The short version
- No published trial has tested an intermittent, cycled or washout dosing schedule for any adaptogen against continuous use, so every cycling protocol in circulation is opinion rather than evidence.
- A 2026 systematic review of 23 randomised studies in 2,317 healthy adults found standardised ashwagandha root extract at 125 to 600mg a day, for durations from a single dose to 180 days, left hepatic, renal, haematological, endocrine and cardiovascular biomarkers within normal ranges with no attributable serious adverse events.
- That same review flags that long-term data beyond 180 days is limited, so continuous use has been studied up to about six months and no further.
- Genuine pharmacological tolerance looks like the 1992 caffeine study, where 300mg three times daily for 18 days produced complete tolerance to caffeine's subjective effects; no adaptogen trial has been designed in a way that could detect that pattern.
- A 2020 case series of ashwagandha-associated liver injury found a latency of two to twelve weeks, meaning the documented risk window is early in exposure — so an on-off schedule re-enters it repeatedly rather than avoiding it.
It is the most confidently repeated piece of advice in this category. Take your adaptogen for eight weeks, then stop for two. Or five days on and two off. Or three months on and one off. The schedules contradict each other, they are stated with total assurance, and every one of them traces back to somebody's opinion rather than somebody's data.
We searched for trials of intermittent versus continuous adaptogen dosing, of cycling protocols, of washout schedules. There are none. Not for ashwagandha, not for rhodiola, not for holy basil, not for reishi. This is an absence-of-evidence article, and the useful thing we can do is tell you precisely what the absence looks like and which of your two underlying worries it applies to.
What the trials actually ran
A 2026 systematic review looked specifically at the safety and tolerability of standardised, single-ingredient ashwagandha root extract in healthy adults. Twenty-three randomised studies, 2,317 participants, doses from 125 to 600mg a day, and intervention durations running from a single dose up to 180 days.
Across those studies, hepatic, renal, haematological, endocrine and cardiovascular biomarkers stayed within normal clinical ranges. No serious adverse events attributable to the extract were reported. Mild events — gastrointestinal discomfort, headache, transient drowsiness — were infrequent and appeared in both intervention and comparator groups. The reviewers' own caveat is the sentence that matters here: long-term data beyond 180 days is limited.
So continuous daily use has been examined up to roughly six months and looks unremarkable. Past six months there is silence. And cycling has been examined for zero days, in zero people. If someone tells you eight-on-two-off is safer than continuous, ask them what they are comparing it to.
Worry one: does the effect fade?
This is the tolerance question, and it is a legitimate one, because tolerance is real and measurable in substances that produce it. Evans and Griffiths demonstrated exactly what that looks like in 1992. Thirty-two healthy people abstained from dietary caffeine, then took either 300mg of caffeine three times daily or placebo for 18 consecutive days. Afterwards, a test dose of caffeine produced significant subjective effects in the group that had been on placebo — and no significant subjective effects in the group that had been taking it. Complete tolerance to a central nervous system effect, in under three weeks.
That is the signature. A crossover, a chronic dosing phase, and a test dose that stops working. Nothing in the adaptogen literature looks like that — but you should be careful with the comfort you take from it, because nobody has designed a study capable of catching it. Almost every trial in this field is a single fixed dose against placebo for eight or twelve weeks, measured once at the end. That design cannot detect a fading effect. It is not evidence of no tolerance; it is evidence of nobody looking.
The one relevant hint runs against the tolerance story. A 12-week randomised placebo-controlled trial in 50 adults aged 65 to 80, taking 600mg a day, found the WHOQOL-BREF total score rose from 140.53 at baseline to 161.84 at the end, with significant improvements in sleep quality (p < 0.0001) and mental alertness (p = 0.034) against placebo. Twelve weeks of continuous use and the curve was still moving in the right direction. One small trial in one population, but it is the opposite of what a fading-effect model predicts.
Worry two: does a break make it safer?
Here the cycling logic gets the shape of the risk backwards. A 2020 case series described five patients — three in Iceland, two through the US Drug-Induced Liver Injury Network — who developed liver injury attributed to ashwagandha-containing supplements. All five developed jaundice. The latency was two to twelve weeks.
