Exogenous Ketones and Cognition: What the 2026 Meta-Analysis of 29 Trials Found
A systematic review and meta-analysis in Frontiers in Nutrition pools 38 trials of ketone esters, MCT, and ketone salts against placebo. Here is what it found on cognitive performance, dose-response, and where the evidence is still thin.
Research context: This article discusses a peer-reviewed systematic review and meta-analysis and the primary trials it draws on. It is intended for informational and educational purposes only and does not constitute medical advice. Consult a qualified healthcare professional before making changes to diet, lifestyle, or supplementation.
A scattered literature gets its first large pooled answer
Exogenous ketone products, ketone esters, ketone salts, and medium-chain triglyceride (MCT) oils, have circulated in endurance sport and nootropic circles for years on the strength of a straightforward mechanistic story: the brain can run on ketone bodies as well as glucose, so raising blood ketones with a drink or a capsule should give neurons an alternative fuel source when glucose metabolism is impaired or under strain. The trial evidence behind that story has been scattered across small studies in wildly different populations, from healthy young adults doing working-memory tasks to people with Alzheimer's disease taking MCT oil for months. No one had pooled it.
A systematic review and meta-analysis published in Frontiers in Nutrition in April 2026 by Bruno Bonnechère, Elizabeth B. Stephens, Amy C. Boileau, Martin Ducker, and Brianna J. Stubbs, drawing on researchers affiliated with Hasselt University and the Buck Institute for Research on Aging, changes that. The review identified 38 randomised controlled trials (41 distinct protocols) covering 1,602 participants, and pooled 29 of those protocols from 26 studies, 1,117 participants, into a formal meta-analysis. It is registered on PROSPERO (CRD42023471727) and is the largest quantitative synthesis of exogenous ketone and cognition trials published to date.
What went into the pool
The review included randomised, placebo-controlled trials of exogenous ketone products across four population groups: healthy adults, people with mild cognitive impairment (MCI), people with Alzheimer's disease, and people with other neurological conditions including Parkinson's disease, multiple sclerosis, and traumatic brain injury. Trial duration ranged from a single acute dose to six months, with most longer protocols running 1.5 to 3 months. Of the 1,602 participants enrolled across the full systematic review, 1,538 (96 percent) completed their assigned protocol.
The interventions were not a single product category. MCT oil was the most heavily represented, 1,054 participants across 19 studies, typically caprylic acid (C8:0) and capric acid (C10:0) at doses of 6 to 56 grams per day. Ketone esters, mainly (R)-3-hydroxybutyl-(R)-3-hydroxybutyrate, appeared in 15 studies covering 300 participants, at doses ranging from 573 mg/kg to a total of 162.5 grams. Ketone salts (beta-hydroxybutyrate salts), mostly tested in athletic-performance contexts, contributed a much smaller 44 participants across 2 studies. Two branded caprylic-triglyceride formulations (AC-1202 and AC-1204) added 565 participants across 2 US trials, and a single study of coconut oil in an Alzheimer's population contributed 44 more. Cognitive outcomes spanned standard instruments including the Mini-Mental State Examination, the ADAS-Cog, the Trail Making Test, the Stroop task, the Wechsler Memory Scale-Revised, digit span, an N-back working-memory task, the Boston Naming Test, verbal fluency, and processing-speed measures.
The pooled result
Across the 29 pooled protocols, exogenous ketone supplementation produced a statistically significant improvement in cognitive performance relative to placebo: standardised mean difference (SMD) 0.29, 95 percent confidence interval 0.16 to 0.41, p < 0.001. That is a small-to-moderate effect by conventional benchmarks, consistent with the authors' own description of the result as "modest." Heterogeneity across the pooled studies was high (I-squared = 91 percent), which the authors flag directly as a limit on how far the pooled estimate can be generalised. A funnel plot and Egger's test found no evidence of publication-bias asymmetry (Egger's intercept 2.65, standard error 0.89, p = 0.08), and a sensitivity analysis did not identify outlier studies distorting the pooled figure.
