Neuro-Ophthalmology & Neuro-Otology
Aug. 17, 2022
At vero eos et accusamus et iusto odio dignissimos ducimus qui blanditiis praesentium voluptatum deleniti atque corrupti quos dolores et quas.
Cluster headache is a primary headache disorder characterized by severe, unilateral pain that is orbital, supraorbital, or temporal and 15 minutes to 3 hours in duration occurring 1 to 8 times a day. Accompanying symptoms include ipsilateral: conjunctival injection, lacrimation, nasal congestion, aural fullness, rhinorrhea, facial sweating, eyelid edema, and miosis. The current update provides the latest understanding and management of this disorder, including the latest clinical trial data.
• Cluster headache is a relatively common, very severe form of primary headache that belongs to the family of trigeminal autonomic cephalalgias.
• Cluster headache involves dysfunction of central nervous system elements concerned with pain control and with links to circadian and circannual mechanisms.
• Acute cluster headache can be treated with oxygen, intranasal triptans (sumatriptan and zolmitriptan), and injected sumatriptan, and for episodic cluster headache noninvasive vagal nerve stimulation.
• Medicines or strategies useful in the preventive management of cluster headache include verapamil, lithium, corticosteroids, topiramate, melatonin, greater occipital nerve injection, the calcitonin gene-related peptide (CGRP) monoclonal antibody galcanezumab for episodic cluster headache, and noninvasive vagal nerve stimulation.
• Sphenopalatine ganglion stimulation has been demonstrated to be useful for the treatment of medically refractory chronic cluster headache.
Cluster headache has been recognized for over 350 years (169); an excellent clinical description can be found in Van Swieten’s textbook (156) and a reasonable description in the Spanish literature (273). It has been known by many names, notably Horton headache in North America (154; 152; 153) and migrainous neuralgia in the United Kingdom (141; 24; 181; 28). Other names that have probably described substantially similar syndromes, perhaps now called the trigeminal autonomic cephalalgias (182), include ciliary neuralgia (298), erythroprosopalgia of Bing (25), and hemicrania periodic neuralgiforms (269). Symonds recognized it as a particular variety of headache (349). Sphenopalatine ganglion neurosis (337) and Vidian neuralgia (360) were described as affecting females and were probably migraine with a facial distribution and cranial autonomic symptoms, given data that confirm the common appearance of cranial autonomic features in migraine (163). The periodicity of the attacks led to the current nomenclature (142; 143; 144), cluster headache (177), which seems to describe it well and respect its biology.
There are two major forms of cluster headache, namely, episodic cluster headache and chronic cluster headache (144). The term “attack” is generally used to refer to a distinct episode of pain, whereas a bout or period is a series or “cluster” of attacks. Thus, episodic cluster headache is characterized by periods lasting 7 days to 1 year separated by remission periods lasting 3 months or longer. Chronic cluster headache is characterized by the absence of remission for 1 year or short remissions of less than 3 months. It should be noted that the time for remissions has increased from 1 month in ICHD-2 to three months in ICHD-3. Both classifications assume no preventive treatments. Patients’ cluster headaches may evolve from the episodic to the chronic form with sporadic, late-onset attacks, a high frequency of cluster periods, and short-lived duration of remission periods when the headache is still the episodic form, all correlating with the transition to the chronic form (357). Chronic cluster headache may be seen de novo (primary), or it may evolve from the episodic form (secondary); differences between the two include earlier cluster headache onset and duration of attacks varying more frequently between 120 and 180 minutes (356), although in practice this distinction seems to have no clinical utility (70). Cluster headache is extremely disabling when patients are experiencing a bout (56). Even with chronic cluster headache, cycling can occur (155), which could be called “micro-cycling”; a similar phenomenon is seen in SUNCT (155). Somewhat more than half of patients report that they can predict the onset of their bout, including when they note what are described as “shadows” (275).
