Cluster Headache Drugs: Medications That Cause or Treat It

Updated May 26, 2026

Image: HealthRX.com clinical image

Cluster headache is a primary headache disorder, not a symptom of another disease, and its drug list splits into two unrelated categories: medications that abort or prevent attacks, and medications that can provoke an attack in someone who already has the condition. The article below covers both, with generic names given first because "cluster headache treatment" is often searched alongside specific brand names (Emgality, Imitrex) that map onto different regimens than their more familiar uses.

A note on the evidence in this draft. The original source material for this page cited specific trial sample sizes, exact percentages, and journal references for many claims. Those identifiers could not be independently verified through primary-literature lookup for this revision. The clinical substance of the claims below (oxygen and injectable sumatriptan as first-line acute treatments, verapamil as first-line prevention, galcanezumab's FDA approval for episodic cluster headache, and vasodilators as attack triggers) reflects well-established clinical teaching in headache medicine. The specific numbers, study sizes, and citation links attached to individual trials should be confirmed by the clinical reviewer against the primary trial reports before this page is published.

The direct answer

For an acute cluster headache attack, high-flow oxygen and subcutaneous sumatriptan are the two first-line treatments, chosen because both act within the 15-to-180-minute attack window that makes oral drugs impractical. For prevention, verapamil is the first-line agent worldwide despite limited randomized-trial data, because decades of clinical use and guideline consensus support it; galcanezumab (Emgality) is the only drug FDA-approved specifically for prevention of episodic cluster headache, and it has not shown benefit in the chronic form. Certain vasodilating drugs, most notably nitroglycerin, sildenafil-class PDE5 inhibitors, and alcohol, can trigger an attack during an active cluster bout even though they do not cause the disorder itself.

What cluster headache is, briefly

Cluster headache belongs to the trigeminal autonomic cephalalgias, a category defined by the International Headache Society's ICHD-3 classification. Attacks cause severe, strictly one-sided pain around the eye or temple, accompanied on the same side by tearing, eyelid drooping, pupil constriction, nasal congestion, or facial sweating. Attacks last 15 to 180 minutes untreated and cluster in bouts lasting weeks to months, often with a striking daily and seasonal pattern.

Diagnosis is clinical, based on the pattern and accompanying autonomic features. A brain MRI is generally recommended once to exclude secondary causes such as a pituitary lesion or cavernous sinus abnormality, since cluster-like pain can occasionally be produced by structural disease. No blood test or imaging study confirms the diagnosis itself. Diagnostic delay is a recognized problem in this condition because the pattern is often mistaken for migraine, sinus disease, or dental pain.

Roughly 85 to 90 percent of patients have the episodic form, with bouts separated by remissions of three months or longer. The remaining patients have chronic cluster headache, defined as attacks persisting more than a year without a remission of that length. This distinction matters pharmacologically: the one FDA-approved preventive for cluster headache works in the episodic form and has not demonstrated benefit in the chronic form.

Stopping an attack in progress

Cluster attacks build in minutes and resolve within three hours, so an effective acute treatment has to act fast. Two options meet that bar with strong trial support.

High-flow oxygen, delivered at 12 to 15 liters per minute through a non-rebreather mask for 15 to 20 minutes, is considered established, first-line therapy with no dose ceiling and no systemic side effects. A randomized crossover trial found it substantially more effective than room air for producing pain relief. The main practical obstacle is logistics: patients need a home oxygen concentrator or cylinder capable of the required flow rate, and coverage for this equipment varies by insurer and by state.

Sumatriptan 6 mg by subcutaneous injection is the other first-line acute option, with trial data showing a majority of treated attacks reaching pain freedom within 15 minutes, compared with a much lower rate on placebo. The injectable route matters because oral triptans are absorbed too slowly for an attack that may already be resolving by the time an oral tablet takes effect. Sumatriptan nasal spray and zolmitriptan nasal spray are alternatives for patients who decline injections, with slower onset, generally judged effective around 30 minutes.

Oral analgesics, NSAIDs, and opioids have no established role in aborting a cluster attack; they act too slowly, and opioids carry a real risk of medication-overuse headache with frequent use. Ergotamine has historical use but is largely supplanted by triptans and oxygen.

