brand logo

Cluster headache, an excruciating, unilateral headache usually accompanied by conjunctival injection and lacrimation, can occur episodically or chronically, and can be difficult to treat. Existing effective treatments may be underused because of underdiagnosis of the syndrome. Oxygen and sumatriptan have been demonstrated to be effective in the acute treatment of cluster headaches. Verapamil has been shown to be effective for prophylaxis. For cluster headache completely refractory to all treatments, surgical modalities and newer interventions such as the implantation of stereotactic electrodes may be useful. Patients should be encouraged to avoid possible triggers such as smoking or alcohol consumption, especially during the cluster period. The intensity of cluster headache pain leads to ethical concerns among researchers over the use of placebo, making randomized controlled trials difficult. As new technology and genetic studies clarify the etiology of cluster headache, it is possible that more specific therapies will emerge.

The diagnosis and optimal management of cluster headache remain challenging. This most painful of primary headaches affects 0.1 percent of adults.1 The male-to-female ratio has diminished from 6.2 to 1 in the 1960s to 2.1 to 1 in the 1990s.2 Men may first experience cluster headache in their early 20s, with peak onset in their 40s. In one study,3 the most frequent age at onset for women was in their 60s. Cluster headache may be underdiagnosed in black women,4 but ethnic differences in prevalence have not been studied. Having a family history of headaches, smoking, head injury, or shift work has been associated with cluster headache.

Key clinical recommendationLabelReferences
The first-line treatments for acute cluster headache are oxygen or sumatriptan, or a combination of the two.A1518
Less well studied alternatives for acute treatment include intranasal dihydroergotamine, intranasal lidocaine, and intranasal capsaicin.B1921
Verapamil, in a dosage of 360 to 480 mg daily, can effectively reduce the number of attacks during a cluster headache period.A24,26
Less well studied alternatives for prophylaxis include prednisone and antiepileptic drugs; they should only be considered if verapamil is not tolerated or not effective.B26,32,28

Clinical Features and Classification

In 2004, the International Headache Society published new criteria for diagnosing cluster headache. To fulfill criteria for diagnosis, patients must have had at least five attacks occurring from one every other day to eight per day; and attributable to no other disorder.5 In addition, headaches must cause severe or very severe unilateral orbital, supraorbital, or temporal pain lasting 15 to 180 minutes if untreated, and be accompanied by one or more of the following: ipsilateral conjunctival injection or lacrimation, ipsilateral nasal congestion or rhinorrhea, ipsilateral eyelid edema, ipsilateral forehead and facial sweating, ipsilateral miosis or ptosis, or a sense of restlessness or agitation.5 Episodic cluster headache is defined as at least two cluster periods lasting seven to 365 days and separated by pain-free remission periods of one month or longer. Chronic attacks recur over more than one year without remission or with remission lasting less than one month.5

Called “suicide headache” because of its severity and “alarm clock” headache because of its periodicity, cluster headache is characterized by unilateral excruciating pain (a hot-poker or stabbing sensation) in the ocular, frontal, or temporal areas. Pain often radiates to the upper teeth, jaw, and neck. Associated signs include ptosis, ipsilateral lacrimation, conjunctival injection, and rhinorrhea. The pain usually is unilateral, with 60 percent of patients reporting headaches on the right side, but 14 percent of patients report a side shift during an attack, and 18 percent report involvement of different sides in subsequent attacks.6

Other symptoms include facial flushing or pallor, tenderness on palpation of the ipsilateral carotid artery, bradycardia, and abnormal feeling of scalp hairs. The absence of aura, nausea, or vomiting has helped distinguish cluster from migraine headaches, but recent studies indicate that 14 percent of patients with cluster headache experience aura, 51 percent have a personal or family history of migraine, 56 percent report photophobia, 43 percent report phonophobia, and 23 percent report osmophobia.7 Thus, the presence of aura, nausea, vomiting, or photophobia should not rule out a diagnosis of cluster headache.8 A characteristic feature of cluster headache, noted by 93 percent of patients in one study,7 is restlessness, with behaviors such as pacing and rocking the head and trunk with head in hands.9 Most of these headaches last 15 minutes to three hours and recur at the same time of day, often at night. Many attacks begin during the first rapid-eye-movement sleep phase. Patients may report a seasonal pattern, with spring and autumn peaks.