The attack consists of the rapid onset of headache that builds to a peak in about 10 to 15 minutes and lasts for approximately 30 to 180 minutes, although longer attacks are not infrequent (364). The headache is almost always unilateral. In order of decreasing frequency, common sites of pain are orbital, retro-orbital, temporal, supraorbital, and infraorbital. The head pain occasionally switches sides, and in extremely rare cases it can be bilateral. The pain is described as the worst experienced by humans (31). Typically, the pain is in the trigeminal nerve distribution, although extra-trigeminal pain, especially in the suboccipital area, is well recognized. Rarely, attacks may occur without cranial autonomic symptoms, although in each such case we have seen, the patients have had typical agitation and restlessness. Cranial autonomic symptoms can also occur alone (220), typically in patients who have been treated with nerve root sections (120). The number of attacks per day varies, usually from one to three, but the International Headache Society classification allows from one every second day to eight per day. More frequent or treatment-resistant attacks should lead to consideration of other trigeminal-autonomic syndromes, such as paroxysmal hemicrania. Migrainous features can be seen in cluster headache, such as nausea, photophobia, and phonophobia, in as many as half of cases (15; 382; 340; 105). Moreover, between 6% (328; 58) and 14% (16) of patients report typical migraine aura with their cluster headache. Hemiplegic aura has been reported (332) as has prolonged aura (183). This needs careful evaluation from a diagnostic viewpoint. Cluster headache is well-recognized in children (237; 180; 209), and this author has seen a 3-year-old child with a convincing story, although it may begin as late as 89 years of age (98). About one third of patients report improvement of cluster headache attacks with sexual activity (137). The recognized male preponderance probably contributes to misdiagnosis in females (208; 340; 105). Authors have noted premonitory symptoms, such as dull ache or cranial autonomic symptoms, before an attack (339; 42).
One remarkable aspect of cluster headache is its periodicity, with attacks often occurring at almost the same time every day (76; 186); this phenomenon has resulted in suggestions that human circadian pacemaker physiology is involved (19). Nocturnal attacks occur during both REM and non-REM sleep periods (176). Sleep studies show reduced REM sleep in cluster headache (18). Attacks are characterized by excruciating pain that is regarded by sufferers, without exception, as the worst pain that they have ever experienced. The pain is associated with ipsilateral cranial parasympathetic overactivity, lacrimation, conjunctival injection and nasal stuffiness or rhinorrhea, and an ipsilateral partial (third neuron) Horner syndrome (72; 73) with ptosis and miosis. These symptoms are equally likely in episodic and chronic cluster headache (108). In exceptional cases, the cranial autonomic symptoms can result in obvious swelling (11) or even nose bleeds. Voice change, which is likely a cranial autonomic symptom, may be seen in cluster headache (329; 163). Although less common than in migraine, photophobia and phonophobia occur in cluster headache (367) and when present are more often lateralized to the side of the pain (155).
Neurologic examination may reveal mild ptosis and miosis on the side of the headache, especially during or immediately following the attack, and impaired trigeminal sensation from time to time, although the latter will trigger a search for a lesion. Ipsilateral tenderness of the carotid artery, periorbital swelling, and congestion of the conjunctiva are also noted. During a bout, a subtle but convincing, mild, partial Horner syndrome ipsilateral to the pain is often present. Alcohol, nitroglycerin, and histamine can induce attacks during cluster periods (154; 75). Nitroglycerine-induced attacks have the same phenotype as spontaneous attacks and allow dissection of disease features, such as pre-pain onset of cranial autonomic symptoms (375). Allodynia and interictal pain are seen in cluster headache (182; 219; 380) and, when present, the patient almost invariably has a personal or family history of migraine.
The prognosis of episodic cluster headache is generally good. As patients age, cluster headaches usually disappear. Approximately 10% of patients have the chronic variety of cluster headache, which tends to run a difficult and often intractable course. Transformation from the episodic to the chronic form, and less commonly from the chronic to the episodic variety, is well recognized. In a follow-up study with a range of 2 to 17 years, 27% of 60 newly diagnosed patients had only one attack over a mean follow-up of 8.9 years (336). A number of patients with cluster headache have obstructive sleep apnea (176; 41; 263; 264; 133), and it can be clinically useful to inquire about such symptoms and investigate appropriately.
A 42-year-old male presented with a 13-year history of strictly left-sided headache. The pain centered on the eye and felt as though the eye was being pulled out with sharp prongs. Attacks of pain lasted 90 minutes and were associated with left-sided photophobia and a sense of restlessness. He had left-sided lacrimation and a sense of aural fullness. The attacks came in bouts for 9 weeks every 18 months to 2 years, and he was pain-free between times. He had a normal general and neurologic examination.
