Subscription is FREE for qualified healthcare professionals in the US.
11 Articles in Volume 13, Issue #6
Ask the Expert: Cash Patient on High-Dose Oxycodone With Negative Urine Screens
Cluster Headache: Providing Relief for a Debilitating Disorder
Editor's Memo: Keeping the Trust in Difficult Times
Gout: New Guidelines for Managing An Ancient Disease
History of Pain: A Brief Overview of the 17th and 18th Centuries
Letters to The Editor: Guidelines for Opioid Prescribing, Drug Legislation
Long-term Opioids, Sickle Cell Disease, and Pain Patches
Lumbar Spinal Stenosis: A Review of the Treatment Options and Modalities
Malabsorption of Opioid Medications
Non-Opioid Pharmaceutical Treatment of Cancer Pain
Treatment of Postherpetic Neuralgia With Low Level Laser Therapy

Cluster Headache: Providing Relief for a Debilitating Disorder

Abortive and prophylactic therapies work well
to stem severity and frequency of attacks.

Headache continues to be the most common pain complaint of patients presenting to general practitioners, internal medicine specialists, and neurologists. Although the various primary headache types (vascular: migraine and cluster; nonvascular: tension-type headache) may appear alone or in various combinations in the headache patient, cluster headache (CH) is felt by many experts to be the most debilitating. CH sufferers frequently are misdiagnosed and subject to inappropriate treatment. Because this disorder often is associated with self-directed violence, even suicide, it is imperative to diagnose and treat the problem quickly and efficiently. This article will review the incidence, diagnosis, and treatment of CH from a pain management viewpoint.

Prevalence of CH

CH is considered the most common type of trigemino-autonomic headache, with estimates ranging from as low as 5 to 69 patients per 100,000 people to as high as 240 to 900 patients per 100,000 people—or between 0.5 and 2 million Americans.1 Despite its prevalence, patients with CH face difficulty in receiving a correct diagnosis. In one study, the average delay from the onset of the CH to diagnosis was 6.6 years, with patients seeing an average of 4.3 physicians and receiving 3.9 incorrect diagnoses prior to receiving the correct diagnosis.2

CH primarily has been a disorder of males, with various studies showing a predominance in males of between 84% and 88%, but change in the gender ratio of CH has been noted over the last several decades.3 When the incidence was analyzed in association with significant lifestyle changes, including employment rate, level of school education, smoking, and coffee and alcohol intake, the male to female ratio fell from 6.2:1 for patients with CH onset prior to 1960, to 3.0:1 for those with onset in the 1990s.

The onset of CH most commonly occurs between 20 and 40 years of age, with a sharp rise between 25 and 30 years of age. Less than 20% of CH patients have their first attack before the age of 20, and rarely is it found in children between 7 and 14 years of age, and then only in boys. Although CH is rare in children, Arruda et al studied 3 pediatric cases over a 10-year period and found, in their small sample, that sustained, long-term medical or spontaneous remissions do occur in children with early onset of CH.4 On the other hand, Manzoni studied a group of 73 late-onset patients—with onset of CH after age 50—and found chronic CH was more common in the older patients and the cluster period lasted longer.5 There also was a male to female ratio of 1.4:1 in this group, showing that development of late-onset CH was higher in older women and there was a higher number of women with chronic CH.

CH has been found to occur in families as well as in twins. Although the current perception is that 80% or more of patients with migraine have a positive family history of migraine, only 3% to 7% of CH patients have a positive family history of CH.6 It also has been postulated that CH may be an autosomal dominant inherited disorder in some families.