Subgroup comparisons. None of the subgroup differences reached statistical significance on their own interaction tests, which means the review cannot say with confidence that one supplement type, population, or dosing window clearly outperforms another. But the point estimates are still informative. Immediate, single-dose protocols showed a larger effect (SMD 0.39, 95% CI 0.16 to 0.62) than longer-term protocols lasting more than 13 days (SMD 0.19, 95% CI 0.12 to 0.26), interaction p = 0.11. Ketone esters showed a larger point estimate (SMD 0.37, 95% CI 0.14 to 0.60) than MCT (SMD 0.15, 95% CI 0.06 to 0.25), interaction p = 0.083. Healthy participants showed a larger effect (SMD 0.36, 95% CI 0.18 to 0.54) than people with MCI or Alzheimer's disease (SMD 0.15, 95% CI 0.01 to 0.30), interaction p = 0.077. Testing under an acute cognitive stressor such as exercise made little difference either way (interaction p = 0.89).
Dose-response. This is the part of the paper that adds the most beyond a simple pooled effect. Meta-regression found a statistically significant positive association between daily ketone dose and cognitive improvement, both overall (beta 0.0020, SE 0.0005, p < 0.001) and within the healthy-adult subgroup specifically (beta 0.0022, SE 0.0005, p < 0.001). The association was steeper again within the Alzheimer's disease subgroup (beta 0.0231, SE 0.0032, p < 0.001). Total cumulative dose across the whole intervention period and overall intervention duration were not significantly associated with outcome (p = 0.42 and p = 0.36 respectively), which points toward daily dosing level, not simply how long someone stays on a product, as the variable that tracks with cognitive benefit in this dataset.
Safety. The authors report that exogenous ketones "have no currently known safety concerns" as a product category. Gastrointestinal symptoms occurred in some trials, most often with MCT and at higher doses, and appeared to be reduced by gradual dose escalation rather than starting at a full target dose immediately.
Why the effect looked different by population
The subgroup split between healthy adults and people with MCI or Alzheimer's disease is worth sitting with rather than collapsing into a single number. In the healthy-adult trials, the clearest signals came from immediate, single-dose ketone ester protocols tested around working memory and processing speed, sometimes alongside an exercise or metabolic-stress challenge, with mixed results specifically during acute exertion. In the MCI and Alzheimer's trials, the pattern looked more like a slower-building, longer-duration MCT or ketone-ester effect on standard clinical instruments: improvements in MMSE scores, episodic memory, language, and executive-function measures over weeks to months, with some suggestion in the underlying literature that the benefit is more pronounced in people who do not carry the APOE4 allele, a genetic variant associated with impaired brain glucose metabolism in Alzheimer's disease.
That distinction lines up with the underlying rationale for testing ketones in neurodegenerative disease in the first place. Alzheimer's disease is associated with reduced glucose uptake in the brain years before clinical symptoms appear, a phenomenon covered in more depth in our review of metabolic dysfunction and dementia risk. Ketone bodies use a different set of transporters to cross the blood-brain barrier than glucose does, which is the mechanistic argument for why raising blood ketones might partially bypass an impaired glucose-uptake bottleneck rather than simply adding calories. The pooled dose-response finding, a steeper effect per unit dose specifically in the Alzheimer's subgroup, is consistent with that "filling a specific metabolic gap" model, though the authors are careful to describe it as an association from a highly heterogeneous dataset, not a demonstrated mechanism.
How this connects to the broader mitochondrial-cognition picture
Ketone bodies do not act on the brain in isolation from cellular energy metabolism generally. Once acetoacetate and beta-hydroxybutyrate reach neurons, they are metabolised inside the mitochondria to generate ATP, feeding into the same broader system covered in our review of mitochondria and cognitive performance. Some of the proposed benefit of ketone metabolism over glucose metabolism is that it can generate ATP with somewhat less oxidative byproduct per unit of energy produced, though this meta-analysis was not designed to test that mechanism directly and reports only the downstream cognitive outcomes, not biochemical intermediates. Blood ketone concentrations, notably, were not measured in all of the included trials, which the authors list as a specific limitation: it means some studies could not confirm that their intervention actually raised circulating ketones by the amount assumed, let alone tie that rise quantitatively to the cognitive result observed.
Quality of the underlying trials
The review scored included studies on the PEDro scale, a standard tool for assessing randomised-trial methodological quality, and found an average score of 7.9 out of 10 (standard deviation 1.6), which the authors characterise as generally high quality. The majority of included trials scored in the high-quality range (7 to 10), with only 7 studies falling into the moderate range (4 to 6). The domains with the weakest compliance across the trial set were allocation concealment and blinding, a common weak point in supplement trials generally, since a distinctly flavoured ketone drink can be difficult to fully mask against a genuinely indistinguishable placebo.