The etiology of cluster headaches is not known. It seems likely that some inherited basis for the problem exists (309; 308; 242; 243; 197; 348; 307; 374), although it has proved hard to study (84). Two genome-wide association approaches have identified the same four loci (importantly, not causal variants, including other loci not overlapping), with five of the seven loci identified previously reported in migraine (140; 267). Cluster headache is probably not related to the familial hemiplegic migraine CACNA1A gene (136; 334) or to nitric oxide synthase genes (335) or to methylenetetrahydrofolate reductase C>T polymorphism (320). There are reports of a polymorphism in the hypocretin receptor 2 gene (289; 288; 319), which would be attractive because of the localization in the hypothalamus (338), although the finding could not be reproduced by others (21; 274), and the CSF levels of hypocretin-1 are unaltered in cluster headache (38).
The three major aspects of the pathophysiology of cluster headache are the trigeminal distribution of the pain, the autonomic features, and, more important, the inherent periodicity of the attacks and bouts (120). Ekbom’s observation of a patient suffering an acute cluster headache demonstrating angiographic changes in the internal carotid artery suggested a pathological focus in the region of the cavernous sinus (76). The arguments for this locus for the disease have been set out clearly (249; 139). There are no consistent markers of inflammation when measured in the periphery (294; 342) and signs of cavernous sinus inflammation when studied with SPECT/MRI (317). Three lines of evidence suggest that the cavernous sinus is not the site of the basic biological problem. The clinical manifestations outlined above, in particular the periodicity, the neuroendocrine changes, and the functional neuroimaging results (48; 57), suggest that cluster headache involves biological dysfunction in the posterior hypothalamic grey matter pacemaker neurons.
Pain transmission in cluster headache. The pain of cluster headache is a first division of trigeminal phenomenon. Many of the autonomic features are due to seventh nerve activation, whereas the remaining changes are due to a transient cervical sympathetic deficiency. It is both clinically likely and experimentally clear (126) that a trigeminal-autonomic reflex underlies the pain expression of cluster headache. There are neuropeptide changes in the cranial circulation in both calcitonin gene-related peptide (CGRP) and vasoactive intestinal polypeptide (118; 91), in tears (162) and nociception (86) release during headache that reflect the trigeminal and autonomic dimensions of the syndrome. Interestingly, CGRP can trigger attacks in episodic cluster headache patients in a bout and not out of one (369), whereas CGRP levels are associated with disease activity (341). Interestingly, pituitary adenylate cyclase activating polypeptide (PACAP) triggers attacks at about the same rate as vasoactive intestinal peptide (370), without attendant release of CGRP (276). Human studies have shown that this process does not seem to be active in cluster headache (235), at least as far as experimental plasma protein extravasation is concerned (250). Interestingly, in rats with 5-HT1D immunoreactivity there are documented primary afferents fibers that innervate postganglionic neurons in the sphenopalatine ganglion, suggesting a site for modulation by triptans (158). Taken together, the clinical and experimental evidence has suggested the term “trigeminal autonomic cephalalgias” as an umbrella term for cluster headache and related conditions (128). Drummond argues for a peripheral sympathetic autonomic disturbance (74), and minimal systemic autonomic disturbances are reported (366). For cluster headache, a crucial aspect of the dysfunction appears to lie within the ipsilateral hypothalamic grey (228; 96), and it is clear that the carotid flow changes are driven by the ophthalmic division of the trigeminal nerve and are not due to cluster headache as such. Further evidence of brain mechanisms in cluster headache patients is the fact that the nociceptive flexion reflex threshold is asymmetrically reduced (312). More challenging is the observation of a typical cluster headache patient whose attacks persist after trigeminal root section (120; 160). In comparison to migraine, it is notable that there is no defect in habituation of the nociceptive blink reflex in cluster headache (151), with lateralized facilitation of central pain pathways (150). Although the site of action of sumatriptan has never been fully resolved, human (59) and animal (253) evidence continues to develop the concept of a possible central site of action.
