"Cycle it — eight weeks on, a few weeks off" is among the most repeated pieces of creatine advice in gyms and on product labels. It has the sound of a rule. This page traces where it came from, what was tested in people, and what current labels actually print, because the three do not line up.
The amounts and schedules below are those given in studies and printed on labels, each with its source and date. They record what was administered and declared; none of them is a suggestion.
Where the idea came from
Creatine gets into muscle through a dedicated transporter protein. In 1998, M. L. Guerrero-Ontiveros and Theo Wallimann at ETH Zürich reported that chronic creatine supplementation in rats down-regulated the expression of that transporter in skeletal muscle. If the same happened in people, long-term use would gradually close the door the supplement comes in through, and a break might reopen it. That is the mechanism behind cycling, and the paper is the published source of it.
The dose is the detail that travelled less well than the conclusion. A 2003 paper co-authored by Wallimann himself describes the animal work as using 1–2 g per kg of body weight a day for three months — "between 5 and 10 times" the amounts usually used in human studies.
The human test
That 2003 paper is the direct check. Mark Tarnopolsky's group at McMaster University measured the transporter in human muscle before and after supplementation, in three experiments:
| Who | Creatine given | Length | Transporter result |
|---|---|---|---|
| Young men, weight training (11 creatine, 8 placebo) | 0.125 g/kg a day | 2 months | Protein content unchanged |
| Men and women over 65, weight training (14 creatine, 14 placebo) | 0.075 g/kg a day | 4 months | Protein content unchanged |
| Young men and women (14 creatine, 13 placebo) | about 0.18 g/kg a day | 8–9 days of loading | mRNA unchanged |
Muscle total creatine rose in every creatine group, so the supplement was reaching the muscle. The authors' conclusion: supplementation "resulted in increases in skeletal muscle total creatine without reductions" in transporter protein, and loading "did not decrease" transporter mRNA.
For scale, 0.125 g/kg is 10 g a day for an 80 kg person. The study ran four months at the longest, so it does not speak to years of use, and nobody has repeated the measurement over a longer span. Within what was measured, the mechanism that cycling is meant to counter did not appear.
What has not been tested
No randomised trial has compared cycled with continuous creatine — on strength, muscle size, muscle creatine or any safety marker. MuscleLedger searched PubMed on 11 November 2026 with two phrasings (given at the foot of this page), which returned 40 and 22 records, and read every title. Most concern creatine during cycling exercise or intermittent sprints; the two closest to the question are described below. None randomises people to an on-and-off schedule against daily use.
The consensus documents are silent in the same way. The International Society of Sports Nutrition's 2017 position stand on creatine and its 2021 paper answering common questions and misconceptions were both searched in full text for "cycl": the word appears only for chemical cyclisation, cycling as a sport, the menstrual cycle and the drug cyclosporine. Neither paper recommends or discusses an on-and-off schedule.
So the practice has a hypothesis from rats, a human study that did not confirm the hypothesis, and no trial of the practice itself.
What a break actually does
Three biopsy studies timed how fast a raised muscle store drains after the last dose, and all three put it at about four weeks; the detail is on what happens when you stop creatine. One further study looked at what a cycle does when the supplement is restarted.
Eric Rawson and colleagues (2004) followed a single resistance-trained man through two loading bouts of 20 g a day for five days, 30 days apart:
- Muscle phosphocreatine rose 45% after the first bout and fell 22% over the 30-day break — leaving it still 23% above the original baseline when the second bout began.
- After the second bout it reached a level similar to the first. The break did not let the muscle hold more.
- Urinary creatine losses were 63.2 g in the first bout and 63.4 g in the second, out of 100 g swallowed each time. Uptake was neither blunted by the earlier bout nor improved by the rest.
- Body mass stayed 2.0 kg above the starting weight throughout the break.
It is one person, and the authors call it a case study. What it shows is narrow but relevant: a month off did not fully empty the store, and coming back did not produce a bigger response than the first time.