Migraine-like features are common in patients with CH. Approximately 25% of CH patients experience migraine features during an attack, a percentage that is higher than that found for migraine in the general population.7,8 It has been reported that these patients experience significantly more allodynia (P=0.022) than CH patients without migraine.7

Studies have also reported that CH patients have a greater incidence of smoking, drinking, and illicit drug use than non-CH patients or controls.9,10 Key facts about CH from the United States Cluster Headache Survey are outlined in Table 1.11

Clinical Features

Romberg first described CH in 1840.12 The term CH was taken from Kunkle’s work, which noted its clustering pattern.13 In 1939, Horton et al conducted a great deal of work on the subject, and proposed the nosological entity known as Horton’s histaminic cephalalgia because of the apparent, at that time, relationship to histamine.14

The major clinical features of CH are listed in Table 2. The CH may be preceded by an emotional prodrome, such as hypomania or euphoria. These premonitory signs may precede the CH episode by days or weeks. The headache itself may begin with unilateral or bilateral cervical pressure. Patients may report a feeling of warmth over the ipsilateral (same side) temporal region for several minutes before the headache begins. Mild pain may occur around the ipsilateral eye or temple, and it may be associated with nasal stuffiness. As the pain settles behind the eye, the patient may experience throbbing pain of increasing intensity that may “stab” the patient in the temple, the teeth, or the face. The “stuffed nostril,” ipsilateral to the pain, may now begin to exhibit a clear discharge.

The patient typically develops autonomic nervous system signs, including ipsilateral lacrimation (tearing), redness of the conjunctiva, ptosis, and miosis. The partial Horner’s syndrome of ptosis and miosis does not include anhidrosis. Interestingly, numerous patients, when asked for a description of the pain, state that “it feels like there is someone in my head, behind my eye, trying to push my eye out with a hot poker.”

The physical appearance of the cluster patient has been described as “leonine.”9 Patients are felt to have a particularly masculine physiognomy and physique. The noted facial characteristics include thick, coarse skin, with a peau d’orange appearance, and extreme wrinkling of the forehead and face with deep furrows, which may be secondary to heavy tobacco use, CH, or both. Facial telangiectasias are frequently seen. Freitag noted that these “tall rugged male types frequently may be seen with a diminutive wife who, between she and her husband, has the dominant personality, frequently answering questions for the patient.”15 Female CH patients also tend to have a “masculine” appearance, with very creased, or furrowed, square faces.16

Pattern of Headaches

Typical patients experience between 1 and 10 headaches per day, with most averaging 1 to 3 per day. The average duration of the attack is 45 minutes, but it may be shorter or last as long as 3 hours. The greater duration of headache may indicate that the diagnosis must be reviewed (ie, longer time may indicate unrecognized migraine). The peak pain intensity may occur within 1 minute, while the pain may last from 10 to 60 minutes or longer before gradually decreasing.

The attacks most often occur after working hours, most commonly 90 to 120 minutes after the patient falls asleep, during the first assumed rapid eye movement (REM) stage of sleep.17 The pain of CH most often is seen over the oculofrontal or oculotemporal regions, unilaterally. It does not switch sides. Some reports indicate an association with CH and obstructive sleep apnea.18 If the sleep apnea is addressed, in some cases, the CH improves.19 (See below for more regarding hypoxia and CH.)

The frequency of CH periods appears to be related to seasonal photoperiod changes (length of daylight), with the cluster periods increasing during shortening or lengthening photoperiods. Cluster periods peak within 2 weeks following the shortest and the longest days of the year (typically in July and January) and decrease within 2 weeks after daylight savings and standard time changes. CH may be provoked by use of vasodilators, including alcohol, histamines, and nitroglycerine. Changes in patients’ sleep/wake (diurnal) cycle and extreme stress may begin a cluster cycle.

During an attack, CH patients have been know to pace the floors and become agitated to the point of violence. They may strike their head into a wall or assail the wall with their fists. Some may be driven to suicide by the pain, as well as the fear of future attacks.