Neuroendocrine changes in cluster headache. Kudrow pointed out that testosterone levels are altered in cluster headache patients during a bout (172; 174). Leone and colleagues identified reduced responses to stimulation by thyrotropin-releasing hormone (196), and interesting observations have been made of disordered circadian rhythm for cortisol, growth hormone, luteinizing hormone, and prolactin (40; Waldenlind and Gustafsson 1987). Despite suggested involvement of the hypothalamic-pituitary axis, somatostatin infusion does not trigger acute cluster headache even with suppression of growth hormone (201). One area involved in human clock systems is the suprachiasmatic nucleus in the hypothalamic grey, which sits at the base of the third ventricle (245; 244). Melatonin is produced by the pineal gland and has a strong circadian rhythm, which is regulated by the suprachiasmatic nucleus. Connections between the retina and the hypothalamus are thought to provide light cues for the circadian rhythm (148). The characteristic nocturnal peak of melatonin secretion is blunted during the active phase of cluster headache (371), whereas urinary excretion of 6-sulphatoxymelatonin, the main metabolite of melatonin, is reduced both within and between bouts compared with controls (195). This suggests a defect with some permanent nature possibly amenable to treatment (188). Involvement in some part of nitric oxide mechanisms in cluster headache is supported by the finding of elevated levels of nitric oxide-oxidation end products in the CSF of patients with cluster headache (346). Latencies of the endogenous event-related-potential components are significantly increased during the cluster period as compared with outside the cluster period and with healthy subjects (89), further contributing to the notion of a central nervous system dysfunction.
Functional imaging and other physiological studies of cluster headache. Acute cluster headache triggered by nitroglycerin produces brain activations on positron emission tomography (PET) that fall into three categories: areas generally associated with pain, an area that seems specific to cluster headache, and vascular structures (228). The anterior cingulate was significantly activated, as would be expected, because in most human PET pain studies, activation of the anterior cingulate is observed, perhaps as a part of the affective response. Its degree of activation seems to track the subjective reporting of pain (231) and can be seen in spontaneous attacks (344) and when using fMRI (247). Activation was also noted in the frontal cortex and insulae and ventroposterior thalamus contralateral to the side of the pain. In addition, activation in the ipsilateral basal ganglia was observed. This is not the first observation of basal ganglia changes associated with pain (44; 63). This may simply relate to movement, the wish to move that is common in cluster patients, or even some deliberate inhibition of movement. Moreover, after occipital nerve stimulation (see Management section), there was normalization of pain matrix areas and lower pons but not of the posterior hypothalamic region. Remarkably, the perigenual anterior cingulate cortex was hyperactive in occipital nerve stimulation responders compared to nonresponders (211). It has been shown that pain responses to nasal ammonia are stronger in cluster headache patients in the hypothalamus than in controls (318). The use of resting state connectivity analysis of abnormal patterns of hypothalamic connectivity with areas active in pain has been reported (287); this methodology will require more work. It may have a parallel in cortical hyperexcitability changes reported in episodic cluster headache (50).
The only activated area that is particular to cluster headache is a region near the base of the third ventricle, in the region of the posterior hypothalamic gray matter. One may speculate that orexinergic neurons play some role at this level (149), and indeed, data demonstrate reduced orexin levels in the CSF of cluster headache patients (17). Activation of the hypothalamic gray matter in this region is not seen in migraine (02), nor in experimental first (ophthalmic) division of trigeminal nerve head pain (233). This region is not activated when the patient is out of a bout (229) but, remarkably, seems to have some structural difference when compared to control brains, notably a subtle increase in gray matter volume in the posterior hypothalamus (227). Moreover, a PET scan of a patient with both cluster headache and migraine who had a migraine at the time of the PET scan only showed brainstem, and not hypothalamic, activation (13). An increase of 20 Hz in the local field potential power in the region of the posterior hypothalamus has been noted in a patient undergoing implantation who had a spontaneous attack of cluster headache (29). These observations have led to implantation of deep brain stimulating electrodes into this portion of the brain with excellent outcomes in otherwise intractable patients (104; 193). Stimulation of the posterior hypothalamic region using the implanted electrodes produces a unique pattern of activation and de-activation in the brain, suggesting that the therapeutic effect is specific and not head pain generic (234). Similarly, 1H-MRS studies have reported reduced N-acetylaspartate levels, implying local neuronal loss or dysfunction in the posterior hypothalamic region (207; 372) whereas resting-state network fMRI demonstrated increased connectivity between hypothalamic and thalamic nuclei in cluster headache (296). Supporting a brain basis for cluster headache, 11C-diprenorphine PET studies reveal reduced pineal opioid receptor availability (345) and triggering of attacks from emotional trauma (311). It is to be expected that discussion remains as to both the mechanism and effective site of stimulation. It has been noted through indirect means that the optimal site for stimulation is located posterior to the hypothalamus (100; 265), although the clear anatomical definition of the posterior hypothalamus in humans seems unclear if one considers the basis on which the structure has been defined. It has been shown using high-resolution T1-weighted MRI that cortical thickness in cluster headache patients is reduced compared to controls, perhaps related to disease duration, ie, as a consequence not a cause (321), and has been refined to the middle cingulate, temporal, and lingual sulcal regions (62). The demonstration that peripheral provocation (135; 240) or activation of cranial autonomic symptoms do not trigger attacks of cluster headache reinforces the view of the condition as being driven from the central nervous system (119).