What 525 labels direct
MuscleLedger read the directions and warnings on every on-market label with "creatine" in the product name in the NIH Office of Dietary Supplements' Dietary Supplement Label Database on 11 November 2026: 525 labels. A text search for cycle, weeks-off, break and time-limit wording matched 29; each was read, and four were false matches (bicycle cycling, the ATP cycle, a "loading cycle").
| What the label says | Labels |
|---|---|
| Directs an on-and-off schedule, with both periods stated | 13 |
| Mentions a cycle with no off period ("run an 8-week cycle"; "if cycling") | 2 |
| Caps continuous use at 8 weeks, with no restart instruction | 6 |
| States that no cycling is needed | 4 |
| Says nothing about cycling or a time limit | 500 |
Twenty-one of 525 — 4% — direct a cycle or a time limit. The 13 full schedules, as printed:
| On | Off | Labels (brand as recorded, DSLD entry year) |
|---|---|---|
| 4–6 weeks | 4–6 weeks | RunFit Nutrition (2023), Nature's Craft (2023), Power Blendz (2024) — identical wording on all three |
| 4 weeks | at least 4 weeks | Nutriforce Sports (2019), Betancourt Nutrition Creatine Micronized (2013) |
| 8 weeks | 4 weeks | Inner Armour (2015) |
| 8 weeks | 2 weeks | Allmax Essentials (2025) |
| 10 weeks | 2 weeks | ILN Iron Labs CEE (2023) |
| 6 weeks | 2 weeks | Betancourt Phosflex (2012) |
| 4 weeks | 2 weeks | MRI Anabolic Switch (2012), Betancourt Creatine Chewies (2013) |
| about 4–5 weeks | 1 or 2 weeks | SAN Performance Creatine (2024) |
| 30 days | 7 days | Axis Labs (2018) |
Three observations from the table.
The schedules do not agree. "On" runs from 4 to 10 weeks and "off" from 7 days to 6 weeks. A rule derived from evidence would converge; these do not, which is what a convention passed between label writers looks like.
Most of the breaks are too short to do what a break is for. Seven of the 13 direct two weeks off or less. The biopsy studies put the drain at about four weeks, one of them found the store still raised at two weeks, and the 2004 case found it still 23% up at 30 days. If the purpose of time off is to return the muscle to its unsupplemented state, a two-week break ends before that has happened.
The four "no cycling needed" labels are all alternative forms — two creatine hydrochloride products, one creatine ethyl ester and one buffered creatine — printing the absence of a cycle as a selling point. Since no trial shows monohydrate needs cycling either, the comparison has nothing on the other side. Those forms are compared on types of creatine.
The reasons labels give
Two of the 13 print a reason. Nutriforce and Betancourt both state that "creatine effectiveness plateaus after 4-5 weeks of continuous use", without a citation. A longer statement on the Betancourt label says researchers found an increase in muscle glycogen "vanished after 5 weeks despite continued creatine supplementation", and concludes from that that cycling "is still a good idea".
The nearest published finding MuscleLedger could open is a 2003 trial from Maastricht (van Loon and colleagues, 20 participants): after loading at 20 g a day for five days, a 2 g a day maintenance amount for up to six weeks "resulted in a decline" in the raised muscle phosphocreatine and total creatine. The same trial reports that repeated-sprint performance was improved after both the short and the longer period. Read plainly, that is a maintenance amount too small to hold the loaded store in those participants — a question about how much, covered on how much creatine trials use — and it is not evidence that a break restores anything. The label's inference from a fading effect to a need for time off is the label's own.
The six labels that cap use at eight weeks give no reason at all.
What continuous use looks like in the trial record
The absence of cycling trials does not mean long continuous use is unstudied. The 2017 position stand summarises clinical trials in which patients took "up to 30 g/day for 5 years", and the Huntington disease trial described on the high-dose trial record gave up to 40 g a day for up to 48 months. Those were daily, uninterrupted regimens in monitored patients with specific diseases. They were not designed to compare schedules, and they cannot say whether a healthy lifter is better off with or without breaks — only that years of daily use is the pattern the long trials followed. What the renal studies measured is on creatine and kidneys.