Classification of CH

Trigeminal autonomic cephalalgias (TACs) comprise a subgroup of primary headache disorders that present with lateralized, severe pain associated with ipsilateral autonomic features. CH is a key TAC syndrome, along with paroxysmal hemicranias; short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT)/short-lasting unilateral neuralgiform headache attacks with cranial autonomic features (SUNA); and hemicrania continua.20,21

There are 2 main classifications of CH: episodic and chronic, with the chronic form divided into primary and secondary forms. Episodic CH is the more common form, affecting at least 80% of the CH population. It is characterized by cluster periods that last between 1 and 5 months (average 3 months). Each CH period is followed by a symptom-free remission period of 6 to 24 months (average 12 months). It is rare, but possible, for an episodic cluster patient to have an isolated CH attack during their remission period. Chronic CH is defined as the absence of a remission period for at least 1 year. This also is associated with an increased frequency of headache and a greater resistance to treatment. The chronic form of CH accounts for only about 10% of cases, with a rate varying between 5% and 22% in various series.22

People who begin their CH and never obtain a remission are termed primary chronic CH patients. People who begin with episodic CH and then develop shorter and shorter remission periods until they no longer have remission are termed secondary chronic CH patients.

It also appears that it may be possible to track some of the factors that appear to predict the evolution of episodic CH to chronic CH. These factors include: onset at an older age, particularly in females; the presence of 4 or more accompanying symptoms (mostly seen in men); cluster periods lasting more than 8 weeks (mostly seen in men); remission periods lasting less than 6 months; and cluster periods having a higher frequency with sporadic CH.23


The clinical diagnosis of CH should be derived from history and neurological examination. Therefore, the performance of diagnostic tests is necessary only to help in the diagnosis of a particularly difficult or questionable case. It also is true, that “as you treat them, so shall you diagnose them.” If a patient is diagnosed with CH and treatment does not help at all, the diagnosis must be questioned.24

There are several tests that have been used to determine the pathophysiology of CH. These include facial thermography, used to identify the “cold spot” in the supraorbital region of the forehead, and transcranial doppler ultrasound and temporal artery ultrasound studies used to measure cerebral blood flow. The latter test has shown middle cerebral artery velocity decreases 30 minutes after the onset of the headache attack on the ipsilateral side; additionally, 50 minutes after headache onset, there is a decrease in blood velocity on the contralateral side.

The temporal artery velocity, as seen with ultrasound, is also lower and thought to be secondary to associated carotid artery vasomotor reactivity. There are no specific electroencephalographic (EEG) findings indicative of CH. Neuroimaging of TACs shows that the posterior hypothalamus appears to play a major role in the initiation and maintenance of these headaches. This is further implied by the usefulness of posterior hypothalamus stimulation for the treatment of these disorders.25,26 (See sidebar for more on pathophysiology of CH, page 64.)

Psychological testing using the Minnesota Multiphasic Personality Inventory (MMPI) indicates that CH patients are very similar to migraine patients but differ from controls. Both migraine and CH patients demonstrated the “conversion V” pattern, which indicates high scores on the “hypochondriasis” and “hysteria” scales (scales 1 and 3) and relatively lower scores on scale 2, or depression. Both headache groups appeared to show obsessive-compulsive behavior and were also perfectionistic. It does not appear that CH patients differ greatly in psychological profiling from other chronic headache patients (migraine and tension type).

Differential Diagnosis

The most common misdiagnoses of CH include: migraine; trigeminal neuralgia; sinusitis; “psychogenic headache”; cervicogenic headache; and chronic paroxysmal hemicranias (Table 3).

It should not be difficult for the clinician to differentiate migraine headache from CH. Migraine also is unilateral, but the pain is throbbing and the entire “half-head” (hemicrania) is involved. Unlike patients with CH, the migraineur may experience a “side shift,” with the headache being either left or right sided. The duration of migraine, ranging from 4 hours to 72 hours, is much longer than that of CH. There are no specific autonomic signs or symptoms, except for those found in an autonomic aura that may precede the migraine. This may include hyper- or hyposexuality, depression, euphoria, marked increased or decreased appetite, and extreme fatigue or high energy. Allodynia secondary to central sensitization is not uncommon.

Trigeminal neuralgia most specifically is a “face pain,” not a “headache.” A major diagnostic aspect is that there are typically intra- or extra-oral or facial trigger zones that will induce the pain when the patient is touched or excited. The pain attributes are electrical and lancinating, and they last seconds. Stories of patients who come to the physician with only half of their face shaved or made-up are not fiction. These patients are afraid to touch anywhere near a trigger zone. CH sufferers have no such trigger.27

The pain of sinusitis is typically fairly constant, not episodic, and may involve the sinuses above the eye, as well as the maxillary sinuses, but the pain is not behind the eye, and tends to be bilateral. Moreover, many patients with possible sinusitis actually have migraine, and are treated successfully with a triptan.