Cluster headache is diagnosed by history, and the key feature is a pattern of recurrent bouts of near-daily attacks lasting for days, weeks, or months. Patients fearing an attack may be afraid to go to sleep. Precipitants of cluster headache include hypoxia, which may occur with sleep apnea. Vasodilators such as nitroglycerin, alcohol, and carbon dioxide may trigger a headache during a cluster period.10

Although similar to cluster headache, paroxysmal hemicrania headaches are briefer and are treated effectively with indomethacin.5 Orbital myositis may mimic cluster headache, but the headache has a longer duration.


Positron emission tomographic (PET) scanning and functional magnetic resonance imaging are helping to clarify the poorly understood etiology of cluster headache. The basic pathophysiology is in the hypothalamic gray matter.11 In some families, an autosomal dominant gene may be involved, but calcium channel activity or nitric oxide sensitivity alleles have not been identified.12 Carotid and ophthalmic artery vasodilation and an increased sensitivity to vasodilator stimuli during an attack may be triggered by trigeminal parasympathetic reflexes. Abnormal heart rate variability and increased nocturnal lipolysis during attacks and in remission reinforce the theory of an autonomic function abnormality with increased parasympathetic drive and decreased sympathetic function. Attacks often begin during sleep, implicating a disorder of circadian rhythm.13 An increased incidence of sleep apnea in patients with cluster headache suggests that periods of reduced oxygenation of key tissues may trigger an attack.14


Cluster headache treatment requires a dual strategy. Acutely, the attack must be aborted or subdued. Concurrently, prophylaxis is initiated to suppress the recurrent headaches expected throughout the remaining cluster period; prophylaxis is continued for the expected duration of the cluster period and then tapered off. Patients with chronic cluster headache require long-term prophylaxis. In patients with intractable headaches, more aggressive intervention, including surgery, may be required.

DrugDosage and routeReported adverse effectsEfficacyComments
Oxygen7 L per minute for 15 minutes via face maskNoneIn a double-blind crossover study15 of 19 patients, 56 percent reported complete or substantial relief compared with 7 percent in the placebo group.Treatment of choice
Sumatriptan6 mg subcutaneously; may repeat in 24 hoursLocal skin reactions, fatigue, nausea, vomiting, dizziness, chest symptoms, throat symptoms, burning sensations, paresthesiasIn a double-blind RCT16 of 49 patients, 46 percent were free of pain within 15 minutes compared with 10 percent in the placebo group.
In an open study17 of 138 patients, 96 percent of attacks were relieved at 15 minutes.
Contraindicated in patients with coronary artery disease, uncontrolled hypertension, or angina
20-mg nasal sprayNoneIn a double-blind RCT18 of 118 patients, there was a 57 percent response compared with a response of 26 percent with placebo. Pain-free rates at 30 minutes were 47 percent compared with 18 percent with placebo.
Intranasal dihydroergotamine0.5-mg nasal spray bilaterallyNoneIn a double-blind trial19 of 25 patients, the severity of attacks was reduced. There was no effect on frequency or duration of attacks.Intravenous and intramuscular routes have been used, but their efficacy has not been substantiated in controlled clinical trials.
Intranasal lidocaine1 mL of 10 percent lidocaine placed with a cotton swab bilaterally for 5 minutesUnpleasant tasteIn a double-blind, placebo-controlled trial20 of 15 patients, there was a decrease in pain after 5 minutes, and nine of nine treated patients were free of pain at 35 minutes.Place as close to sphenopalatine fossa as possible.
Intranasal capsaicinPlace via cotton swab in ipsilateral nostril twice a day for seven daysBurning sensation (decreases after five applications)In a double-blind, placebo-controlled trial21 of 15 patients, there was a reduction of headache severity in the treatment group at eight to 15 days.There were more patients with episodic cluster headache in the treatment group and more patients with chronic headache in the control group.


