Cluster headache is the one headache condition on this site where an edible is the wrong tool for the attack itself, and the arithmetic settles that before the evidence does. The International Classification of Headache Disorders defines an untreated attack as 15 to 180 minutes of severe, strictly one-sided pain around the eye or temple, arriving anywhere from once every other day to eight times a day. Swallowed THC starts working 30 to 90 minutes after the dose and peaks at two to three hours, per Grotenhermen's pharmacokinetic review. Plenty of attacks are finished before the gummy has started.
The treatments that beat that clock are high-flow oxygen and injected sumatriptan. In Cohen, Burns and Goadsby's 2009 JAMA trial, 100% oxygen at 12 litres a minute through a face mask left 78% of treated attacks pain-free at 15 minutes, against 20% on high-flow air. The cannabis research, for anyone hoping it changes the picture, is two surveys and a case report, mostly about smoked cannabis, and it splits roughly three ways: some better, most unsure, some worse. This is the second Neurological page on the site, after multiple sclerosis spasticity, and its position is narrow on purpose.
Can a THC edible stop a cluster headache attack?
Not on any timeline that helps. An untreated attack lasts 15 to 180 minutes, and swallowed THC takes 30 to 90 minutes to begin working and two to three hours to peak. The only published report of oral THC aborting attacks is a single 2009 case, and its abstract gives no dose or timing.
That case deserves a fair hearing, because it is the entire oral evidence base. Robbins and colleagues at the Montefiore Headache Center described one patient, refractory to multiple acute and preventive drugs, who aborted attacks with recreational marijuana and then found dronabinol, the prescription synthetic THC taken by mouth, "equally effective." One person, no onset time in the abstract, and the full text behind a paywall. Every other patient-reported benefit in this literature comes from inhaled cannabis, which Grotenhermen puts at peak effect in 15 to 30 minutes, about when the oxygen mask has finished its work.
Fast-acting gummies close some of the gap on paper. Wana lists 5 to 15 minutes of onset for its fast-acting Quick Relief line and names pain relief as the effect, then tells first-time buyers on the same page to wait at least an hour to see how they feel, "yes, even if the product is fast-acting." A pain gummy that asks for an hour of patience is a strange thing to sell into a condition where the attack can start and finish inside that hour. Our fast-acting edibles comparison explains how those formats work. For this condition the relevant fact is older: Leroux's team called in 2013 for controlled trials before anyone recommends cannabis for cluster headache, and the headache society's 2016 guideline lists no cannabis trial.
Is it safe to take cannabis with verapamil?
Nobody has studied the pair, so the answer rests on labels and mechanism. Verapamil is a CYP3A4 drug and can raise THC levels, while the Epidiolex label found CBD left a sensitive CYP3A4 test drug unchanged. The bigger issue is the ECG monitoring high-dose verapamil already calls for, which a prescriber runs.
Verapamil is the preventive most cluster patients end up on. The American Headache Society calls it "generally regarded as the maintenance prophylactic therapy of choice" while grading it Level C, possibly effective, on one class II and one class III trial at 360 mg a day. In the clinic the doses run far higher. Cohen, Matharu and Goadsby audited 217 London outpatients started at 240 mg and raised by 80 mg every two weeks, with an ECG at each step, to a ceiling of 960 mg. Of the 108 whose ECGs were in the notes, 19% developed an arrhythmia and 36% bradycardia, 13 had first-degree heart block, and one needed a permanent pacemaker. The authors strongly recommended ECG monitoring for everyone on the drug. A 2016 Delphi panel of 22 heart-rhythm specialists agreed only on an ECG before treatment starts, which tells you how settled the monitoring question is.
The usual cannabis worry runs: CBD inhibits CYP3A4, verapamil is cleared by CYP3A4, so verapamil levels climb. Half of that holds. Verapamil's label lists CYP3A4 first among the enzymes that metabolize it, alongside 1A2, 2C8, 2C9 and 2C18. The CBD half does not survive the Epidiolex label, where 750 mg of CBD twice a day left midazolam, a sensitive CYP3A4 test drug, unchanged. The same label does report a roughly 2.5-fold rise in everolimus, a drug moved by both CYP3A4 and the P-glycoprotein pump, and names CBD a moderate CYP2C19 inhibitor. All of that was measured at 1,500 mg of CBD a day, which is a hundred and fifty 10 mg gummies.
