One Randomized Trial and a Pile of Survey Data: Where Cannabis for Headache Actually Stands

Until this trial appeared, there was no randomized, placebo-controlled trial of cannabis for any headache disorder. There is now exactly one, in acute migraine, using vaporized flower. Vaporized cannabis containing 6 percent THC plus 11 percent CBD beat placebo for pain relief at two hours, 67.2 percent versus 46.6 percent.

That single result does not extend to tension-type headache, cluster headache, or headache prevention, and it sits next to a separate and well-replicated observation: people with chronic migraine who use cannabis carry a substantially higher burden of medication overuse headache.

Headache patients are among the most experimentally inclined people in medicine, because the conditions are common, disabling, and frequently undertreated. Many arrive having tried cannabis before they ever raise it with a clinician.

The clinical task is not to approve or forbid. It is to say accurately which headache type has evidence behind it, which does not, what a reasonable trial of cannabis looks like, and what pattern of use turns a rescue option into a driver of more headache days.

Nathaniel M. Schuster and colleagues at the University of California San Diego ran a randomized, double-blind, placebo-controlled crossover trial in adults with migraine, published in Headache. Ninety-two participants were enrolled, and 247 migraine attacks were treated. Each participant treated up to four separate attacks, one each with vaporized cannabis flower containing 6 percent THC, 11 percent CBD, 6 percent THC plus 11 percent CBD, or placebo flower, in randomized order with at least a week between treated attacks.

The combination arm was the one that worked. THC plus CBD beat placebo for pain relief at two hours (67.2 percent versus 46.6 percent, odds ratio 2.85), for pain freedom (34.5 percent versus 15.5 percent), and for freedom from the most bothersome symptom (60.3 percent versus 34.5 percent). Benefits on pain freedom and most bothersome symptom freedom held at 24 and 48 hours. There were no serious adverse events.

The THC-only arm beat placebo for pain relief but not for pain freedom or most bothersome symptom freedom. The CBD-only arm did not separate from placebo on any of the three endpoints. That is a meaningful negative result for anyone treating an acute migraine with a CBD product.

This is a real finding and a narrow one. It describes inhaled flower at specified potencies, treating an attack that has started, in people who already had migraine and consented to inhale cannabis under observation. It is not a statement about edibles, tinctures, topicals, prevention, or any headache type other than migraine.

The largest dataset on cannabis and headache is not a trial. Carrie Cuttler and colleagues at Washington State University analyzed archival records from Strainprint, a symptom-tracking app, covering 12,293 sessions where cannabis was used for headache and 7,441 sessions where it was used for migraine. Published in The Journal of Pain, the analysis found self-reported severity reductions of roughly 50 percent after inhalation.

That number gets quoted often and deserves several caveats at once. There was no placebo and no blinding, users chose their own product and dose, and people who find something unhelpful stop logging it. The same analysis found evidence of tolerance: effectiveness diminished over time and users escalated their doses.

For tension-type headache and cluster headache, the controlled literature is essentially empty. Claims that specific cannabinoids or specific products treat those conditions are extrapolations, not findings. A trial of a cannabidiol-enriched cannabis herbal extract in adolescents with chronic migraine that has not responded to other therapies has been designed and registered, and it is a dose-escalation tolerability study rather than an efficacy trial.

The 2012 randomized trial by Luigi Alberto Pini and colleagues in Modena is a genuine controlled result, but it tested nabilone, a synthetic THC analogue, not cannabis, in 30 patients with long-standing intractable medication overuse headache. Nabilone 0.5 mg daily outperformed ibuprofen 400 mg on pain intensity and daily analgesic intake, and reduced measured medication dependence. Twenty-six patients completed. It is a small, single-center study that has not been replicated at scale.

Medication overuse headache is a recognized secondary headache disorder in which frequent use of acute headache treatment produces more headache days rather than fewer. It is the reason headache specialists limit how many days per month a patient uses a triptan or a combination analgesic, and it is the single most important structural risk in treating headache with any rescue medication.

Niushen Zhang and Yohannes W. Woldeamanuel at Stanford reviewed 368 charts of adults with chronic migraine of at least one year’s duration, seen at their headache clinics between 2015 and 2019. Medication overuse headache was present in 81 percent of current cannabis users compared with 41 percent of non-users, an adjusted odds ratio of 6.3.

Their analysis also found a bidirectional association between cannabis use and opioid use: patients using one were more likely to be using the other. Hierarchical clustering separated the sample into two groups, one with roughly nine times the cannabis use, nine times the opioid use, and nearly twice the medication overuse burden of the other.

This is a case-referent study in a tertiary headache clinic, so it cannot establish that cannabis causes medication overuse headache. Patients with the worst headache burden use more of everything. But the size of the association, its consistency with the general behavior of acute headache treatments, and the fact that it was found in exactly the population most likely to reach for cannabis make it a risk worth naming out loud.

Route determines how quickly a drug arrives and how long it lasts, and headache care depends heavily on both. Inhaled routes reach peak blood concentration within minutes and fade over a few hours. Oral routes take considerably longer to arrive, produce a different metabolite profile, and last longer. The details of what reaches the bloodstream from each route are covered in our review of measured cannabinoid absorption by method.

