Why This Question Matters

Migraine is one of the most common reasons people use medical cannabis. In one survey of over 1,400 medical cannabis users, about a third said they used it for headache or migraine, and most of them inhaled it. Clinicians get asked about this all the time, and until recently the answer was that nobody had run a proper randomized trial. Advice was built on surveys, chart reviews, and animal studies.

That changed with a placebo-controlled trial published in Headache, and a new systematic review in Cannabis and Cannabinoid Research has now put that trial alongside every other double-blind randomized trial of cannabinoids for migraine, medication overuse headache, and low back pain. The review graded the certainty of each finding, which is what makes it useful. It tells you which claims are supported and which are still guesses.

What the Review Looked At

The authors searched electronic databases, trial registries, and other sources for double-blind randomized clinical trials in adults. They accepted isolated cannabinoids, synthetic cannabinoids, vaporized products, and standardized plant extracts. The review was registered in PROSPERO, and each included trial was assessed with the Cochrane Risk of Bias 2 tool. Because the trials were too different to pool, the authors synthesized them narratively and rated certainty using GRADE.

Five trials is a small evidence base, and only one of them is a migraine trial. That is worth keeping in mind through everything that follows. The migraine conclusion is only as strong as that one study, and the review's moderate-certainty rating reflects both the quality of that trial and the fact that it has not been replicated.

What each trial tested and how the review rated the certainty of its findings.

The Migraine Trial in Detail

The migraine trial was run at the University of California San Diego and published in Headache. It was randomized, double-blind, placebo-controlled, and used a crossover design, meaning each participant treated up to four separate migraine attacks, one with each of four treatments in a random order, with at least a week between attacks. The four treatments were vaporized cannabis flower supplied by the National Institute on Drug Abuse: 6% THC (THC-dominant), 11% CBD (CBD-dominant), 6% THC plus 11% CBD, and a placebo flower with the THC and CBD chemically removed.

Ninety-two adults with migraine were enrolled, 83% of them women, with a median age of 41. About 30% had chronic migraine. A third had never used cannabis. Participants treated attacks within 4 hours of onset when pain was moderate or severe, using a handheld vaporizer and a standardized four-puff procedure guided by a smartphone app. They agreed not to use any other acute treatment before or within 2 hours after vaporizing. In total, 247 attacks were treated.

The primary outcome was pain relief at 2 hours. The authors chose that rather than the more usual pain freedom because prior survey data only supported relief, and because they worried that THC's known anti-nausea effect could make a symptom-freedom outcome look positive for the wrong reason.

All three main comparisons were statistically significant, and the results held up in the intention-to-treat analysis (which counts missing data as treatment failure) and in a sensitivity analysis restricted to surveys completed within the intended time window. Response rates were similar in people with episodic and chronic migraine.

The pattern on nausea deserves a note. THC plus CBD cleared light and sound sensitivity but did not beat placebo on nausea, partly because nausea resolved at high rates in both groups (76% with THC plus CBD, 71% with placebo). The authors point to this as evidence that the symptom-freedom result reflects an effect on the migraine attack itself rather than THC suppressing nausea.

THC, CBD, or Both

The 2 by 2 design of the trial lets you separate the two cannabinoids, and the separation is informative.

Modified intention-to-treat rates from the migraine trial. Placebo rates: pain relief 47%, pain freedom 16%, most bothersome symptom freedom 34%.

THC on its own got people to mild pain about as often as the combination did, but it did not get them to pain-free or symptom-free, and its benefit did not last. CBD on its own did nothing detectable at 2 hours, though at 1 hour all three active treatments beat placebo on pain relief. The combination was the only treatment that hit every endpoint and held through 24 and 48 hours.

The authors' proposed explanation is pharmacological. CBD acts as a negative allosteric modulator at the CB1 receptor, which blunts THC's psychoactive effects. In the trial, that showed up as lower reported highness, euphoria, and cognitive impairment with the combination than with THC alone, at similar or better efficacy. Whether CBD also adds a direct antimigraine effect at this ratio is not something the trial can answer.

Side Effects and Blinding

There were no serious adverse events with any of the four treatments. The side effects that did occur were the ones you would expect, and they tracked with THC content.