What is established, and what is not
- Established: the transporter down-regulation that cycling is meant to prevent was reported in rats at 5 to 10 times human doses, and was not found in human muscle after two and four months of daily creatine.
- Established: a raised muscle store takes about four weeks to drain, and in one case study was not fully drained at 30 days.
- Established: in that case study a second loading bout after a month's break reached a similar level to the first, with near-identical urinary losses.
- Not established: that cycling improves the response to creatine, protects the body's own production, or reduces any risk. No trial has tested it.
- Not established: the reverse — that cycling does harm beyond the weeks spent with a lower store. That has not been tested either.
- Not measured: creatine transporter content in people after more than four months of continuous use.
Decisions about starting, stopping or pausing any supplement alongside a medical condition or prescription belong with a clinician.
How this page was put together
PubMed searches run on 11 November 2026: (creatine[ti]) AND (cycling[tiab] OR intermittent[tiab] OR "repeated"[ti]) AND (supplementation[tiab]) AND (washout[tiab] OR "creatine transporter"[tiab] OR loading[tiab]) — 40 records; and creatine[ti] AND supplement*[tiab] AND (intermittent[ti] OR cyclic[ti] OR "on-off"[tiab] OR "cycled"[tiab] OR "cyclical"[tiab] OR "discontinuous"[tiab]) — 22 records. A trial indexed under different wording could have been missed, which is why the claim is stated as "none found". The four study summaries come from their PubMed abstracts; the two society papers were read in open-access full text. The label census used the Dietary Supplement Label Database API (v9): the search "creatine", filtered to on-market records with creatine in the product name, each label's directions and precautions read by text search and every match checked by hand. A cycling instruction phrased in words the search did not include would be counted among the 500. "On the market" is the database's flag, and entries date from 2011 to 2025, so a label may have changed since. Brand names are as the database records them; no product was tested (how we work).
Sources
- Guerrero-Ontiveros ML, Wallimann T. Creatine supplementation in health and disease. Effects of chronic creatine ingestion in vivo: down-regulation of the expression of creatine transporter isoforms in skeletal muscle. Mol Cell Biochem 1998;184:427–437. PMID 9746337 (abstract read 2026-11-11).
- Tarnopolsky M, Parise G, Fu MH, et al. Acute and moderate-term creatine monohydrate supplementation does not affect creatine transporter mRNA or protein content in either young or elderly humans. Mol Cell Biochem 2003;244:159–166. PMID 12701826 (abstract read 2026-11-11).
- Rawson ES, Persky AM, Price TB, Clarkson PM. Effects of repeated creatine supplementation on muscle, plasma, and urine creatine levels. J Strength Cond Res 2004;18:162–167. PMID 14971966 (abstract read 2026-11-11).
- van Loon LJ, Oosterlaar AM, Hartgens F, et al. Effects of creatine loading and prolonged creatine supplementation on body composition, fuel selection, sprint and endurance performance in humans. Clin Sci 2003;104:153–162. PMID 12546637 (abstract read 2026-11-11; DOI 10.1042/CS20020159 verified).
- Kreider RB, Kalman DS, Antonio J, et al. International Society of Sports Nutrition position stand: safety and efficacy of creatine supplementation in exercise, sport, and medicine. J Int Soc Sports Nutr 2017. PMID 28615996 · full text (searched 2026-11-11).
- Antonio J, Candow DG, Forbes SC, et al. Common questions and misconceptions about creatine supplementation: what does the scientific evidence really show? J Int Soc Sports Nutr 2021. PMID 33557850 · full text (searched 2026-11-11).
- NIH Office of Dietary Supplements, Dietary Supplement Label Database, API v9, search "creatine" and 525 individual label records, read 2026-11-11; examples: labels 325793, 53444, 206168, 18086, 305148, 182728.
