Primary endpoint
The main prespecified outcome the study is designed around. Ideally the sample-size calculation is tied to it, but check the protocol or registration rather than assuming.
Evidence literacy · Research methods
How to Read an Exogenous Ketone Study
Ten checks that work on any ketone paper — then one Veech Ketone Ester trial read all the way through, so you can see what each check turns up in practice.
Short answer
Short answer
Start with the exact molecule and the dose. Then ask who took it, against what, what was measured, and whether the finding the headline leans on was the one the study set out to test.
Three common shortcuts in reading ketone research are treating “ketones” as one thing, treating a blood ketone number as a result, and quoting a secondary or exploratory finding as if it were the main one. The ten checks below are built to catch all three. They apply to any exogenous ketone study — the worked example uses a Veech Ketone Ester trial because that is the molecule this site indexes.
The method
Ten checks, in the order a paper answers them.
Methods first, Results second, then funding, tolerability and the published correspondence.
- 1Start with the exact molecule.Find the chemical name of what was given. “Ketone ester,” “KME,” “BHB” and “exogenous ketones” are categories, and different molecules in each category have different results.Where: Methods — the drink or product description, and who supplied it.
- 2Check dose and timing.Grams or g/kg, how often, how long before or after exercise, and for how many days. A protocol is part of the finding — the same molecule at a different dose or time is a different experiment.Where: Methods — the treatment schedule, often a figure.
- 3Check population and training status.Professional cyclists, trained cyclists, recreational runners and healthy adults are different study populations. Results do not automatically generalize from one to another. Note how many were recruited and how many were analysed.Where: Methods — participants; Table 1.
- 4Check the comparator and the context.What was the placebo, and was it inert? Were participants fed or fasted, and was carbohydrate given alongside? Was the study blinded, and did the authors check the blinding?Where: Methods — control drink, dietary control, blinding.
- 5Separate primary, secondary and exploratory endpoints.The Methods and any trial registration say which outcomes were prespecified. A secondary or exploratory finding can be informative, but its weight depends on whether it was prespecified, how many outcomes were tested, statistical power and multiplicity. Do not silently promote it to the study’s primary result.Where: Methods — aims and statistics; the registration record if one exists.
- 6Separate the biomarker from the outcome.Blood BHB going up is exposure — it confirms the drink worked as chemistry. Performance, cognition and recovery are outcomes, and they have to be measured on their own.Where: Results — the BHB time course versus the performance or cognition tables.
- 7Weigh statistical against practical significance.Look past the p-value to the size of the effect and its confidence interval. The practical meaning of a 2% change depends on the task, athlete level, normal day-to-day variability and the uncertainty around the estimate. Report the size of the effect, not just the p-value.Where: Results — effect sizes, CIs, the between-group comparison.
- 8Check adverse events and tolerability.Tolerability affects real-world applicability. Record symptoms, withdrawals and whether the tested dose is practical outside the laboratory.Where: Results — tolerability scores, withdrawals; Discussion.
- 9Check funding and conflicts — without assuming bias.Who paid, who supplied the drink, and whether the authors hold a commercial interest. A sponsor-funded trial can be well designed; an independent one can be weak. Funding tells you where to look harder, not what to conclude.Where: Funding, Competing interests, Acknowledgements — usually the last paragraphs.
- 10Ask whether the protocol resembles real use.Supervised lab training, fixed diets and standardized tests are what make a trial clean. They are also what separate it from a normal week. Say which parts transfer to your situation and which do not.Where: Methods — design; Discussion — limitations; the correspondence that followed publication.
A citation is a claim about one study. These ten questions are how you check it against the paper.
Then match it to the product
Check 5, in detail
Not every result in a paper carries the same weight.
Trials measure many things. Three tiers tell you how much a given number was designed to bear.
Secondary endpoints
Additional prespecified outcomes. They can be meaningful, but interpretation depends on the study design, statistical power, multiplicity and whether the analysis plan specified how they would be tested.
Exploratory and post-hoc
Exploratory outcomes may be prespecified but are generally hypothesis-generating. Post-hoc analyses are defined after the data are available. Both deserve more caution than a prespecified primary analysis — and they are not the same thing.
Some papers state a single primary endpoint; many exercise studies state an aim and report a panel of outcomes. When the hierarchy is not explicit, read the aims paragraph and the statistical-analysis section, and treat anything the authors call preliminary as Tier 3. Appearing in the abstract does not move a finding up a tier.
Worked example
One Veech Ketone Ester trial, read with all ten checks.
The 2019 overload-training study is one of the most-cited Veech Ketone Ester papers — and the one behind the “+15%” figure. Here is what a careful read turns up, including the parts marketing usually leaves out.