Five documented cases against enormous consumption is genuinely rare. But read the latency again. The documented window of risk is early, not late. Someone running eight weeks on and two weeks off is re-entering the first weeks of exposure several times a year rather than escaping anything. If your reason for cycling is safety, the schedule does not deliver it. What does deliver something is knowing the warning signs — jaundice, dark urine, persistent nausea, unexplained itching — and stopping and seeing a doctor if they appear, at any point.
The four honest reasons to take a break
None of them are pharmacological, and all of them are good.
- To find out whether it is doing anything. This is the best reason by a distance. Stop for three weeks. If you cannot tell, you have your answer and you have saved yourself a subscription.
- To attribute a side effect. If something has changed and you are on four things, stopping one is how you find out which.
- Cost. An entirely sufficient reason that nobody needs to dress up as physiology.
- Because the reason you started has passed. These are tools for a period of load. When the period ends, so can the tool.
What we'd actually tell you
Take it daily, at the dose the trials used, for the length the trials ran — eight to twelve weeks — and then stop and see what you notice. That is not a cycle. It is a trial, run on the only participant whose result you care about, and it answers a question that no cycling schedule addresses.
One thing worth adding, because it is the mistake we see most often: running two or three at once makes all of this unanswerable. If you are taking ashwagandha and rhodiola and a mushroom blend together, no break tells you anything, because you cannot attribute either the effect or its absence. One plant, one dose, one stretch of weeks.
And if you are asking about cycling because you suspect the thing has stopped working, consider the less flattering possibility first — that the initial change was novelty and expectation, which fade on their own schedule. That happens with capsules and it happens with slow breathing too. It is not a reason to be cynical. It is a reason to judge these things at eight weeks rather than at week one, and to hold the cheap interventions and the expensive ones to the same standard.
Good questions
Do I need to cycle ashwagandha?
There is no evidence that you do. No trial has ever compared a cycled schedule against continuous use, for ashwagandha or any other adaptogen. Continuous daily use has been studied up to 180 days without biomarker abnormalities or serious adverse events. If you want to stop periodically, do it to test whether the supplement is doing anything, not because a schedule requires it.
Do adaptogens stop working over time?
Nobody has looked properly, so we cannot tell you no. Almost every trial uses one fixed dose against placebo for eight or twelve weeks and measures once at the end, which is a design that cannot detect a fading effect. One 12-week trial in older adults showed quality-of-life scores still improving at the end, which points the other way.
Is it safe to take ashwagandha every day for a year?
Nobody knows, and we will not call that silence reassurance. The safety review covers durations up to 180 days and explicitly flags the lack of longer data. If you have been taking it for six months, you are past the point where trials can tell you anything, which is a reasonable moment to stop and see whether you still need it.
Does taking a break protect my liver?
Probably not, and it may do the opposite of what people intend. The documented cases of ashwagandha-associated liver injury had a latency of two to twelve weeks, so the risk window sits early in exposure. An eight-on, two-off schedule re-enters that window several times a year. What actually helps is knowing the warning signs and stopping if they appear.
How long should I stop for before deciding it wasn't working?
Three weeks off, after eight to twelve weeks on, is enough for most people to notice a difference if there was one. Judge it against how you felt at the start of the whole period rather than against yesterday. And run one thing at a time — if you are taking three supplements together, stopping one tells you nothing you can attribute.
Can I cycle several adaptogens by rotating them?
You can, and you will learn nothing from it. Rotation makes attribution impossible in both directions: you cannot tell which plant helped, and you cannot tell which one caused a side effect. There is also no trial data on any rotation scheme. If you are genuinely experimenting, run one plant at one dose for eight weeks, then the next.
Sources
- Coope OC, Willems MET, Levington A, et al. Back to the Roots: Safety and Tolerability of Standardised Ashwagandha (Withania somnifera) Root Extract in Healthy Adults-A Systematic Review of Biomarkers and Adverse Events. Pharmaceuticals (Basel), 2026. View study
- Evans SM, Griffiths RR. Caffeine tolerance and choice in humans. Psychopharmacology (Berl), 1992. View study
- Kelgane SB, Salve J, Sampara P, Debnath K. Efficacy and Tolerability of Ashwagandha Root Extract in the Elderly for Improvement of General Well-being and Sleep: A Prospective, Randomized, Double-blind, Placebo-controlled Study. Cureus, 2020. View study
- Björnsson HK, Björnsson ES, Avula B, et al. Ashwagandha-induced liver injury: A case series from Iceland and the US Drug-Induced Liver Injury Network. Liver Int, 2020. View study
These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.