The treatments of choice for acute cluster headache are oxygen (7 L per minute for 15 minutes), sumatriptan, or a combination of the two (Table 1).1521 Both therapies appear to be underprescribed. Sumatriptan is available as a subcutaneous injection or as a nasal spray.1618 It is contraindicated in patients with ischemic heart disease, uncontrolled hypertension, or peripheral vascular disease, and it should not be combined with ergotamine. Tachyphylaxis does not occur when the drug is used for long periods. Higher sumatriptan doses (12 mg subcutaneously) were not found to be more effective and had more side effects than 6-mg doses.18

Intranasal dihydroergotamine can reduce attack severity, but it does not reduce attack frequency or duration. This nasal spray has fewer adverse reactions and better bioavailability than oral ergotamine.19 Intranasal lidocaine has been shown to decrease pain after five minutes, with nine of nine treated patients free of pain after 35 minutes.20 Intranasal capsaicin, which is hypothesized to deplete substance P from sensory nerve terminals, is used in treatment of other pain states, including herpes zoster. In a study of patients with cluster headache, it reduced the severity of headaches after seven days of treatment.21

The effectiveness of melatonin is unclear because of conflicting studies.22 Its role, if any, will be in the initial prevention of attacks, theoretically by resetting the circadian rhythm.


Oral sumatriptan has not been shown to be effective prophylactically.23 Verapamil in dosages of 360 to 480 mg daily is one of the few treatments for episodic cluster headache tested in a randomized controlled trial (RCT) and found effective in reducing attack frequency (Table 2).13,2431 This treatment also is underused, with only 4 percent of patients with cluster headache reporting prophylactic verapamil use.

DrugDosage and routeAdverse effectsEfficacyComments
Verapamil120 to 160 mg orally three times dailyHypotension, bradycardia, atrioventricular block, dizziness, fatigue, nausea, constipationIn a double-blind RCT24 of 30 patients, 12 of 15 patients had a reduction in frequency of attacks within two weeks.
In a double-blind study25 of 30 patients who received verapamil versus lithium, 50 percent of patients improved during first week of treatment.
Prednisone50 to 80 mg orally daily tapered over 10 to 12 daysIncreased appetite, insomnia, nervousness, hyperglycemia, dizziness, headacheIn a double-blind study,26 17 of 19 patients experienced improvement in pain, and the treatment group had a lower frequency of headaches.
In a retrospective study,27 14 of 19 patients had more than 50 percent headache relief.
In trial,26 improvement occurred within two days.
Recurrences often occurred toward the end of the taper.27
Should take concurrently with another prophylactic medication.
Divalproex600 to 2,000 mg dailyNausea, somnolence, dizziness, insomnia, anorexia, weakness, thrombocytopenia, alopecia, weight gainIn an open-label study,28 nine of 15 patients experienced complete disappearance of pain, and two others markedly improved.Use caution in patients with renal or hepatic insufficiency.
In another open-label study28 of 26 patients, headache frequency decreased by 54 to 59 percent.*
No RCTs.
Topiramate25 mg orally daily for seven days, then increase dose by 25 mg daily every week to a maximum dosage of 200 mg dailyParesthesias, cognitive effects, drowsiness, dizzinessIn a retrospective chart review29 of patients with migraine and cluster headache, nine of 12 patients with cluster headaches exhibited moderate to substantial improvement. No RCTs.
Ergotamine2 to 4 mg daily in divided dosesVertigo, pruritus, nausea, paresthesias, weakness, cardiac valvular fibrosis, retroperitoneal or pleuropulmonary fibrosis, angina, myocardial infarction; may cause withdrawal symptoms if suddenly discontinuedNo RCTs. Anecdotal evidence by experienced neurologist suggests effectiveness.13 Best for nocturnal attacks; contraindicated in peripheral vascular disease, hypertension, and cardiac disease; caution with renal or hepatic insufficiency; should not take concurrently with sumatriptan
Methylergonovine maleate0.2 mg orally three or four times dailyHypertension, nausea, vomiting, diarrhea, leg cramps, dyspnea, dizziness, tinnitus, nasal congestion, diaphoresis, palpitations, thrombophlebitis, hematuria, water intoxication, abdominal cramping, weight gain, paresthesias, amenorrhea; hallucinations in high dosesIn a retrospective cross-sectional study,30 19 of 20 patients reported a decrease of more than 50 percent in headache frequency; 15 of 20 patients reported a reduction in headache intensity.Caution in patients with peripheral vascular disease, cardiac disease, or renal or hepatic insufficiency; only for use in refractory cases; should never be used continually for longer than six months; contraindicated in pregnant and hypertensive patients; possibility of retroperitoneal, cardiac, and pleuropulmonary fibrosis
Melatonin10 mg orally at bedtimeNone reportedIn a double-blind, placebo-controlled study31 of 20 patients, headache frequency was reduced in five of 10 patients in the treatment group. A more recent pilot study19 of nine patients did not show any difference in response between the treatment and placebo groups.