During acute cluster headache, activation or signal on the PET studies is observed in the region of the cavernous sinus (228) and has been shown, using MRA, to be due to internal carotid artery vasodilation (230). The same region is activated during capsaicin-induced experimental head pain (229; 232), but when capsaicin is injected into the skin innervated by the mandibular division of the trigeminal nerve or the ipsilateral leg, no carotid dilation is observed (232). Moreover, after trigeminal root section, attacks of tearing without pain are well recognized (120; 206). These data imply that the activation of the carotid does not relate specifically to cluster headache, but that it is a trigeminovascular autonomic reflex to first division pain. The flow changes are, therefore, an epiphenomenon of the trigeminal activation, and not part of the disease generation process in cluster headache.
Cluster headache, predominantly a disease of males (in a ratio of about 4:1) (215; 14), has an approximate prevalence of between 0.01% and 0.4% in the general population (79; 55; 355; 88; 350; 324) and between 8% and 10% in a headache clinic population (217). The most comprehensive population study coming from the Vaga project indicates that prevalence is 0.2% of the population (333). It has been claimed that cluster headache incidence is decreasing; however, this seems unlikely (27). Female cluster headache attacks tend to be rather similar to male cluster headache attacks (248; 303; 365). Community-based studies suggest general under-treatment of the disorder (295). The disorder has been clearly recognized in China (373). About 10% of cases have onset after the age of 50 years (216). Chronic cluster headache seems less common in China (68; 381).
Cluster periods start spontaneously. During a cluster period, trigger factors that precipitate headache include alcohol ingestion and nitric oxide donors or promotors, such as nitroglycerin (75) and sildenafil (60; 87). These should be avoided to prevent cluster headache attacks. Migraine triggers, such as chocolate and cheeses, have no known triggering influence on cluster headache. It has been suggested that childhood exposure to cigarette smoke may be a predisposing factor to cluster headache, although the data are limited (304), and cessation has no effect (95).
The differential diagnosis for cluster headache lies between other trigeminal autonomic cephalalgia and the so-called secondary cluster headache syndromes, where pituitary tumor pathology is an important consideration (202; 203; 92; 93; 314; 165). Unfortunately, cluster headache is often missed (114), and many patients have unnecessary dental (278) and ENT procedures (12). Reasons for missing the diagnosis include gender (ie, a female patient) (12) and the presence of migrainous features, such as photophobia or phonophobia (363). The primary headaches that arise in the differential diagnosis of cluster headache include paroxysmal hemicrania, short-lasting neuralgiform headache attacks with conjunctival injection and tearing (SUNCT), and hemicrania continua (128; 326). These can be differentiated by clinical features, such as length and frequency of attacks, and by the response to indomethacin, which is absent for cluster headache, at least in an open-label study (09) and in the author’s placebo-controlled experience. It is likely that cases of cluster headache labeled as indomethacin-sensitive are misdiagnosed (284). Trigeminal neuralgia is sometimes mistaken for cluster headache, but the attacks are much shorter, and autonomic activation does not seem to be an important issue. Hypnic headache is a more moderate form of headache that is usually seen in the elderly. It is characterized by attacks of moderately severe headache that come on after falling asleep. The attacks can usually be controlled by lithium taken at night or by caffeine consumption (292; 260; 132; 67; 157; 246).
