§ Claim under review · Health
"In a pilot trial of 20 people with Alzheimer's disease (average age 73), those who took 20 g/day of creatine monohydrate for 8 weeks showed improvements in working memory, fluid cognition, reading ability, and attention" Post image text: "Notable / People with Alzheimer's who took 20 g/day of creatine for 8 weeks showed improvements in working memory, fluid cognition, reading ability, and attention"
Verdict
Mostly accurate
Confidence
HighSummary
This claim describes a real study. Researchers at the University of Kansas gave 20 people with Alzheimer's disease, average age 73, twenty grams of creatine monohydrate daily for eight weeks, and every number quoted in the post matches the published paper exactly, including the 11 percent rise in brain creatine and the specific cognitive scores. The important limitation is that there was no placebo group and no comparison group, so the study cannot show that creatine caused the score improvements. The authors themselves wrote that they cannot rule out practice effects or placebo effects, and they framed the results as preliminary evidence for future trials. Two details are softened in the post: the attention result was borderline and described by the authors as only a trend, and four of the seven cognitive subtests showed no improvement at all. The largest gain was on a reading recognition test that is specifically designed to stay stable over time, which is a pattern often seen with repeat testing. The bottom line is that the study is genuine and the numbers are reported honestly, but it was a small feasibility trial, not proof that creatine improves cognition in Alzheimer's.
The readings
key figures from the evidenceaverage increase in brain total creatine concentration
significance of flanker attention/inhibitory control test improvement
Why this verdict
Evidence
The study is real and the numbers in the post match the published paper almost exactly. In this single-arm pilot trial, researchers investigated the feasibility of 20 g/day creatine monohydrate for 8 weeks in 20 patients with Alzheimer's disease, measuring compliance, serum creatine, brain total creatine, and cognition using the NIH Toolbox and the Mini-Mental State Examination.
Nineteen participants achieved the target of 80% or greater compliance, serum creatine was elevated at 4 and 8 weeks, brain total creatine increased by 11%, and cognition improved on global (p = .02) and fluid (p = .004) composites, List Sorting (p = .001), Oral Reading (p < .001), and Flanker (p = .05) tests.
The exact figures cited in the post appear in the paper: brain total creatine concentrations increased from 330.5 ± 36.8 IU to 366.9 ± 57.5 IU (p < .001, Cohen's dz = 1.01) , and brain total creatine concentration increased in 85% of participants, with an overall average increase of 11% . Total cognition improved from 75.3 ± 13.9 to 78.6 ± 13.5 (p = .02), fluid cognition from 59.1 ± 14.7 to 63.5 ± 15.8 (p = .004), list sorting working memory from 66.2 ± 18.7 to 74.2 ± 20.3 (p = .001), and oral reading recognition from 98.0 ± 7.18 to 103 ± 7.23 (p < .001); there was a trending improvement in the flanker inhibitory control and attention test (68.1 ± 15.4 to 73.0 ± 13.9, p = .05), and no change in crystallized cognition or in the other individual cognitive tests.
Participant details also check out: twenty participants who met McKhann criteria for dementia due to AD (age 73.1 ± 6.3 years) were allocated to the 8-week intervention, all 20 completed the study, study partners reported 13 adverse events, all mild, and nineteen participants reported consuming 80% or more of expected doses .
Findings
✓ What's accurate 6
- The study exists, is peer-reviewed, and PMID 40395689 is correct.
- 20 participants, mean age 73.1 years, 20 g/day creatine monohydrate, 8 weeks: all accurate.
- All 20 completed; 19 met the 80% compliance threshold: accurate.
- Brain creatine rose 11% on average, from 330.5 to 366.9 IU, in 85% of participants: accurate.
- Every cognitive score cited in the caption (75.3 to 78.6, 59.1 to 63.5, 66.2 to 74.2, 98 to 103, 68.1 to 73.0) matches the published table exactly.
- The caption itself correctly flags the single-arm, no-placebo, preliminary nature and states causality is not established.
≈ What's misleading 5
- Temporal/certainty overreach in the headline image. The standalone banner text, "People with Alzheimer's who took 20 g/day of creatine for 8 weeks showed improvements in working memory, fluid cognition, reading ability, and attention," reads as a demonstrated drug effect. Without a control group these are within-person score changes over time, not a measured effect of creatine. The caption corrects this, but the image is the part most likely to be screenshotted and shared alone.
- Overstatement of the attention finding. The paper calls the flanker attention result a "trending improvement" at p = .05 with a small effect size (Cohen's dz = 0.33) , yet the claim lists attention alongside the clearly significant results without qualification.
- Selective reporting of the cognitive battery. Four of seven NIH Toolbox subtests did not improve, and one (Picture Vocabulary) trended downward. The crystallized composite showed no change. Neither the banner nor the caption mentions the null results, which creates an impression of broad cognitive improvement.
- "Reading ability" is the weakest evidence of a real drug effect, not the strongest. Oral Reading Recognition had the largest effect size, but it is a crystallized "hold" test. Crystallized tests such as Picture Vocabulary and Oral Reading Recognition measure verbal abilities that are less likely to be affected by brain injury compared to fluid measures, and the crystallized composite can serve as an estimate of premorbid ability. A large short-term gain on a test designed to be stable is more consistent with practice or test-retest effects than with a bioenergetic treatment effect. The authors themselves flag this class of explanation.
- Minor imprecision in the journal name. The paper is in Alzheimer's & Dementia: Translational Research & Clinical Interventions, a companion journal, not the flagship Alzheimer's & Dementia. This does not change the substance.
? What's uncertain 5
- Whether creatine caused any of the cognitive changes. This cannot be determined from a single-arm design and the authors do not claim it.
- The MMSE result. The abstract confirms MMSE was measured at baseline and 8 weeks, but I did not retrieve the specific MMSE values or p value. The paper states there was no change in the other individual cognitive tests.
- Funding sources, supplement sourcing, and author conflict-of-interest disclosures were not retrieved.
- Durability beyond 8 weeks, and whether any effect exists at more common doses such as 5 g/day, are not addressed by this trial.
- Whether a larger randomized controlled trial has since been completed was not established.
Sources
4 of 5 linked to recordsSmith AN, Choi IY, Lee P, et al. "Creatine monohydrate pilot in Alzheimer's: Feasibility, brain creatine, and cognition." Alzheimer's & Dementia: Translational Research & Clinical Interventions, 2025;11(2):e70101. PMID 40395689
ClinicalTrials.gov NCT05383833, "Creatine to Augment Bioenergetics in Alzheimer's (CABA)," PI Matthew Taylor, University of Kansas Medical Center
Taylor MK, et al. Published protocol, Pilot and Feasibility Studies 2024;10(1):42
Conference abstract, Alzheimer's & Dementia (Taylor 2025), reporting mean age 73.1 ± 6.3 years
Smith AN, et al. Muscle strength/size data from the same trial, Front Nutr 2025;12:1670641