Three weeks of overload training: 15% higher sustainable training load and late-endurance power
Poffé C, Ramaekers M, Van Thienen R, Hespel P · The Journal of Physiology · 2019;597(12):3009–3027 · full title in the reference list · PubMed ↗ · Study index
Fit young men completed three weeks of supervised, deliberately excessive cycling training — 28 sessions, ending at roughly three times their usual volume. Half received 25 g of the ester after each session and before sleep; half received a matched control drink. The question was whether the ester would blunt the signs of overreaching.
| Check | What the paper says | Read it as |
|---|---|---|
| 1Molecule | Methods name the compound — (R)-3-hydroxybutyl (R)-3-hydroxybutyrate at 96% — and the supplier. | Clean link. Anyone can check it against a label. |
| 2Dose & timing | 25 g immediately after every session and 25 g 30 minutes before sleep, six days a week, three weeks. Post-dose blood BHB about 2.6 mM within 30 minutes. | Specific and repeated. Not a pre-workout protocol — a result from this schedule does not describe a single dose before exercise. |
| 3Population | Healthy, physically active men, about 21 years old, exercising under 5 h/week, none regular cyclists. Two of 20 recruits did not finish; 18 analysed. | Small, and not trained cyclists. A small sample of physically active young men who were not regular cyclists — nine per group. Not the trained cyclists a product page might imply. |
| 4Comparator & context | Control drink was isocaloric MCT oil, taste- and colour-matched. Everyone also got a standardized protein–carbohydrate recovery drink. Double-blind; after the study, 10 of 18 guessed their group correctly. | Well controlled, with one thing to note. The control was an isocaloric 16.4 g MCT drink rather than a noncaloric placebo. In this study, blood BHB remained unchanged in the control condition. It was still a nutritional comparator, so its composition belongs in the interpretation. The blinding check is useful context because it shows whether masking plausibly held. |
| 5Endpoint hierarchy | The stated aim was overreaching prevention. Reported: training load, a 30-minute time trial, a 90-second sprint, a 2-hour endurance test, heart rates, hormones, energy intake, glycogen, bone, immune and stress measures. | A wide panel. Performance was one of many outcomes. The 15% figures come from sustainable training load in week 3 and the final 30 minutes of the 2-hour test — read those, not the abstract, to know exactly what improved. |
| 6Biomarker vs outcome | BHB rose to about 2.6 mM after each dose. Separately, week-3 training load and 2-hour-test power were about 15% higher than control, with a confidence interval of roughly +5 to +52 W on the power result. | Both columns present. The paper measured exposure and outcomes separately, which is what lets it be cited for an outcome at all. |
| 7Statistical vs practical | The 30-minute time trial improved within the ester group over time, but the between-group difference was not statistically significant at any point. The 2-hour-test and training-load differences were. | Read the between-group line. Two performance tests, two answers. The accurate summary is “higher sustained power under overload,” not “faster time trials.” |
| 8Tolerability | GI discomfort rose slightly across the study in both groups and did not differ between them. One recruit left over the protein–carbohydrate drinks, not the ester. | Tolerated at this dose and schedule. Three doses a day for three weeks without a tolerability gap versus control is itself a finding. |
| 9Funding & conflicts | Funded by a public research foundation. Ester purchased from the manufacturer. Authors declare no competing interests. | No commercial sponsor listed. Funded by Research Fund Flanders, ester purchased from TdeltaS, no competing interests declared. That removes one potential commercial conflict; it does not replace the other nine checks. |
| 10Real-world fit | Every session was supervised in a laboratory and volume tripled by week 3. Food intake outside the supplied drinks was not fixed: by week 3 the ester group was eating more total energy — predominantly additional carbohydrate — while the control group had developed an energy deficit. The authors noted that the late-endurance performance difference coincided with this higher carbohydrate intake. Three letters to the editor raised statistical and interpretive questions; the authors replied in the same journal. | A deliberate-overload model with an open pathway. It shows what the ester did under extreme, supervised training in physically active non-cyclists. The energy-intake difference does not erase the between-group result, but it makes the pathway to that result harder to attribute to one factor alone. Reading the correspondence is part of reading the record — the links are below. |
Every entry above is taken from the open-access paper and its linked correspondence. Where a number is approximate on this page, the exact value and confidence interval are in the paper's Results.
What this paper supports
Higher sustainable training load and late-endurance power, under this protocol
Veech Ketone Ester, 25 g after training and before sleep, three weeks of supervised overload, fit young men, versus an MCT control.
What it does not
“Ketones improve performance”
Not a pre-workout result. Not a time-trial result versus control. Not a result for other molecules, other doses, or athletes who train normally.
The checks at work
What the same checks turn up in other ketone trials.
Four papers, one check each — including a negative result, because a reading method that only finds good news is not a method.
One more context check: in a human pharmacokinetic study of the Veech Ketone Ester, eating a meal before the drink lowered the peak blood BHB by about a third. Fed versus fasted is part of check 4, and it changes what the same dose does. Stubbs et al. · Front Physiol 2017.