Cervicogenic headache is unilateral headache secondary to facet joint abnormalities or other diatheses originating from, typically, the second and third cervical vertebral regions. This pain commonly is constant, but it may be intermittent and depend on movement of the neck. There is no clustering noted.

The term “psychogenic headache” typically describes a waste-basket diagnosis made by clinicians who do not know how to determine the actual headache diagnosis. In spite of the obvious clinical aspects of CH, this diagnosis is still used. The diagnostic aspects of psychogenic headache should include the facts that it is constant, generalized, nonfluctuating, and while the patients give very vivid descriptions of the severity of the headache, it appears to observers to be mild. There is, typically, a total non-responsiveness to medications and other treatments.

Chronic paroxysmal hemicrania (CPH), a TAC, may reasonably be felt to be CH. However, the CPH attacks may occur up to 38 to 40 times per day and last seconds or several minutes. This is obviously in contradistinction to CH. The site of the two pain problems may be quite similar. Finally, SUNCT syndrome has also been incorrectly diagnosed as CH.

Approaches to Management

The medication therapy for CH may be considered either abortive or prophylactic. Abortive agents work better, clinically, in episodic cluster patients. The typical episodic CH patient may have more than one cluster attack per day. For this reason, it is appropriate to use prophylactic agents during an episodic cluster period and abortives for breakthrough headaches. This means that the clinician should know the average length of a patient’s cluster period and when it typically may start. That way, the patient would begin prophylactic medications prior to the expected onset of the episodic cluster period, and then they should slowly wean off the prophylactic medications when the episodic period ends.

Abortive Agents

A clumsy, but very useful, abortive agent is 100% oxygen given via face mask at 7 liters per minute over 15 minutes. Generally, it results in the cessation of the CH within 15 minutes.2 Oxygen is best supplied via a large tank in the patient’s home, typically in their bedroom (Table 4). For daytime attacks, the use of a small tank of oxygen is useful but problematic. It should be noted that there is a clinical suspicion that overuse of oxygen will increase the frequency of CH.

Sumatriptan (6 mg, given subcutaneously) is another first-line abortive agent for CH. Because of its manner of use and its fast onset of action, sumatriptan was thought to be the only triptan useful for management of CH. However, a number of studies have shown that some patients do better with other triptans, such as rizotriptan.28 An intranasal triptan formulation (zolmitriptan) also is recommended as a first-line therapy.29 Evans et al found that sumatriptan may inhibit transient potential receptor vanilloid 1 (TRPV1) ion channels in trigeminal neurons.30 The study supported previous work indicating that TRPV1 channels in trigeminal nociceptors play a role in craniofacial pain. Sumatriptan inhibited TRPV1 channels, which are functional in neurons projecting to cerebral dura, suggesting “a specific role for these channels in migraine or cluster headache.”30

Lidocaine, in a 4% aqueous solution applied by a dropper or a pledget to the sphenopalatine ganglion (via the ipsilateral nostril), is another abortive agent. In the past, the sympathomimetic agent cocaine was used in this fashion and gave good relief from CH. The main problem with the use of intranasal lidocaine is that for best results, the patient must lie down and get into a proper position with their head to enable the medication to reach the ipsilateral sphenopalatine ganglion. This agent may be problematic to use, depending on where a patient is and when they need the medication.31

Many CH patients experience a cluster attack routinely during what is presumed to be the first REM stage of sleep—90 to 120 minutes after they fall asleep. A rapidly acting dihydroergotamine (DHE) given intranasally is an excellent abortive, but should not be used within 24 hours of a triptan. In my experience, taking two ergotamine tablets, or oral triptans, immediately before going to sleep will enable the medication to reach a peak plasma level and typically abort this headache.