The migraine trial used vaporized flower for a defensible reason. An attack that has already begun needs something that acts fast, and gastric stasis during migraine makes oral absorption unreliable, which is the same reason nasal and injectable triptan formulations exist.

No controlled trial has tested an oral cannabis product, a tincture, or a topical for any headache disorder. Published tables that assign onset times and durations to those formats for headache are reporting numbers that the literature does not contain.

The practical implication is that the trial result cannot be transferred format to format. A patient who takes a 10 mg edible for a migraine is not doing a smaller version of what the trial tested. They are doing something untested.

For a patient with migraine who wants to try cannabis as an acute treatment, the evidence supports a specific and limited version of that plan: an inhaled product containing both THC and CBD, used at the start of an attack, at the lowest dose that produces relief.

The frequency limit matters more than the dose. Any acute headache treatment used on more than roughly ten days per month puts a patient in the range where medication overuse headache becomes a concern, and there is no reason to assume cannabis is exempt. Tracking headache days and treatment days on a calendar is the cheapest and most useful thing a headache patient can do.

Cannabis is not a substitute for preventive treatment. A patient having frequent attacks needs a prevention conversation, and the trial evidence for established preventives, including CGRP-targeted agents, is far stronger than anything cannabis currently has.

Disclosure to the treating clinician is not optional. Cannabis interacts with a range of medications, is relevant to anesthesia and to psychiatric care, and changes how a headache history should be interpreted. A patient who conceals daily cannabis use makes their own diagnostic picture harder to read.

The migraine trial is strong evidence for a narrow question. Randomization, double blinding, placebo flower, a crossover design that uses each participant as their own control, and standard headache endpoints assessed at two hours are the right tools for testing an acute treatment. For inhaled THC plus CBD in acute migraine, this is credible evidence of benefit.

Everything else is weaker by design. App-based self-report cannot separate drug effect from expectation, and a chart review in a headache clinic cannot separate cannabis from the severity of the headache disorder that prompted it. Those studies describe patterns. They do not test treatments.

A single positive crossover trial in 92 participants is a starting point, not a settled question. Blinding is inherently difficult when the active drug is intoxicating and the placebo is not, and participants can frequently tell which arm they are in. The trial reported no serious adverse events, which is reassuring for safety and says nothing about whether unblinding inflated the effect.

The Strainprint analyses are often cited as though 50 percent symptom reduction were a trial result. They are not. Users selected their own products, logged voluntarily, and had every reason to record sessions that worked. The most reliable finding in that dataset may be the tolerance signal, because there is no obvious reporting bias that would create it.

The medication overuse association comes from a tertiary headache clinic, which enriches for the most refractory patients. The direction of causation is not established, and the authors did not claim that it was.

None of this evidence shows that cannabis prevents headaches, reduces monthly headache days over time, or treats tension-type or cluster headache. No controlled trial has tested any of those questions.

It does not show that oral, sublingual, or topical cannabis products help headache. The only positive controlled result used vaporized flower, and route is not a cosmetic detail in acute headache treatment.

It does not establish a dose. The trial used flower of specified potency, not a milligram dose, and it did not compare doses. Nor does it show how cannabis compares with a triptan, a gepant, or an NSAID, because no trial has run that comparison.

Headache medicine has spent the past decade absorbing genuinely new mechanisms, most visibly CGRP-targeted agents developed on the back of large randomized programs. Against that standard, cannabis for headache is an early-stage question with one positive trial, and it should be described that way rather than as an established option.

At the same time, a substantial number of patients are already treating headaches with cannabis and reporting that it helps. Their experience is data of a kind, and the appropriate response is neither dismissal nor endorsement. It is a structured conversation about which headache type, which format, how often, and what the calendar shows after three months.

For years I told headache patients the truth, which was that nobody had run the trial. Now one exists, it was positive for inhaled THC plus CBD in acute migraine, and I can say that. What I still cannot say is that cannabis prevents migraine, helps tension headaches, or belongs anywhere near a cluster headache protocol.

The part of this conversation I care most about is frequency. Headache medicine learned the hard way that treating attacks too often makes the disorder worse, and I see no argument for why cannabis would be the exception to a rule that applies to triptans, combination analgesics, and opioids. When a patient tells me cannabis is the only thing that touches their headaches and they are using it daily, that is not a success story. That is the pattern that needs unwinding, ideally with a headache specialist involved.

The most useful thing most headache patients can do is keep a calendar. Headache days and treatment days, three months, written down. It converts an argument about whether something works into a question that can be answered.

Inhaled cannabis containing both THC and CBD beat placebo for acute migraine in one well-designed crossover trial. CBD alone did not. No controlled evidence supports cannabis for headache prevention, tension-type headache, or cluster headache, and observational data link cannabis use in chronic migraine to a much higher burden of medication overuse headache. Treat it as an acute option for migraine with a strict frequency limit, not as a daily therapy.

Carry forward two things: the combination arm worked in acute migraine, and the CBD-only arm did not. Carry forward the frequency caution as firmly as the efficacy finding. Do not carry forward the 50 percent symptom reduction figure from app data as if it were a trial result, and do not extend the migraine finding to other headache types or to formats the trial did not test.