The documented direction runs the other way. Herdegen and Cascorbi's 2023 review lists verapamil among the CYP3A4 inhibitors that can raise plasma THC and its active metabolite 11-hydroxy-THC, and puts clinically relevant interactions above 30 mg THC or 300 mg CBD a day. Grotenhermen lists raised heart rate and blood pressure changes among the acute physical effects of too much THC. Nobody has measured what those do to a heart already slowed by verapamil, and this page does not guess.
Alcohol is the documented trigger. ICHD-3 notes that during a cluster period, and at any time in chronic cluster headache, attacks can be provoked by alcohol, histamine or nitroglycerin, and beer was the most common alcohol trigger among the 1,134 respondents to Rozen and Fishman's US Cluster Headache Survey. No source classifies cannabis as a trigger. The closest evidence is the minority in both surveys below who said it made attacks worse. Anyone combining the two between bouts will find the sedation arithmetic in our edibles and alcohol guide.
What does the research on cannabis for cluster headache show?
Two patient surveys and one case report, and the surveys split. In a French clinic sample, a quarter of those who tried cannabis on attacks reported some efficacy, half variable effects and a fifth negative ones. In a Dutch sample, two thirds noticed no change. Unpredictable is the honest summary.
Leroux and colleagues surveyed 139 patients at two French headache centres. Sixty-three, 45.3%, had used cannabis at some point, and 27 had tried it on attacks. Of those 27, 25.9% reported some efficacy, 51.8% variable or uncertain effects and 22.3% negative effects. The authors noted that fewer than a third described relief "following inhalation," and concluded that cannabis should not be recommended for cluster headache unless controlled trials of synthetic selective cannabinoids show a more convincing benefit.
The Dutch study is bigger and flatter. De Coo and colleagues at Leiden sent questionnaires to 756 people with cluster headache and analysed 643. Lifetime cannabis use was 29.5%, against 22.7% in the Dutch general population. Eighty-six had used cannabis during a cluster episode, and on attack frequency 17.4% reported fewer attacks, 67.4% no change and 15.1% more. On attack length, 15.1% said shorter, 66.3% no change and 5.8% longer, with 12.8% unsure. The paper does not report how the cannabis was taken.
Both surveys carry a confound no questionnaire removes. Leroux's cannabis users were more likely than non-users to smoke tobacco, the Dutch cluster group smoked more than the general population, and Rozen's survey describes most US respondents as chronic heavy smokers. Every self-reported cannabis effect in this population is measured in people who often light a cigarette too, and none of these studies can separate the two.
Will a bedtime edible prevent night-time cluster attacks?
No study has tested that, and the timing works against it. US survey respondents reported most attacks between early evening and early morning, peaking between midnight and 3am. Swallowed THC lasts four to 12 hours, so a sedating bedtime dose is still working at the hour the person has to run an oxygen regulator or an injector.
Clusterbusters, the US patient advocacy group, spells out what that 2am treatment involves. Its oxygen page calls for no less than 15 litres a minute through a non-rebreather mask, from a regulator that reaches that flow, and notes that at 15 litres a minute a home E tank holds about 35 minutes of oxygen, fewer than two 20-minute sessions. The group says oxygen, properly used, aborts most attacks within 15 minutes. That is a tank swap and a flow setting in the dark, with a clock running, and no one has studied how well people manage it with 11-hydroxy-THC on board.
No trial has tested any cannabis product as a cluster preventive. The frequency figures from the Dutch survey above, two thirds reporting no change, are the closest thing to data.
How is cluster headache different from migraine?
By what the person does during the attack. ICHD-3 describes cluster patients as usually unable to lie down and pacing the floor, with strictly one-sided pain around the eye and same-side tearing, redness, nasal congestion or a drooping lid. Migraine lasts 4 to 72 hours and is worsened by routine activity, so that patient keeps still.
Clockwork is the other tell. Pain recurs on the same side through a single cluster period, attacks come in bouts lasting weeks or months with remissions between them, and 10 to 15% of patients have the chronic form with no remission. Men are affected about three times as often as women. The distinction matters here because the acute protocol on our migraine page assumes an attack long enough for a fast-onset product to catch, and cluster attacks rarely allow it. Getting the diagnosis takes a while: only 21% of Rozen's respondents were correctly diagnosed at their first presentation, and the average delay ran past five years. Anyone self-treating a "migraine" that makes them pace is worth a neurologist's second look.