Intranasal butorphanol, an opioid agonist-antagonist, is a good abortive agent, but caution must be taken to avoid rebound headaches. Common side effects of this medication range from a burning feeling in the sinuses to “weird” and “hyper” feelings and dysphoria; some patients may experience hallucinations. In my experience, butorphanol’s addiction potential is markedly underrated.

Prophylactic Agents

Prophylactic treatment is used for both episodic and chronic CH (Table 5, page 62). It is more difficult in the chronic CH patients who may need to periodically change their medications, which would depend on the patient. Some patients will do fine throughout a year, and then need to have a medication change, probably, as there might be a bit of tachyphylaxis going on. But later on, those patients may be switched back to the original medication. Most chronic CH patients, as well as some episodic cluster patients, may need a combination of prophylactic medications to achieve relief. It is, therefore, incumbent upon the clinician to understand the various aspects of these medications, alone and in combination with other medications.

Verapamil, a calcium channel blocker, is felt to be a first-line prophylactic medication—useful in both episodic and chronic CH. It may be used in combination with other medications, but carefully. Verapamil may take about 2 weeks to be effective in episodic CH, at an average dosage of 240 to 350 mg per day. In chronic CH, it may take up to 4 to 5 weeks to show a response, and the average dose in those cases is 572 mg per day, given in divided dosages. Time-release formulations are useful. In 2007, it was noted that verapamil may be associated with cardiac arrhythmias in CH patients, so baseline electrocardiogram is obligatory with increasing doses.32 Another calcium channel blocker, nimodipine, also is useful at a dosage of 120 mg per day.

Corticosteroids (oral or intramuscular) can be useful in the prophylaxis of episodic CH. Patients are given a 10-day burst treatment with prednisone, starting at 80 mg per day and weaning down to 0 mg over 10 days. Some clinicians have used prednisone on a daily basis at dosages ranging from 10 mg to 80 mg per day (average 40 mg). Responsiveness was good, but in a significant number of patients, headaches returned as the steroid dose was decreased. Low-dose ergotamine (ergots) or lithium carbonate may be used concomitantly with prednisone. Both ergots (Cafergot) and DHE—another form of ergot—are used, along with triptans, to abort migraine, as well as CH. One cannot mix the two within a 24-hour period.

Lithium carbonate is an excellent prophylactic treatment for CH. The typical dosage is between 300 and 900 mg per day, given in divided doses. It may be started as a single drug, and if efficacy is lost, a second drug, such as verapamil, can be added. The drug action is not well understood, although it appears to work in the hypothalamus, possibly in the region of the supraoptic nucleus, and it may help by restoring chronobiological homeostasis. It does have serotonergic properties. Blood levels of patients on lithium should be monitored for signs of toxicity. However, CH patients appear to metabolize lithium carbonate differently than do bipolar patients. Therefore, CH patients may take more lithium without the side effects seen in bipolar patients, including shakiness, dizziness, and subjective cognitive changes.

Two anticonvulsants are widely known to be effective in managing migraine, chronic daily headaches, and CH. The first, valproate sodium given at doses of 500 to 750 mg per day, may be used alone or, carefully, in combination with ergots and possibly lithium, for chronic CH prophylaxis. The agent usually is well tolerated in the lower doses used for headaches; however, the FDA recently issued a warning that valproate sodium can cause decreased IQ scores in children whose mothers took the medication during pregnancy. The agency now reports that these agents are contraindicated in pregnant women for the prevention of migraine headaches.33 Topiramate, the second anticonvulsant, has been used successfully in several studies for the prophylactic treatment of CH.34,35

Other, older treatments include methysergide (4 mg/d to 10 mg/d, given in divided dosages). It may be used as monotherapy for CH. However, its use is limited to 5 to 6 months at a time, due to the possibility of development of retroperitoneal fibrosis as well as fibrotic changes in the lungs and heart valves. It should not be used for repeated bouts of CH because tachyphylaxis may develop, making it ineffective. In some cases, the addition of a corticosteroid in these patients can restore the effectiveness of the drug. All patients should receive a chest x-ray and intravenous pyelogram to rule out any end-organ damage from prolonged or repeated use of the medication.

