to beta-amyloid showed less death when pre-treated with creatine.4

None of this translates automatically to humans. Mouse models of Alzheimer's are notoriously poor predictors of human trial outcomes — dozens of interventions that cured Alzheimer's in mice have failed in humans. But the mechanism is specific enough, and the new human trial data on brain creatine penetration is solid enough, that the hypothesis is now worth testing properly.

Why 20g instead of 5g?

The standard creatine dose for athletic performance is 3–5g/day. The CABA trial used 20g/day — four times that amount. This is worth explaining because some of the coverage treated it as a generic protocol that anyone could follow.

The rationale was mechanistic. Brain creatine uptake via the SLC6A8 transporter is slow and limited compared to muscle uptake. Previous healthy-adult studies (Dechent et al., 1999; Lyoo et al., 2003) showed that brain creatine increases of 5–15% required sustained supplementation at higher doses than are needed for muscle effects. In an Alzheimer's population where the transporter might be further compromised, the researchers chose a higher dose to maximise the probability of detecting any brain creatine increase at all.

The 20g/day is also divided — 10g twice daily with food, not a single bolus. This improves absorption and reduces GI side effects. At this dose, creatine will cause water retention and some GI discomfort in the first week. That is expected and transient. It is not appropriate for people with kidney disease, and anyone on existing medications should check for interactions before attempting anything like this dose independently.

This dose is not a recommendation for the general public

20g/day was chosen for a specific scientific reason in a monitored clinical trial with weekly check-ins and blood draws. The 5g/day dose is appropriate for general cognitive use in healthy individuals — particularly vegetarians, or anyone on insufficient sleep. Do not take "20g worked in an Alzheimer's trial" as a reason to quadruple your creatine intake. The trial dose was set to maximise brain uptake measurement sensitivity, not to optimise outcomes.

The India angle — why this study is worth watching here

India context

India has an estimated 8.8 million people living with dementia, of whom Alzheimer's disease accounts for the majority. That number is projected to cross 17 million by 2050 as the population ages. Access to the existing pharmacological treatments — memantine, donepezil — is improving but remains uneven outside major cities.

The creatine story matters in India for a second reason too: widespread vegetarian dietary patterns mean large fractions of the Indian population are already at lower baseline brain creatine than omnivores. The cognitive protection hypothesis for creatine is strongest precisely in people who eat no animal protein — which describes a substantial share of the at-risk older population in states like Gujarat, Rajasthan, and Karnataka.

The CABA trial does not prove anything about prevention. But if a larger, properly controlled trial confirms even modest cognitive slowing in Alzheimer's patients, the low cost, high safety profile, and wide availability of creatine monohydrate in India make it potentially relevant at a population scale that expensive disease-modifying biologics cannot reach.

None of that is settled yet. It is worth watching.

What happens next

The CABA trial was designed to answer: "Is it worth running a real trial?" The answer is yes.

The University of Kansas team has stated they intend to proceed to a randomised, placebo-controlled, double-blind trial based on these results. That trial will need several hundred participants, a placebo arm receiving an inert powder, blind cognitive assessors, a longer duration (likely 12–24 months), and biomarker endpoints beyond just cognitive scores — amyloid PET, tau, and brain volume measures are the realistic targets.

The cognitive improvements in the CABA pilot may or may not replicate when controlled. The brain creatine increase almost certainly will — it is a pharmacokinetic result with no plausible placebo mechanism, and it is consistent with what was seen in healthy adults. That result is robust. The cognitive signal is the question the next trial is built to answer.

Timeline: a properly powered phase II RCT in this population takes three to five years from funding to publication. The earliest we should expect controlled human evidence on creatine and Alzheimer's cognition is 2028–2030, realistically.

What to make of it right now

The honest answer to "should someone with Alzheimer's take creatine?" is: we do not know yet. This trial showed the supplement reaches the brain in this population and that the dose is tolerable. It did not show it slows progression, improves daily function, or changes disease trajectory. Those things require a controlled trial that does not exist yet.