At least five attacks fulfilling the following criteria:
• Severe or very severe unilateral orbital, supraorbital, or temporal pain lasting 15 to 180 minutes if untreated.
• Headache accompanied by one of the following:
- ipsilateral conjunctival infection or lacrimation
• Attacks have a frequency from one every other day to eight per day
• Attacks are not attributed to another disorder
Episodic cluster headaches also have at least two cluster periods lasting 7 days to 1 year and separated by pain-free remission periods of 3 months or more.
Chronic cluster headaches also have attacks that recur for more than 1 year, without remission periods or with remission periods lasting less than 3 months.
It could be argued that secondary cluster headache should be divided into lesions that produce trigeminal-autonomic activation, which one might call pseudo-secondary cluster headache, where episodicity is not a key feature, and true secondary cluster headache. The latter is likely to be due to lesions in and around the diencephalon. A secondary cluster headache should be suspected when the clinical features of the headache are atypical. Atypical features include lack of periodicity, inter-attack pain (although this can be seen with frequent attacks), poor response to medications, and neurologic signs other than a partial Horner syndrome. Aural fullness has been removed as a cranial autonomic symptom being regarded as not adding to the classification; this author dissented from that view. Interestingly, diagnosis and management are considerable concerns to general practitioners and neurologists alike (36).
Other primary trigeminal autonomic cephalalgias
• Paroxysmal hemicrania
• SUNCT/SUNA (short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing/cranial autonomic features) syndrome
Similar secondary headaches
• Tolosa-Hunt syndrome
Secondary trigeminal autonomic cephalalgias – cluster-like headaches
- Carotid artery dissection (300; 213; 293; 131; 359; 37) or aneurysm (134)
- Pituitary tumors (353; 134; 203; 257; 92; 93; 204; 314)
- Maxillary sinusitis (239)
• Posttraumatic or surgery:
- Facial trauma (182)
- Impacted wisdom tooth (299)
- Cervical Syringomyelia and Chiari malformation (322)
Given that cluster headache is a lifelong condition as far as can be determined, and that cluster headache mimics can be extremely difficult to dissect clinically from primary cluster headache, some initial investigation seems appropriate. An MRI scan of the brain, with pituitary views, and screen for prolactin, TFTs, and IGF1/GH (122), would be a reasonable screening investigation if there is no suspicion of the any of the lesions reported above, which can be sought on their own merits. The results of investigations need to be interpreted in a clinical context.
The management of cluster headache can be divided into the management of the acute attacks of pain and preventive management to reduce the number of attacks or control the attacks during a bout (122; 258; 147; 64). Some surgical options are available, and these may afford a more permanent solution; however, they are not guaranteed to work, nor are they without significant morbidity. Medication overuse headache has now been reported in patients with cluster headache who all had migrainous biology (272); practitioners should be aware of this complication. The author has seen a similar problem with regular cannabis use, which is generally not helpful in cluster headache (199).
Preventive treatments. The options for preventive treatment in cluster headache are determined largely by the bout length not by the designation of episodic versus chronic cluster headache. Preventives may be regarded as short-term or long-term based on how quickly they act and how long they can be safely used. Most experts would now favor verapamil as the first-line preventive treatment, although for some patients with short bouts, limited courses of oral corticosteroids alone (159; 08), or in combination with verapamil (266), ergotamine (106), or methysergide (54) can be very useful strategies.
Verapamil has been recognized as a useful option for some time (107). It is superior to placebo (190) and compares favorably with lithium (35). What has clearly emerged from clinical practice is the need to use higher doses than had initially been considered and certainly higher than those used in cardiological indications. Although most patients will start on doses as low as 40 mg twice daily, doses up to 960 mg daily are now employed (268). Side effects, such as constipation and leg swelling can be a problem (325), but more difficult is the issue of cardiovascular safety. Verapamil can cause heart block by slowing conduction in the atrioventricular node (330), as demonstrated by prolongation of the A-H interval (256). Given that the PR interval on the ECG is made up of atrial conduction and A-H and His bundle conduction, it may be difficult to monitor subtle early effects as the verapamil dose is increased. This question needs study in this group of patients, but for the moment it seems appropriate to do a baseline ECG and then repeat the ECG 1 to 2 weeks after a dose change, usually in 80 mg increments, when doses exceed 240 mg daily (49; 184). We have also seen a malignant skin lesion related to verapamil usage, lymphomatoid papulosis (01), which may prove an issue in the future. Interestingly, the mechanism of action of verapamil in cluster headache has not yet been identified; an effect on the circadian clock is attractive (30).