What the authors say the evidence does not yet show
The authors are explicit about the limits of what a pooled analysis with 91 percent heterogeneity can support. They name several specific gaps: a lack of standardised cognitive endpoints across trials, meaning the "same" outcome domain was sometimes measured with different instruments in different studies; a relative paucity of long-term studies, particularly for ketone esters, which skew toward short and acute protocols; limited long-term follow-up data generally; and the point noted above, that blood ketone concentrations were not consistently measured, so dose-response conclusions rest on administered dose rather than confirmed physiological exposure in every study. They describe the temporal dynamics of the cognitive effect, how quickly it appears and how long it persists after stopping supplementation, as poorly characterised by the current literature.
Key takeaways
A meta-analysis of 29 pooled trial protocols and 1,117 participants found a small-to-moderate, statistically significant improvement in cognitive performance with exogenous ketone supplementation compared to placebo (SMD 0.29), with a positive dose-response relationship that held up in both healthy adults and people with Alzheimer's disease. None of the subgroup differences (supplement type, duration, or population) reached statistical significance on their own, so this analysis cannot say that ketone esters definitively outperform MCT, or that one population responds meaningfully better than another; those are directional signals in the point estimates, not confirmed findings. High heterogeneity (91 percent) and the incomplete measurement of actual blood ketone levels across included trials are real constraints on how confidently any of this generalises. The authors' own framing, that exogenous ketones show a modest, dose-related cognitive benefit and warrant further investigation in longer, better-powered trials, is the accurate summary. It is evidence of a real, replicable, if modest, signal rather than of a settled cognitive-enhancement effect.
FAQ
Does this meta-analysis prove exogenous ketones improve memory?
It shows a statistically significant pooled improvement in cognitive performance across a broad set of outcome measures (SMD 0.29), not a proven memory-specific effect. The included studies used different cognitive tests measuring different domains, and heterogeneity across studies was high (91 percent), which limits how precisely the finding generalises to any single outcome like memory specifically.
Are ketone esters better than MCT oil for cognition?
The point estimate for ketone esters was larger (SMD 0.37) than for MCT (SMD 0.15), but the difference between the two subgroups did not reach statistical significance (interaction p = 0.083). The honest reading is that both showed a positive association with cognitive performance in this dataset, with a hint, not a confirmed finding, that ketone esters may have a larger effect.
Does a higher dose mean a bigger cognitive benefit?
Meta-regression found a statistically significant positive association between daily ketone dose and cognitive improvement, both overall and within the healthy-adult and Alzheimer's disease subgroups specifically. Total cumulative dose and overall duration were not significantly associated with outcome, suggesting the daily dosing level tracked more closely with benefit than simply staying on a product longer.
Is this meta-analysis relevant to healthy people or only to Alzheimer's disease patients?
Both were included and analysed as separate subgroups. Healthy adults showed a larger point estimate (SMD 0.36) than people with MCI or Alzheimer's disease (SMD 0.15), though this subgroup difference also did not reach statistical significance. The trial designs differed meaningfully between the two groups, acute single-dose protocols in healthy adults versus longer MCT or ketone-ester courses in the MCI/Alzheimer's trials, so the two bodies of evidence are not directly comparable outcome-for-outcome.
Primary sources: Bonnechère B, Stephens EB, Boileau AC, Ducker M, Stubbs BJ. "The effect of exogenous ketone bodies on cognition across health and disease: a systematic review and meta-analysis." Frontiers in Nutrition 2026;13:1802531. DOI: 10.3389/fnut.2026.1802531. PMID: 42063954. PROSPERO registration: CRD42023471727. This article summarises published trial data and is not a substitute for individualised medical advice.
Disclaimer
This article is for educational and research-context purposes only. It does not constitute medical advice and is not a substitute for consultation with a qualified healthcare practitioner. The trials and meta-analysis discussed reflect published research findings, not personalised clinical guidance. Always consult a doctor before beginning, modifying, or stopping any supplement or dietary intervention, particularly if you have a diagnosed cognitive or neurological condition, are pregnant or breastfeeding, or take prescription medications.