Preventing attacks during a bout

Acute treatments stop individual attacks but do nothing to reduce how many attacks occur during a bout that can bring one to eight episodes a day for weeks. Preventive medication is used specifically to shorten and quiet the bout.

Verapamil is the calcium channel blocker used first-line worldwide for cluster headache prevention, endorsed by headache society guidelines despite a comparatively small base of randomized-trial evidence. Typical dosing ranges from 240 mg to 960 mg per day, titrated slowly upward. Baseline ECG monitoring is required before starting and with each dose increase, because verapamil at higher doses can prolong the PR interval and cause bradycardia or heart block. This is not a drug to titrate without cardiac monitoring, and patients with pre-existing conduction disease need cardiology input before starting.

Galcanezumab (Emgality) is a monoclonal antibody against CGRP and is, as of this writing, the only medication with an FDA-approved indication specifically for prevention of episodic cluster headache. Its cluster headache dosing regimen (a 300 mg loading dose given as three consecutive 100 mg injections at the start of a bout) is different from its migraine-prevention dosing, and this distinction should be specified clearly on any prescription to avoid pharmacy or insurance confusion. In a separate trial in chronic cluster headache, galcanezumab did not meet its primary efficacy endpoint, so its approved use does not extend to the chronic form. A related CGRP antibody, fremanezumab, likewise failed to meet its primary endpoint in a chronic cluster headache trial, a pattern that has led headache specialists to suspect that episodic and chronic cluster headache may respond differently to CGRP-targeted therapy, though the underlying reason is not established.

Lithium carbonate has decades of use in chronic cluster headache, based mainly on open-label experience rather than large placebo-controlled trials. It requires serum lithium level monitoring plus thyroid and kidney function surveillance, and carries tremor, polyuria, and hypothyroidism as recognized side effects.

Topiramate is generally reserved as a second- or third-line option when verapamil or lithium cannot be used or tolerated. Evidence is largely open-label. Cognitive slowing and weight loss are the side effects that most often limit its use.

Melatonin, typically dosed around 10 mg at night, has shown benefit in a small placebo-controlled trial and has an appealing safety profile, which makes it a reasonable low-risk adjunct, particularly given the circadian disruption believed to underlie cluster headache. It is not a substitute for verapamil or another primary preventive in an active bout.

Bridging the gap before prevention works

Verapamil and other preventives take two to three weeks to reach an effective dose. Bridge therapy covers that window.

Oral corticosteroids, commonly a short prednisone course tapered over two to three weeks, are widely used as a bridge, though attacks often recur as the taper proceeds. Headache society guidance recommends limiting steroid bridges to once or twice a year given the risks of adrenal suppression and bone loss with repeated courses.

Greater occipital nerve block, an injection of local anesthetic with or without corticosteroid near the occipital nerve on the affected side, has randomized-trial support as an effective bridge that can meaningfully reduce attack frequency for at least several weeks.

Inpatient intravenous dihydroergotamine is used for refractory bouts in a hospital setting. It is contraindicated within 24 hours of triptan use and in patients with coronary artery disease or uncontrolled hypertension, which limits its use to carefully selected patients.

Medications and substances that can trigger an attack

A distinctive feature of cluster headache is that certain drugs reliably provoke an attack during an active bout but typically do not do so during remission. This is useful clinically: a patient who develops a headache after a glass of wine during a bout, but tolerates the same drink during remission, is showing a textbook pattern.

Nitroglycerin and other nitric oxide donors are the best-documented pharmacological trigger; provocation studies have shown that sublingual nitroglycerin reliably induces a cluster-like attack in patients during an active bout and does not do so during remission. This observation is part of why the nitric oxide pathway is considered central to cluster headache biology. Patients who need nitrates for angina and who are in an active cluster bout may need cardiology input to consider alternative antianginal therapy; calcium channel blockers, already used as a cluster preventive, can sometimes serve double duty.

Alcohol is the most commonly reported trigger during active bouts, with even small amounts capable of provoking an attack within an hour. Patients in a bout are generally advised to avoid alcohol entirely; outside a bout, most patients tolerate alcohol without problems.

PDE5 inhibitors (sildenafil, tadalafil, vardenafil) have been reported to trigger cluster attacks in case reports, plausibly through the same nitric oxide/cGMP pathway as nitroglycerin. Patients with cluster headache who need a PDE5 inhibitor for erectile dysfunction should discuss timing relative to their bout with the prescribing clinician rather than assume it is safe by default.