Prednisone often is used in prophylaxis, starting at a dosage of 50 to 80 mg daily and tapered over 10 to 12 days. A 1975 study that used a double-blind control methodology led to improvement in 17 of 19 patients compared with placebo, but the study quality was limited.26 A recent nonrandomized study of cluster headache prophylaxis that used three days of 250 mg of intravenous methylprednisolone followed by a prednisone taper indicated fewer headaches in the active phase than with previous treatments in a group of 14 men.32 In small studies, the anti-epileptic drugs divalproex and topiramate were found to be useful.28 Gabapentin and baclofen have been tried with some success in case reports and nonrandomized trials.28 However, no RCTs have been conducted to determine conclusively the effectiveness of antiepileptic drugs.

Historically, ergotamine in a dosage of 2 to 4 mg per day has been a common agent for episodic prophylaxis, but no RCT of oral ergotamine has been reported. Ergotamine and sumatriptan should not be taken concurrently. In addition, patients with peripheral vascular disease should not take ergotamine. A recent review summarizes management guidelines for patients who continue to use ergotamine or who cannot tolerate verapamil or other prophylactic medications. Patients who experience attacks at night may be given 1- or 2-mg tablets orally or by suppository before bedtime.13 When headaches start consistently at the same time each day, patients may be advised to take ergotamine 30 to 60 minutes before the time of their usual attack. A combination of verapamil and ergotamine may be considered.13

Methysergide previously was used for cluster headache prophylaxis, but its use has been discontinued in the United States because it can cause retroperitoneal, cardiac, and pleuropulmonary fibrosis. Methylergonovine maleate should be restricted to use in refractory cases and should never be used for more than six months continuously. Methylergonovine, probably the active metabolite of methysergide, decreased cluster headache frequency by more than 50 percent in 19 of 20 patients in a pilot study.32 Contraindications include hypertension and pregnancy. Methylergonovine should be used with caution, if at all, in patients with cardiac or peripheral vascular disease, or hepatic or renal insufficiency. Hallucinations can occur because it is a lysergic acid derivative.30


Verapamil is used in the treatment of chronic cluster hadache.24 Lithium has been used as long-term prophylaxis for years on the basis of case series demonstrating effectiveness (Table 3).25,3335 In an RCT of lithium in episodic cluster headache, substantial improvement occurred in both groups: six (43 percent) of 14 patients in the placebo group and eight (62 percent) of 13 in the lithium group. The study, which was stopped because superiority over placebo could not be demonstrated, challenges physicians to look at the positive treatment effect of placebo.34

