"Does caffeine cancel creatine" is one of the oldest arguments in the supplement aisle, and unusually for a gym belief it has a real paper behind it. What it does not have is a second laboratory.
The finding, and where it came from
In 1996, nine healthy men completed six days of placebo, creatine, or creatine plus caffeine, with phosphorus magnetic resonance spectroscopy of the calf and an isokinetic fatigue test before and after. Muscle phosphocreatine rose 4 to 6 per cent on creatine and on creatine plus caffeine alike. Dynamic torque rose 10 to 23 per cent on creatine — and did not change at all on creatine plus caffeine. The paper's title states the conclusion: Caffeine counteracts the ergogenic action of muscle creatine loading (Vandenberghe K, Gillis N, Van Leemputte M, et al., J Appl Physiol 1996;80(2):452–7, PMID 8929583).
Six years later the same department published the mechanism. In ten men, using electrically stimulated contractions to remove voluntary effort from the measurement, creatine shortened muscle relaxation time by about 5 per cent, three days of caffeine lengthened it by about 10 per cent, and the two together produced no significant change (Hespel P, Op't Eijnde B, Van Leemputte M, J Appl Physiol 2002;92(2):513–8, PMID 11796658).
Both papers come from the Department of Kinesiology at the Katholieke Universiteit Leuven. Hespel and Van Leemputte are authors on both. The whole interference literature is nineteen men, one faculty, and in thirty years no independent group has reproduced it.
The trials that found the opposite
Two studies gave caffeine after creatine loading, as a single dose, and found it still worked.
At the University of Luton, 14 trained men ran to exhaustion at 125% of VO₂max before and after six days of creatine loading. One hour before the post-loading trials they took caffeine at 5 mg/kg or placebo. Time to exhaustion was 222.1 seconds with caffeine against 198.3 with placebo and 200.8 at baseline — the paper is titled Caffeine is ergogenic after supplementation of oral creatine monohydrate (Doherty M, Smith PM, Davison RC, et al., Med Sci Sports Exerc 2002;34(11):1785–92, PMID 12439084).
In Taiwan, 12 physically active men loaded creatine at 0.3 g/kg/day for five days and then performed six 10-second sprints after either 6 mg/kg of caffeine or placebo. Mean and peak power were significantly higher with caffeine than with placebo (Lee CL, Lin JC, Cheng CF, Eur J Appl Physiol 2011;111(8):1669–77, PMID 21207054).
The variable that separates them
Lay the four trials out and one design feature sorts them exactly:
| Trial | n | Caffeine given | Result |
|---|---|---|---|
| Vandenberghe 1996 (Leuven) | 9 | 5 mg/kg/day for 6 days, during loading | creatine's torque gain abolished |
| Hespel 2002 (Leuven) | 10 | 5 mg/kg/day for 3 days | creatine's relaxation effect abolished |
| Doherty 2002 (Luton) | 14 | 5 mg/kg once, after loading | caffeine still ergogenic |
| Lee 2011 (Taiwan) | 12 | 6 mg/kg once, after loading | caffeine still ergogenic |
Every trial that dosed caffeine for days found interference; every trial that gave one dose found none. That is a clean split on a single variable, and it is the most useful thing in this literature — but it is four small trials with two laboratories on one side, so it is a hypothesis that fits the data rather than a settled mechanism.
It also exposes the gap. Nobody has tested the pattern people actually follow: daily creatine maintenance alongside a daily pre-workout coffee, over weeks. The days-of-caffeine trials ran during a 20-gram loading phase; the single-dose trials ran after loading ended. Neither is the habit.
The largest trial, and why it cannot adjudicate
Fifty-four physically active men were randomised to five days of creatine alone, creatine plus 300 mg of caffeine anhydrous, creatine plus 8.9 g of coffee yielding 303 mg of caffeine, or placebo, then retested on bench press, leg press and repeated cycle sprints. Strength improved in every group, with no interaction by treatment (Trexler ET, Smith-Ryan AE, Roelofs EJ, et al., J Strength Cond Res 2016;30(5):1438–46, PMID 26439785).
It is the biggest and best-designed trial in the field, it is the only one that separated coffee from caffeine powder, and it cannot settle the question — because it did not detect an effect of creatine alone either. A trial that fails to reproduce the thing being interfered with cannot measure interference. Its useful contributions are elsewhere: serum creatinine rose in every creatine group and not in placebo, which is the laboratory artefact discussed on what creatine means in bloodwork, and four participants on creatine plus caffeine reported mild gastrointestinal discomfort, with none in the other groups.
Additivity, tested once
In 2024, 23 endurance-trained men and women loaded creatine or placebo for five days and then continued a maintenance amount while testing caffeine at 3 mg/kg, sodium bicarbonate, both, or placebo. Creatine raised 15-second sprint power by 34 watts (95% CI 10 to 58). Caffeine raised 6-minute time-trial power by 12 watts (95% CI 5 to 18). Neither added to the other: caffeine contributed +10 W to the sprint (CI −7 to 24) and creatine −6 W to the time trial (CI −15 to 4) (Moesgaard L, Jessen S, Christensen PM, et al., Scand J Med Sci Sports 2024;34(4):e14629, PMID 38646853).
Each improves a different thing, and stacking them does not improve either further.