Melatonin may be helpful as a cluster headache preventive (191; 279; 285; 113). Similarly, topiramate has proved effective in a number of cases (378; 102; 313; 179; 192; 290); a controlled trial is required.
Galcanezumab, a CGRP monoclonal antibody (22) useful in the preventive treatment of migraine (65), at 300 mg s/c, is superior to placebo in reducing attacks of cluster headache in patients with the episodic type (125). Interestingly, a patient global impression reporting “much better” experienced a median weekly reduction of attacks of 43% over the first 3 weeks, validating a 50% responder rate as patient centric (171). A randomized controlled trial of galcanezumab in preventive treatment of chronic cluster headache reported out negative (66), whereas open label use suggests there may be a cohort who respond (306); the latter is the author’s clinical impression. It is well tolerated and offers a clear mechanism-based approach to therapy (39).
Because of its side effects, lithium carbonate is mainly used for the prophylactic treatment of chronic cluster headache, but it is sometimes employed in the episodic variety (77; 78; 173; 224). The usual dose of lithium is 600 to 900 mg per day in divided doses. Lithium levels should be obtained within the first week and periodically thereafter. The serum level required for therapeutic response is usually 0.4 to 0.8 mEq/L. Lithium and verapamil were shown to be similar in a comparative trial (35). The side effects of lithium are significant, and its use needs to be carefully monitored. Neurotoxic effects such as tremor, lethargy, slurred speech, blurred vision, confusion, nystagmus, ataxia, extrapyramidal signs, and seizures can occur if toxic levels are reached. Concomitant use of sodium-depleting diuretics should be avoided, as sodium depletion will result in high lithium levels and neurotoxicity. Long-term effects such as hypothyroidism and renal complications must be monitored when chronic cluster headache patients use lithium for extended periods of time. Polymorphonuclear leukocytosis is a common reaction to lithium and is often mistaken for occult infection. Concomitant use with indomethacin can increase the lithium level; an important consideration is a trial in a relatively refractory patient in consideration of paroxysmal hemicrania.
An open study suggested that valproate is effective as a preventive treatment of cluster headache (146), but a subsequent controlled study showed no advantage (85). Melatonin may be helpful as a cluster headache preventive (191; 279; 285; 113). Intranasal capsaicin application has also been used (218), and intranasal civamide has modest effects (313). Induction of nitrate tolerance is not helpful in cluster headache (43). Gabapentin has been reported to be useful in cluster headache in some cases (04; 185; 352). Testosterone therapy has been used with some success for refractory cluster headache (347); there is a report of psilocybin and LSD use (323) and of 2-bromo-lysergic acid diethylamide use (164); each requires more controlled investigations. Given the involvement of circadian biology in cluster headache, the possible use of sodium oxybate as a preventive in chronic cluster headache is an interesting new suggestion (166).
Newer triptans have been suggested to be helpful as preventives in cluster headache, although the studies have not been easy to conduct (251). As a short-term preventive approach, a course of intravenous dihydroergotamine can be helpful (254). Candasartan was not effective in a controlled trial (358).
Acute attack treatment. Cluster headache attacks often peak rapidly and, thus, require a treatment that has a quick onset. Moreover, the established placebo effect in cluster headache (262) makes it necessary to conduct controlled trials.
Many patients with acute cluster headache respond well to treatment with oxygen inhalation (175; 99). A placebo-controlled study has established clearly that oxygen is superior to air in the treatment of acute cluster headache (47). This should be given as 100% oxygen at 10 to 12 L/min for 15 minutes. It seems important to have a high flow and high oxygen content, whereas hyperbaric oxygen confers no prophylactic benefit above its treatment in acute attacks (261). Patients suggest, and there is some evidence for (280), ultra-high flow oxygen; this needs studying. Some proportion of patients who are treated with oxygen for acute attacks have the attacks rebound quickly when it is stopped (112), and this can limit the utility of oxygen therapy. Remarkably, when studied, high flow oxygen is not clearly helpful for migraine (331).