Histamine was the original experimental trigger used in early cluster headache research. It is not a prescribed drug, but histamine-releasing exposures (certain contrast agents, rapid vancomycin infusion) are worth being aware of during an active bout.

Newer and investigational approaches

Non-invasive vagus nerve stimulation (the gammaCore device) has FDA clearance for acute cluster headache treatment and has shown benefit for episodic cluster headache attacks in sham-controlled research, without a clear benefit demonstrated for chronic cluster headache in the same research program.

Sphenopalatine ganglion stimulation via an implanted device has shown attack relief in patients with chronic cluster headache in early clinical studies, but it is an invasive procedure requiring implantation, available only at specialized centers, and reserved for patients who have not responded to standard preventive therapy.

Psilocybin and related tryptamines are the subject of active patient interest and emerging research, including survey data suggesting that some patients report attack abortion or bout interruption. Randomized controlled trial data are still limited, and these substances remain Schedule I in most jurisdictions. They should not be pursued outside a supervised research or clinical setting, and patients asking about this option should be directed toward legitimate clinical trials rather than self-directed use.

When standard treatment is not enough

An estimated meaningful minority of cluster headache patients are refractory to standard therapy. Refractory chronic cluster headache is generally defined by specialist consensus as failure of adequate trials of multiple preventive drugs, including verapamil at a high dose and lithium at a therapeutic serum level. These patients should be referred to a headache specialist center to consider occipital nerve stimulation, sphenopalatine ganglion procedures, or, in the most intractable cases, hypothalamic deep brain stimulation, as well as clinical trial enrollment.

A refractory course is also a reason to re-examine the diagnosis itself. Other trigeminal autonomic cephalalgias, particularly paroxysmal hemicrania and SUNCT/SUNA, can mimic cluster headache but require different treatment; a headache that responds dramatically to indomethacin is more likely paroxysmal hemicrania than cluster headache.

When to seek urgent evaluation

Standard cluster headache management assumes the diagnosis is secure. Urgent evaluation is warranted, not routine follow-up, if any of the following occur: a new pattern of headache after age 50 with no prior history, new neurological symptoms such as weakness, vision loss, or confusion, a headache that becomes continuous without remission, or a sudden loss of response to treatments (oxygen, sumatriptan) that previously worked well. These features raise concern for a secondary cause rather than primary cluster headache and typically call for repeat or more detailed imaging.

What is established, what is plausible, and what is not established

Oxygen and injectable sumatriptan for acute attacks, and verapamil for prevention, rest on the strongest and most consistent evidence base in this condition, reinforced by long clinical experience even where trial sample sizes are modest. Galcanezumab's FDA approval for episodic cluster headache prevention is a regulatory fact, not just an opinion, but its lack of benefit in chronic cluster headache is also established and should not be glossed over. Lithium, topiramate, and melatonin are plausible options with far less rigorous trial support, useful mainly when first-line therapy fails or is contraindicated. Neuromodulation devices and psilocybin-based approaches are investigational for this condition; they may help some patients, but claims of comparable effectiveness to standard therapy are not established and should not be presented to patients as settled.

Cluster headache drug decision framework

This framework maps the phase a patient is in to the decisions that actually change management. It is meant to organize thinking, not to replace individualized dosing decisions made with a treating clinician.

Situation	What to check first	Reasonable first move	Watch for / escalate if
Attack in progress, no treatment yet	Confirm strictly one-sided pain with autonomic features (tearing, congestion, ptosis) lasting under 3 hours	High-flow oxygen and/or sumatriptan injection if prescribed; avoid oral analgesics as the primary strategy	Pain lasting well beyond 3 hours, bilateral pain, or new neurological signs, reconsider diagnosis
New bout starting, not yet on a preventive	Baseline ECG before starting verapamil; review cardiac history	Bridge therapy (short steroid course or occipital nerve block) started same day as verapamil titration begins	Bradycardia, PR prolongation, or heart block on ECG, do not push verapamil dose further without cardiology input
Already on verapamil, attacks continuing after 2-3 weeks at target dose	Confirm dose adequacy and adherence; repeat ECG at higher doses	Discuss adding or switching to galcanezumab (episodic) or lithium (with level monitoring)	Chronic cluster headache with no response to galcanezumab, do not expect the same benefit seen in episodic disease
On a preventive, considering an unrelated new prescription	Ask specifically about nitrates, PDE5 inhibitors, and alcohol use before prescribing for an unrelated condition	Time PDE5 inhibitor or nitrate use around known bout-free periods where possible; involve cardiology if nitrates are needed for angina during a bout	New attacks starting shortly after starting a vasodilator during a known bout, treat as a probable trigger, not a coincidence
Failing 3+ adequate preventive trials	Confirm ICHD-3 criteria are still met; re-image if the pattern has changed	Refer to a headache specialist center for neuromodulation evaluation or trial enrollment	Diagnosis uncertainty (indomethacin responsiveness, atypical duration), reconsider paroxysmal hemicrania or SUNCT/SUNA before escalating cluster headache therapy further