A six-week resistance-training trial points the same way more faintly. In 28 trained young adults, only the creatine-alone group increased knee-extensor muscle thickness; creatine plus caffeine did not, although with six to nine people per group the trial had little chance of detecting much (Pakulak A, Candow DG, Totosy de Zepetnek J, et al., J Diet Suppl 2022;19(5):587–602, PMID 33759701).
Reading the literature count honestly
PubMed returns 30 records tagged as randomised controlled trials mentioning both creatine and caffeine. Most are not about the interaction. Classified by title and abstract, they are dominated by trials of multi-ingredient pre-workout products — Bang Master Blaster, synephrine blends, amino-acid and B-vitamin formulas — in which creatine is one ingredient on a label and no arm isolates it. One is a 1994 treadmill study of caffeine and ethanol that matched only because it measured creatine phosphokinase, and one is a trial of oral creatine for apnoea of prematurity.
A count of 30 becomes four relevant trials once you read them, which is the difference between a literature and a search result.
The one large human dataset, and it is not about performance
In the Parkinson disease trial that gave 1,741 people creatine 10 grams a day for a planned five years, an analysis of the 1,549 participants with caffeine-intake data found no association between caffeine and disease progression overall — but higher caffeine intake was associated with significantly faster progression among those taking creatine (Simon DK, Wu C, Tilley BC, et al., Clin Neuropharmacol 2015;38(5):163–9, PMID 26366971).
Caffeine intake was observed, not assigned, so this is an association inside a randomised trial rather than a randomised comparison. It is nonetheless the largest human dataset in which the two substances are examined together, it concerns a clinical outcome rather than a sprint, and it is absent from every consumer discussion of the pair. The rest of what is and is not known about creatine and the brain is set out on creatine and the brain, and the safety tables of that same trial are read on is creatine bad for you.
What this page does not say
It gives no instruction about when to consume either substance. It reports what four small trials measured, notes that two of them share a laboratory, and identifies the design variable that separates their results. Anyone wanting the rest of the creatine timing literature will find it on creatine timing; how much the trials administered is on how much creatine should I take.
Key figures
- Interference evidence: 19 men, two trials, one department (KU Leuven)
- Torque gain on creatine: +10 to 23% · with caffeine added: no change
- Relaxation time: creatine −5%, caffeine +10%, together unchanged
- Single-dose caffeine after loading: 222.1 s vs 198.3 s to exhaustion
- Largest trial: n = 54, coffee vs anhydrous, no effect of anything, creatine included
- Additivity trial: creatine +34 W sprint, caffeine +12 W time trial, no cross-benefit
Sources
- Vandenberghe K, Gillis N, Van Leemputte M, et al. Caffeine counteracts the ergogenic action of muscle creatine loading. J Appl Physiol 1996;80(2):452–7 — PMID 8929583 · doi:10.1152/jappl.1996.80.2.452
- Hespel P, Op't Eijnde B, Van Leemputte M. Opposite actions of caffeine and creatine on muscle relaxation time in humans. J Appl Physiol 2002;92(2):513–8 — PMID 11796658 · doi:10.1152/japplphysiol.00255.2001
- Doherty M, Smith PM, Davison RC, et al. Caffeine is ergogenic after supplementation of oral creatine monohydrate. Med Sci Sports Exerc 2002;34(11):1785–92 — PMID 12439084 · doi:10.1097/00005768-200211000-00015
- Lee CL, Lin JC, Cheng CF. Effect of caffeine ingestion after creatine supplementation on intermittent high-intensity sprint performance. Eur J Appl Physiol 2011;111(8):1669–77 — PMID 21207054 · doi:10.1007/s00421-010-1792-0
- Trexler ET, Smith-Ryan AE, Roelofs EJ, et al. Effects of Coffee and Caffeine Anhydrous Intake During Creatine Loading. J Strength Cond Res 2016;30(5):1438–46 — PMID 26439785 · doi:10.1519/JSC.0000000000001223
- Moesgaard L, Jessen S, Christensen PM, et al. No additive effect of creatine, caffeine, and sodium bicarbonate on intense exercise performance in endurance-trained individuals. Scand J Med Sci Sports 2024;34(4):e14629 — PMID 38646853 · doi:10.1111/sms.14629
- Pakulak A, Candow DG, Totosy de Zepetnek J, et al. Effects of Creatine and Caffeine Supplementation During Resistance Training on Body Composition, Strength, Endurance, Rating of Perceived Exertion and Fatigue in Trained Young Adults. J Diet Suppl 2022;19(5):587–602 — PMID 33759701 · doi:10.1080/19390211.2021.1904085
- Vanakoski J, Kosunen V, Meririnne E, et al. Creatine and caffeine in anaerobic and aerobic exercise: effects on physical performance and pharmacokinetic considerations. Int J Clin Pharmacol Ther 1998;36(5):258–62 — PMID 9629989
- Simon DK, Wu C, Tilley BC, et al. Caffeine and Progression of Parkinson Disease: A Deleterious Interaction With Creatine. Clin Neuropharmacol 2015;38(5):163–9 — PMID 26366971 · doi:10.1097/WNF.0000000000000102
- National Library of Medicine. PubMed E-utilities — census of randomised controlled trials mentioning creatine and caffeine, and author affiliations, run 2026-09-19 — pubmed.ncbi.nlm.nih.gov
All sources retrieved 2026-09-19. Literature counts are as of that date.