Percentages of treated attacks in which participants reported each effect.

Free-text reports included sedation, slowed thinking, throat irritation, dry mouth, dizziness, tingling, and increased appetite. Most had faded by 2 hours.

Blinding is the obvious weak point in any cannabis trial, since a participant who feels high has a good idea what they received. The investigators tried to limit this by using low potencies, by telling participants they might feel a placebo high from CBD or placebo flower, and by measuring blinding formally with Bang's blinding index. Across 16 treatment cohorts, only one showed significant correct guessing (placebo given as the fourth attack) and one showed significant incorrect guessing. When the authors added highness and euphoria to the statistical model as covariates, the treatment effects did not change. The review still flagged possible functional unblinding as a limitation, which is fair. The blinding index was calculated on small groups, and a modest tendency toward correct guessing was present for the combination given as the fourth treatment.

Nabilone and Other Products

The only other headache trial in the review tested nabilone, an FDA-approved synthetic cannabinoid pill, in medication overuse headache. It showed a favorable signal for pain outcomes and for reducing analgesic consumption, but the reviewers rated the evidence very uncertain because of the trial's size and design. A separate small nabilone trial in chronic musculoskeletal pain landed in the same category. Neither supports prescribing nabilone for headache on the current evidence.

The low back pain trials are outside the scope of this site, but they matter for the review's overall message. A single 400 mg oral dose of isolated CBD did nothing for acute low back pain. A standardized full-spectrum extract called VER-01 improved chronic low back pain in a phase III trial, with moderate certainty. Put those next to the migraine trial and you get the review's central conclusion: the results depend on which cannabinoid, in what ratio, by which route, for which condition. A positive result for vaporized THC plus CBD in acute migraine says nothing about a CBD gummy for prevention, and a negative result for oral CBD in back pain says nothing about inhaled cannabis for an attack.

What We Still Don't Know

The trial was designed to answer one question: does a single dose of vaporized cannabis treat a migraine attack better than placebo? It answers that question for one product at one potency and ratio. It leaves a lot open.

• Repeated use. Participants treated at most four attacks, a week or more apart. There is no data on what happens with regular use over months, including whether cannabis can drive medication overuse headache the way frequent triptan or analgesic use does, or whether it has any preventive effect.
• Cannabis hyperemesis syndrome. Chronic heavy cannabis use can cause cyclic vomiting, which in someone with migraine can be hard to distinguish from the migraine itself. The trial could not assess this risk.
• Cannabis use disorder and psychiatric effects. People with a history of substance use disorder, bipolar disorder, schizophrenia, psychosis, or moderate to severe depression were excluded. Results do not apply to those groups.
• Real-world products. The study flower came from the NIDA supply program at 6% THC with no terpenes and under 1% minor cannabinoids. Dispensary products are usually far more potent and chemically different. Nobody has tested whether they perform the same way.
• Other routes. Edibles, tinctures, and oral capsules have different onset times and pharmacokinetics. The trial only tested inhalation.
• Pregnancy and breastfeeding. Both were exclusion criteria. Cannabis is not recommended in either.
• Driving and work. A third of attacks treated with THC alone and a fifth treated with the combination involved reported cognitive impairment at 1 hour. That is a practical limit on when the treatment can be used.

The authors of both the trial and the review call for multicenter randomized trials and for long-term studies of repeated use. Until those exist, the evidence covers a single treated attack under controlled conditions.

Key Takeaways

For people with migraine who already use cannabis, or are thinking about it, the trial gives a concrete reference point: a low-potency product with more CBD than THC, inhaled early in an attack, produced pain freedom in about a third of attacks at 2 hours, compared with about one in six on placebo. That is in the same range as some established acute treatments, though no head-to-head trial exists and cross-trial comparisons are unreliable.

For clinicians, the review's main value is in what it rules out. Cannabinoids are not a class with a single answer. The formulation, ratio, route, and indication all change the result, and the strongest migraine evidence is for one specific combination that most patients cannot buy. Legal status varies by state and country, and this is a conversation to have with your own clinician, taking into account your other medications, your mental health history, and how often you would be using it.