Frequently asked questions

What causes cluster headache?

The exact cause is not fully established. PET imaging has consistently shown activation of the hypothalamus on the side of the pain during attacks, which is thought to explain the strong daily and seasonal rhythm of bouts. CGRP release from trigeminal nerve fibers and nitric oxide signaling are both implicated. During an active bout, specific triggers such as alcohol, nitroglycerin, and certain other vasodilators can provoke individual attacks.

How is cluster headache diagnosed?

Diagnosis is clinical, using ICHD-3 criteria: repeated attacks of severe one-sided orbital, supraorbital, or temporal pain lasting 15 to 180 minutes, with at least one same-sided autonomic symptom such as tearing, nasal congestion, ptosis, or restlessness. Brain MRI is generally recommended once to rule out a secondary cause. No blood test or scan confirms the diagnosis on its own.

When should I worry about a cluster-type headache?

Seek prompt evaluation if the pattern changes, if new neurological symptoms appear (weakness, vision loss, speech trouble, confusion), if this is a first episode after age 50, if attacks become continuous without remission, or if treatments that used to work stop working. These can point to a secondary cause that needs urgent imaging.

Is cluster headache the same as migraine?

No. Cluster attacks are much shorter than migraine attacks, strictly one-sided, and accompanied by autonomic features like tearing or nasal congestion rather than the nausea and light sensitivity typical of migraine. People with cluster headache are often restless and pace during an attack, while migraine typically drives someone to lie still in a dark room.

Can cluster headache be cured?

There is no cure. Episodic cluster headache typically remits between bouts, sometimes for years, and some patients eventually stop having bouts altogether with age. Preventive medication reduces how often attacks occur during a bout, and acute treatment can abort individual attacks. A minority of patients develop the chronic form, with attacks persisting for more than a year without a long remission.

Can Emgality (galcanezumab) treat cluster headache?

Galcanezumab is FDA-approved specifically for prevention of episodic cluster headache, using a different dosing regimen than its migraine-prevention use. It has not shown benefit in a separate trial in chronic cluster headache, so its approved use does not extend to that form. Prescriptions should clearly specify the cluster headache regimen to avoid confusion with the migraine dose.

What medications should I avoid during a cluster bout?

During an active bout, most headache specialists advise avoiding nitroglycerin and related nitrates, PDE5 inhibitors (sildenafil, tadalafil, vardenafil), and alcohol, since these can provoke an attack within minutes to hours. If nitrates are needed for a heart condition during a bout, that decision should involve a cardiologist rather than being managed alone.

What is the difference between episodic and chronic cluster headache?

Episodic cluster headache involves bouts lasting roughly a week to a year, separated by remissions of three months or more. Chronic cluster headache means attacks continue for more than a year without a remission of that length. Most patients have the episodic form, and some patients shift between the two categories over time.

Sources referenced in this topic area

The clinical statements above draw on the published headache literature on cluster headache diagnosis, oxygen and triptan therapy for acute attacks, verapamil and galcanezumab for prevention, and case reports on vasodilator triggers. Specific trial identifiers, sample sizes, and journal citations from the prior version of this page could not be independently verified during this revision and should be checked against the primary literature (for example, via PubMed) by the clinical reviewer before publication. Readers and clinicians who want current guideline-level detail should consult the American Headache Society's published treatment guidance and the ICHD-3 criteria maintained by the International Headache Society directly.

For background on how cluster headache presents and is worked up, see the related cluster headache overview.