DrugDosage and routeAdverse effects or complicationsEfficacyGenericcost*Comments
Verapamil120 mg orally three times dailyHypotension, bradycardia, atrioventricular block, dizziness, fatigue, nausea, constipationIn an RCT of 30 patients, 12 of 15 patients had a reduction in frequency of attacks within two weeks.25 $20 to 46 per month
LithiumStart at 300 mg orally three times daily; use blood levels to achieve therapeutic dose.Confusion, dizziness, blurry vision, diabetes insipidus, headache, nausea, polyuriaIn a double-blind crossover study33 of 30 patients, 50 percent of patients responded in two weeks.
In a more recent RCT,34 using 800-mg extended-release tablets daily showed no benefit over placebo.
$11 to 16 per monthNeed close monitoring of lithium levels; test 12 hours after last dose; side effects include tremor and dysuria; check thyroid and renal function before and during treatment.
Microvascular decompressionNAInfection, cerebrospinal fluid leak, postoperative headache requiring lumbar punctureIn a trial35 of 28 patients, 46 percent had success at long-term follow-up.NAOnly used for intractable cases


Microvascular decompression of the fifth cranial nerve with or without section of the nervus intermedius was performed in 30 patients.35 Initially, 77.3 percent of patients experienced relief, but after five years, only 46.6 percent reported continued excellent or good outcome. Repeat surgeries did not add benefit. It is important to choose a surgical intervention that maintains nerve supply so that facial numbness and corneal anesthesia are less likely to occur.


Patients with severe snoring and cluster headache should be evaluated for sleep apnea. In some patients diagnosed with sleep apnea, treatment for apnea also proved to be effective for the headaches.14

Botulinum toxins A and B are being studied as treatments for various headaches, with evidence of some efficacy and mixed reports as to side effects. An ongoing study is addressing their role in chronic cluster headache.

A recent report35 described a patient with intractable cluster headache; PET scanning revealed activation of the posterior inferior hypothalamic gray matter during attacks. A stereotactic electrode was implanted in this area, with a permanent generator placed in a subclavicular pocket. When stimulation was provided at a frequency of 180 Hz, the attacks disappeared after 48 hours. Twice, the stimulator was turned off without the patient’s knowledge, and attacks returned within 48 hours. The latter attacks disappeared 48 hours after the generator was restarted. Thirteen months later, the pain had not recurred.

Behavior and Lifestyle Interventions

Although little research has assessed the effectiveness of psychologic and behavior strategies on cluster headache, these approaches have been supported in the management of migraine and other chronic headaches. Strategies that have proved effective for other headaches target pain intensity and resultant disability, analgesic overuse, adherence to prescribed regimens, and psychiatric comorbidity. The difference between a functional life with pain and a life of significant disability often is related to the patient’s ability to pace activities and regulate emotions (thereby decreasing the fear of pain and subsequent autonomic arousal), engage in healthy behaviors, and function despite the presence of pain.

To maximize the quality of life in patients with cluster headache, strategies including relaxation, biofeedback, smoking cessation, and alcohol intake reduction should be considered. The prevalence of cigarette smoking is higher in patients with cluster headache than in control patients. In one retrospective study,7 however, patients with cluster headache who had stopped smoking did not notice any change in their symptom pattern. Heavy alcohol use is more common in patients who progress to chronic cluster headache compared with those who have episodic headache, although it is unknown whether alcohol is a cause or result of living with chronic pain.

Relaxation training paired with thermal biofeedback is effective in tension and migraine headache treatment. Biofeedback may be effective in patients with post-traumatic headache.36 For patients with cluster headache and a history of head trauma or migraine, these options may be considered. If cluster headache is an autonomic disorder, biofeedback may be valuable. Cognitive behavior interventions with proven value in migraine treatment38 that may be considered in patients with cluster headache include discussing reasonable expectations, keeping simple behavior diaries, and emphasizing personal responsibility. Patients are taught self-assessment and self-regulation skills that help them challenge irrational beliefs such as all-or-nothing thinking, catastrophizing outcomes, or attributional styles that lead them toward ineffective strategies (e.g., “I have no control over the pain”). Providing patients with ongoing compassion, knowledge about cluster headache, coping skills, and a long-term trust relationship is crucial in helping them cope with this difficult and painful condition.

Continue Reading

More in AFP

More in PubMed

Copyright © 2005 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See permissions for copyright questions and/or permission requests.