In reality, especially for chronic CH patients, a combination of extended-release verapamil, lithium carbonate, and valproate sodium may need to be given together to enable the chronic patient to achieve prophylaxis, with triptans used for any breakthrough headaches. As with migraine sufferers, medication overuse headache also has been seen in CH patients who may abuse opioids, ergots, triptans, and even simple analgesics.36

Other Preventive Measures

Histamine desensitization, using intravenous histamine infusions of 5.5 mg of histamine phosphate for 24 hours for 10 days, may be useful in the treatment of intractable CH, as well as those who have lost their ability to respond to lithium carbonate or verapamil.

Surgical procedures may be used for the management of patients who have failed all pharmaceutical treatment options. One preferred treatment is percutaneous radiofrequency lesioning of the ipsilateral gasserian (trigeminal) ganglion. This abolishes the corneal reflex and has a 66% chance of producing pain relief. This will last until the corneal reflex recovers, typically over several years.

Another surgical procedure includes sectioning of the nervus intermedius, which can lead to sterile inflammatory reactions secondary to parasympathetic discharges. Retro-gasserian injection of glycerol has also been useful. Trigeminal root sectioning in the posterior fossa is a last resort, even for treatment-resistant patients, because of the risk of painful numbness. A commonly used surgical procedure several decades ago was the radiofrequency thermocoagulation of the gasserian ganglion.37,38

The placement of an electrode into the posterior inferior hypothalamus appears to be of therapeutic value.39 Four years after the initial work, Schoenen et al found that hypothalamic stimulation was found to be helpful in most, but not all, patients with treatment-resistant chronic CH.40 The same results were noted by another group.41 A recent case report showed that secondary chronic CH was treated well by hypothalamic stimulation in a drug-refractory patient (2-year follow-up).42

Other forms of neurostimulation for chronic CH include occipital nerve stimulation43,44 and greater occipital nerve injections.45,46 High cervical cord stimulation for chronic CH also has been shown to decrease the mean attack frequency, as well as mean intensity and duration of attacks.47 Finally, microvascular decompression of the pterygopalatine ganglion in patients with refractory headaches has been attempted, but it did not provide pain reduction.48


CH is an excruciatingly painful headache diathesis. Its primary etiology remains unclear, but there are various secondary manifestations of autonomic nervous system dysfunction. Correct diagnosis is clinically mandatory and must be followed by appropriate abortive and/or prophylactic medications.

Pathophysiology of Cluster Headache

Cluster headache (CH) is associated with changes in normal neurophysiology. Pathways of CH may involve cranial nerves V, VII, IX, and X. Changes in these nerves may induce the associated signs and symptoms of CH. These changes include hypothalamic dysfunction that can induce chronobiological abnormalities—leading to impaired sympathetic and parasympathetic activity—as well as decreases in autoregulatory function in vasomotor regulation and chemoreceptor responses to hypoxemia. Sustained hypoxemia, secondary to diminished carotid body activity, can create the pathophysiological milieu needed for the cluster attack.

Hypothalamic activation, as noted by positron emission tomography (PET) scanning, indicated an increase in regional cerebral blood flow during CH.1 Other studies have indicated that the hypothalamic activation in CH appears to be specific to CH.2 Goadsby, for example, noted a difference in the structure of the hypothalamus specific to CH patients.3,4

There also appeared to be a generalized hyperfunctioning of pain pathways in CH as indicated by CH patients in whom cerebral blood flow was found to be lower in the contralateral primary sensory region and the thalamic regions. It was felt that the pattern of cerebral activation was changed in CH during the interictal period on the ipsilateral side to the headache, altering the central tonic pain mechanisms.5 Gi proteins, which are known to help modulate pain perception, are found to be reduced in the lymphocytes of migraine and CH patients, possibly representing a biological mechanism for cell hyperexcitability in primary headache disorders.6

During their active cluster cycle, neuroendocrine changes, including decreased plasma testosterone and luteinizing hormone, have been observed in CH patients; these changes possibly are indicative of a change in the hypothalamic-pituitary axis. There is a decreased response by thyrotrophin to thyrotrophin-releasing hormone in patients with CH. A change or loss of circadian rhythmicity of neuroendocrine substances has been noted to include decreased melatonin levels during the night throughout active cluster periods. Also noted, during cluster periods there is a loss of normal circadian rhythms for blood pressure and temperature, beta-endorphins, and plasma cortisol.

