Cannabis withdrawal: how long does it actually last?

Published 12 September 2026 · 6 min read

By Loutra · Sources (9)

The month everyone quotes is a urine test

You stopped a few days ago. You are on edge, you are not sleeping, you are not hungry, and someone told you that "THC takes a month to leave your system". That is the sentence that makes the week feel endless: if the drug needs thirty days to clear, the withdrawal must last thirty days.

The figure is real. It simply answers a different question.

FRANK, the UK government's drug information service — the page the NHS now redirects its own cannabis address to — puts it like this: "If you've used cannabis as a one-off, it will show up in a urine test for around 2 to 3 days afterwards. However, this can go up to a month for regular users."

A urine test. And what that test looks for is not even THC: it is THC-COOH, an inactive breakdown product the body releases slowly, because THC "concentrates in human fat". Marilyn Huestis, whose review of cannabinoid pharmacokinetics is the reference on this, spells out what the result is worth: "A positive urine test for cannabinoids indicates only that drug exposure has occurred. The result does not provide information on the route of administration, the amount of drug exposure, when drug exposure occurred, or the degree of impairment."

The trace and the withdrawal are two different calendars. The second one is shorter, and it starts far sooner.

The real calendar fits inside three weeks

Ulrich Bonnet and Ulrich Preuss pulled together what is known about cannabis withdrawal in Substance Abuse and Rehabilitation. The syndrome is in the DSM-5, and the definition already carries a timeframe: it is diagnosed when three or more symptoms "develop within ~1 week after quitting cannabis use abruptly". The six: irritability, nervousness or anxiety, sleep difficulty, decreased appetite, depressed mood, and one physical symptom off a list — abdominal pain, shakiness, sweating, fever, chills, headache.

In heavy users watched day by day through an inpatient detox, the curve is clean. Physical symptoms start at a mean of 1 to 3 days after the last use, psychological ones at 2 to 10 days. Then, in the authors' words: "mean intensity of irritability, nervousness, restlessness, and anger increased until day 4 and then decreased. Craving and sleeplessness peaked on day 2."

• 1 to 3 daysPhysical symptoms arrive: sweating, shakiness, headaches, feeling sick.
• Day 2Craving and sleeplessness hit their maximum. The worst possible moment to judge the rest.
• Day 4Irritability, nervousness, restlessness and anger peak — and strange dreams, which turn up late, peak here too.
• 2 to 10 daysThe psychological symptoms settle in: low mood, anxiety.
• 14 to 21 daysThe syndrome fades for most people. It "usually lasts up to 3 weeks".
• ~4 weeksBrain CB1 receptors are back to normal, after 26 ± 5 days of monitored abstinence (Hirvonen 2012).

That last line needs a caveat. Jussi Hirvonen and his team scanned the brains of 30 daily smokers and 28 controls: CB1 receptor binding was "~20% lower in neocortex and limbic cortex", and back to normal "in most regions" in the 14 who were rescanned after four weeks of monitored abstinence. The brain's calendar looks like the felt one — but the same authors note that receptor density was "not significantly correlated with withdrawal or craving". Two clocks running side by side are not a demonstrated cause, which is exactly the limit this blog already drew around the third day without nicotine.

It is neither rare nor in your head

Second received idea: cannabis supposedly gives you no "real" withdrawal. A meta-analysis in JAMA Network Open pooled 47 studies and 23,518 regular or dependent users. Prevalence of the withdrawal syndrome: 47% (95% CI, 41% to 52%). Close to one in two.

The breakdown says more than the total. In population-based samples the figure falls to 17%; among people in outpatient care it rises to 54%; among inpatients, 87%. The authors also find daily use and concurrent tobacco use associated with higher prevalence — more on that below.

On intensity, Bonnet and Preuss write that in heavy users, average peak severity "is comparable to that of a moderate depression or alcohol withdrawal syndrome". ⚠️ That comparison is about how much it costs to live through, not about medical risk. Alcohol withdrawal can kill and needs a doctor in advance — that is the subject of the article on stopping drinking suddenly. Nothing equivalent is described for cannabis: here the same authors talk about symptoms "of light to moderate intensity, which can usually be treated in an outpatient setting". And the next sentence in their review is the one that describes your situation: "In outpatient settings, the average discomfort of CWS was similar to that of tobacco withdrawal."

What the NHS and FRANK actually say

The NHS no longer keeps a cannabis page of its own — its old address now redirects to FRANK, which remains the UK's public reference on the subject. On dependence, FRANK answers in one word: "Yes. Heavy cannabis users often get cravings and find it hard not to take the drug – even when they know it's causing them physical, mental or social problems." On what happens when they stop, the list is short and concrete: they can "feel moody and irritable", "feel sick", "find it hard to sleep", "find it hard to eat", "experience sweating and shaking", "get diarrhoea".

No duration anywhere. It is the same pattern as with alcohol: public health services describe symptoms and date almost nothing, because dating commits you. The timeline comes from the literature, not from them. On what to do, though, the NHS is clear: "A GP may be a good place to start. They can discuss your problems with you and help get you into treatment." FRANK's own line is free and confidential on 0300 123 6600.

Part of what you are feeling is not cannabis

This is the piece most timelines skip. Chandni Hindocha and colleagues analysed how 33,687 people in the Global Drugs Survey actually consume: 65.6% used a tobacco-based route of administration, and the geography is stark — 77.2% to 90.9% in Europe, against 4.4% to 16.0% in the Americas. The joint rolled with tobacco is a European habit, not a worldwide constant.

FRANK flags the consequence on the same page: people who mix cannabis with tobacco are "at risk of getting addicted (or staying addicted) to nicotine". So if that describes you, stopping cannabis means stopping two drugs on the same morning — and only one of them has a well-mapped withdrawal with its own week-by-week timeline and a treatment whose efficacy is well established. Sorting out which is which is the single most useful thing you can do in week one.

For cannabis itself, there is no medicine

The Cochrane review published on 30 September 2025 pooled 37 randomised trials and 3,201 participants. It opens on a flat statement: "There are currently no medicines specifically for the treatment of cannabis use disorder." On abstinence at the end of treatment, THC preparations, CBD, N-acetylcysteine, oxytocin and PF-04457845 are "probably ineffective", and the authors conclude that these treatments "should still be considered experimental". The same page says what is left, in one line: "Currently, psychological treatments are the only recommended treatment for cannabis use disorder."

Which, in practice, means the unglamorous version: a GP, a local drug service, and days. Loutra counts those days and asks for one check-in each — and on this subject the counting has a shape worth knowing. The first four days carry almost all the load, the third week is where the thing ends for most people, and nothing in the literature says the month you were dreading was ever the right unit.

Questions people ask

Loutra is free on the App Store

An otter that’s counting on you, a daily check-in, and a breathing round for the minutes that are hard.