Weighing a citation
How much can one citation carry?
After the ten checks, place the paper on this ladder. The rung sets the ceiling on what the citation can support for a product claim.
- 1Controlled human trial · same molecule · measured the claimed outcomeDirect product-specific evidence for that outcome, under that protocol.Direct support
- 2Controlled human trial · same molecule · measured blood BHB onlyShows the molecule delivers ketones. Says nothing about performance, cognition or recovery.Exposure only
- 3Controlled human trial · different molecule in the same categoryContext for the category. The result belongs to the molecule that was tested.Context
- 4Pooled review or meta-analysis across moleculesUseful for the field. Not product-specific unless it reports molecule-level results.Context
- 5Mechanism, animal or cell studiesExplains how something could work. Does not show that it did, in people, at a dose.Hypothesis
- 6Testimonials and athlete endorsementsExperience, not evidence. Read them as marketing.Not evidence
Run the ten checks. Place the paper on the ladder. Then decide what the citation can fairly carry — and hold the product page to that.
The product-match test
From evidence to the bottle
Finished reading the paper? Check the bottle.
If a result matters to you, the next question is identity: does the product you buy actually contain the molecule used in the trial? KetoneAid products use Veech Ketone Ester.
Dose, timing, population and outcome still matter. The point is to start with the same molecule before asking what the study can support.
Go deeper
Read the papers.
Study index
Clinical studies
Every indexed Veech Ketone Ester human study, with molecule, study material and result — filterable by area and outcome.
Browse the studies The match testProduct-specific evidence
The five-link chain between a product and a claim, and the questions that test it.
Run the test The moleculesExogenous ketones
Ketone ester, R 1,3 butanediol, BHB salts and D-BHB free acid — what each one is and what it has shown.
Compare the types The researchVeech Ketone Ester science
The molecule, where it came from, and the human evidence behind it.
Explore VKE scienceReferences cited on this page
- Poffé C, Ramaekers M, Van Thienen R, Hespel P. Ketone ester supplementation blunts overreaching symptoms during endurance training overload. J Physiol. 2019;597(12):3009–3027. PubMed ↗ · Open access ↗ · Study indexWorked example. Funded by Research Fund Flanders (G080117N); no competing interests declared; ester purchased from TdeltaS Ltd.
- Korevaar DA, Cohen JF, McInnes MDF. Ketone ester supplementation in endurance athletes: a miracle drink or ‘spin’? J Physiol. 2019;597(16):4407–4408. PubMed ↗Letter to the editor on the worked example. Authors’ reply: J Physiol. 2019;597(16):4409–4410, PubMed ↗.
- Mullie P, Autier P. Type 1 error. J Physiol. 2019;597(17):4677–4678. PubMed ↗Letter to the editor on the worked example. Authors’ reply: J Physiol. 2019;597(17):4679–4680, PubMed ↗.
- Bellinger P. Does ketone ester supplementation really blunt overreaching symptoms during endurance training overload? J Physiol. 2019;597(21):5307–5308. PubMed ↗Letter to the editor on the worked example. Authors’ reply: PubMed ↗.
- McCarthy DG, Bone J, Fong M, et al. Acute Ketone Monoester Supplementation Impairs 20-min Time-Trial Performance in Trained Cyclists: A Randomized, Crossover Trial. Int J Sport Nutr Exerc Metab. 2023;33(4):181–188. PubMed ↗(R)-3-hydroxybutyl (R)-3-hydroxybutyrate, 0.35 g/kg, 30 min pre-exercise; triple-blind crossover; n = 23. Funded by an NSERC Discovery Grant.
- Prins PJ, Buga A, Storoschuk K, et al. The Effects of 31-Day Exogenous Ketone Consumption on Running Performance, Cognitive Function, Metabolism, Body Composition, Hemodynamics, and Mood in Recreational Runners: A Randomized-Control Trial. J Am Nutr Assoc. 2026;45(6):545–560. PubMed ↗ · Study index
- Falkenhain K, Daraei A, Little JP. The Effect of Novel Exogenous Ketone Supplements on Blood Beta-Hydroxybutyrate and Glucose. J Diet Suppl. 2024;21(1):38–52. PubMed ↗
- Leckey JJ, Ross ML, Quod M, Hawley JA, Burke LM. Ketone Diester Ingestion Impairs Time-Trial Performance in Professional Cyclists. Front Physiol. 2017;8:806. PubMed ↗1,3-butanediol acetoacetate diester — not the Veech Ketone Ester.
- Stubbs BJ, Cox PJ, Evans RD, et al. On the Metabolism of Exogenous Ketones in Humans. Front Physiol. 2017;8:848. PubMed ↗ · Study indexFed vs fasted: a prior meal lowered peak D-BHB by 33% (2.2 vs 3.3 mM).