What the trial does not support: taking 20g/day of creatine without medical supervision based on this study. The participants in CABA had regular check-ins, blood draws at weeks four and eight, and were monitored throughout. The dose that was tested in a clinical setting with that level of oversight is not the same as something someone should start at home based on a headline.

What is reasonable: if you are a vegetarian or vegan, and you are in the population that this research will eventually apply to — or you have a family member who is — 5g/day of plain creatine monohydrate is safe, cheap, and has existing cognitive evidence in vegetarians and sleep-deprived individuals. The Alzheimer's-specific evidence is preliminary. The general cognitive evidence for creatine in depleted populations is not.

The 11% brain creatine increase is genuinely exciting. Not because it cures anything. Because it shows the door is open. Whether walking through that door actually helps is what we are still waiting to find out.

References

1
Smith AN, Choi IY, Lee P, et al. (2026). Creatine monohydrate pilot in Alzheimer's: Feasibility, brain creatine, and cognition. Alzheimer's & Dementia: Translational Research & Clinical Interventions, 11(2):e70101. doi:10.1002/trc2.70101 — The CABA pilot trial. Primary source for all trial data in this article. Published May 19, 2026. Pilot RCT (single-arm)
2
Aksenov MY, Aksenova MV, Butterfield DA, Geddes JW, Markesbery WR. (2001). Protein oxidation in the brain in Alzheimer's disease. Neuroscience, 103(2):373–383. — Creatine kinase activity impairment in Alzheimer's brain tissue. doi:10.1016/S0306-4522(00)00580-7 Post-mortem brain study
3
Zhu S, Li M, Figueroa BE, et al. (2004). Prophylactic creatine administration mediates neuroprotection in cerebral ischemia in mice. Journal of Neuroscience, 24(26):5909–5912. Also: Cooke MB et al. (2020). Sex-specific effects of chronic creatine supplementation on hippocampal-mediated spatial cognition in the 3xTg mouse model of Alzheimer's disease. Nutrients, 12(11):3589. doi:10.3390/nu12113589 Animal model
4
Brewer GJ, Wallimann TW. (2000). Protective effect of the energy precursor creatine against toxicity of glutamate and beta-amyloid in rat hippocampal neurons. Journal of Neurochemistry, 74(5):1968–1978. doi:10.1046/j.1471-4159.2000.0741968.x — Creatine neuroprotection against amyloid-beta in rat neurons. In vitro / ex vivo
5
Dechent P, Pouwels PJ, Wilken B, Hanefeld F, Frahm J. (1999). Increase of total creatine in human brain after oral supplementation of creatine-monohydrate. American Journal of Physiology, 277(3):R698–R704. doi:10.1152/ajpregu.1999.277.3.R698 — Original ¹H-MRS demonstration of brain creatine increase in healthy adults. Controlled study
6
Forbes SC, Cordingley DM, Cornish SM, et al. (2022). Effects of creatine supplementation on brain function and health. Nutrients, 14(5):921. doi:10.3390/nu14050921 — Comprehensive review of creatine and brain health across populations. Systematic review
7
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 — PCr-ATP buffering and CK system in disease context. Review
8
Taylor MK, Burns JM, Choi IY, et al. (2024). Protocol for a single-arm, pilot trial of creatine monohydrate supplementation in patients with Alzheimer's disease. Pilot and Feasibility Studies, 10(1):42. doi:10.1186/s40814-024-01469-5 — Pre-registration protocol paper for the CABA trial. Protocol

Disclosures: Naked Compound participates in the Amazon.in affiliate programme. Some links earn a small commission. No manufacturer provided samples or funding for this content. The CABA trial creatine was donated by Life Extension Inc. — that relationship is disclosed in the primary paper and noted above. Full conflicts policy: conflicts-policy