Disturbance of the sympathetic nervous system ipsilateral to the headache is indicated by the “cold spots” found during facial thermography in cluster patients, both between and during cluster attacks. This suggests that the changes in supraorbital artery hemodynamics possibly are associated with impairment of autonomic regulation. Pupillometric studies have shown a diminished response to tyramine eyedrops, indicating impairment of sympathetic neuronal transmission on the side of the CH. Altered lipolysis was found in patients with CH, both in symptomatic periods and in remission, possibly due to reduced nocturnal sympathetic activity and, therefore, is a possible indicator of central sympathetic dysregulation and hypothalamic dysfunction.7

A number of neurochemical changes have been associated with CH. Histamine initially was reported to be a mediator of CH, but this has been a subject of significant debate. Substance P and associated neuropeptides calcitonin gene-related peptide (CGRP) and vasoactive intestinal polypeptide (VIP) may induce a rostral spread of nociception to the trigeminal nerve. This is supported by the alleviation of a CH attack by the administration of somatostatin, a substance P inhibitor. Levels of CGRP, a marker of activation of the trigeminal pathway,8 are noted to be markedly increased during spontaneous CH3 and during CH induced by the use of nitroglycerin.9 The use of oxygen as well as subcutaneous sumatriptan normalized the level of CGRP.10 VIP is a marker of parasympathetic activity and its ictal increase was noted in the same study.9

Other neurochemicals also change. Mosnaim found decreased plasma levels of methionine-enkephalin in patients.11 Patients were found to have significantly higher levels of brain-derived neurotrophic factor (BDNF) both inside and outside of cluster episodes. BDNF is associated with pain modulation and central sensitization. It also is known to interact with CGRP.12 Patients with CH had significantly increased relative interleukin-2 (IL-2) gene expression levels between headache attacks during active periods and between headache attacks during a CH episode.13 Another study done by Steinberg found no intracranial inflammation, but that conclusion was secondary to their white blood cell single-proton emission computer tomography (WBC-SPECT) evaluation.14 Nitric oxide (NO) has also been considered to play a role in CH.15

The sympathetic nervous system hypofunction noted above appears to be countered by parasympathetic hyperfunction during CH. The efferent impulses from the greater superficial petrosal nerves, which originate in the parasympathetic nuclei in the hypothalamus, may induce CH. Excessive release of acetylcholine from parasympathetic nerve terminals could stimulate histamine release from mast cells and this, along with other neurochemical pain mediators, would stimulate nociceptors directly, leading to increased pain and the release of substance P, inducing more pain, inflammation, and histamine release.

Studies using diffusion tensor imaging found that patients with episodic CH have microstructural brain changes in regions in the pain matrix.16 These included microstructural changes in brain tissue bilaterally in the white matter of the brainstem, the frontal lobe, the temporal lobe, the occipital lobe, the internal capsule and the right side of the thalamus and cerebellum. They also noted changes in the medial lemniscus and central sympathetic pathways. When looking at volumetric approaches, Seifert and his group noted a decrement in grey matter or cortical thickness, specifically in the angular gyrus, and the precentral gyrus contralateral to the headache side.17 The cortical thickness of an area in the primary sensory cortex correlated with disease duration. Another group found decreased grey matter volume in the right thalamus, the head of the right caudate nucleus, the right precentral gyrus, the right posterior cingulate cortex, the bilateral middle frontal gyrus, the left inferior parietal lobule, and the left insula.18 The decreased left middle frontal gyrus volume was, the authors felt, significantly correlated with the duration of the disease.

Last updated on: April 30, 2018
close X