Injectable sumatriptan is an exceptionally useful treatment for patients with cluster headache (118). It is effective and rapid in onset (82), with no dose-response benefit over 6 mg and no evidence of tachyphylaxis (83), although there has been one report of increased headache frequency (302). In general, long-term use of injectable sumatriptan in chronic cluster headache, even on a daily basis, seems well tolerated and retains efficacy (189). Interestingly, only longer attacks and increased attack frequency, which have differential diagnostic implications, seem to predict the small (< 10%) nonresponder rate (115). Sumatriptan is not effective when given preemptively as 100 mg orally 3 times daily (241), although longer acting triptans, such as frovatriptan (110), have been used in open-label fashion. Both intranasal sumatriptan (20 mg) (362) and intranasal zolmitriptan (5 mg) (46; 291) have been shown to be effective in acute cluster headache in placebo-controlled trials. Olanzapine has been reported to be helpful for acute cluster headache (303); this requires a controlled trial.
Dihydroergotamine is effective in the relief of acute attacks of cluster headache (152). Intravenous administration gives rapid relief in less than 10 minutes, whereas intramuscular injection takes longer to be effective. Intranasal administration of dihydroergotamine may also be of benefit (06). Dihydroergotamine can be used repetitively to improve some patients with more difficult cluster headache (223). Clinical experience suggests that ergotamine by inhalation may be useful, although the oral and suppository formulations are too slow acting.
Locally applied lidocaine nasal drops have been reported to be effective in the treatment of acute attacks of cluster headache (168). Patients are told to lie supine with the head tilted backwards toward the floor at 30 degrees and turned to the side of the headache. A nasal dropper may be used and the dose (1 mL of 4% lidocaine) repeated once after 15 minutes. An interesting extension of this concept is the use of cocaine in cluster headache, which works for some patients (277), presumably in the same way as lidocaine (51).
Procedures. Injection of certain nerves and more radical procedures have been employed in patients with frequent or intractable attacks of cluster headache.
Injections of methylprednisolone (120 mg) with lidocaine into the greater occipital nerve ipsilateral to the site of attack have been reported to result in remissions lasting from 5 to 73 days (07). This is a simple, relatively benign procedure that seems to help some patients (05; 03; 200). It has been suggested to be no better than oral steroids (376), a question that needs prospective study. Interestingly, changes in the nociception-specific blink reflex do not predict a clinical effect (34), suggesting that the mechanism is rostral in the CNS. Blockade of the sphenopalatine ganglion using cocaine or lidocaine is a useful temporary measure to give freedom from attacks for a few days, although an endoscopic study reported 2 of 21 patients having a break of 12 months or more (94; 283). The recurrence rate is high, however. In a series, 40% of patients with cluster headache reported improvement after radiofrequency lesion of the ipsilateral sphenopalatine ganglion (97; 255). Microvascular decompression does not seem useful (271).
Initial studies with noninvasive vagal nerve stimulation (nVNS) have been promising (259). It has been shown to be superior to standard of care for the prevention of attacks of cluster headache in an open-label study (109). Two randomized sham-controlled studies reported nVNS to be effective in the acute treatment of cluster headache in patients with the episodic type and not the chronic type (327; 119). Data from two randomized sham-controlled trials with sphenopalatine ganglion stimulation were positive (316; 129). This may evolve as the treatment of choice for medically refractory chronic cluster headache, as strong longer-term follow-up (161; 20) and cost effectiveness (282) data are now available.
Given the progress in neuromodulatory approaches, destructive procedures are, over time, becoming unacceptable.
A small number of patients with cluster headache are intractable to treatment and require consideration for surgical treatment (130). Accepted favorable indications for surgical treatment include (1) strictly unilateral headache, (2) total resistance to medical therapy or significant contraindications to the use of effective medical therapy, and (3) a stable personality profile without the potential for addictive behavior.
One destructive procedure is radiofrequency thermocoagulation of the trigeminal ganglion (270; 225). The procedure is essentially the same as that done for trigeminal neuralgia, except that both the first and second divisions of the trigeminal nerve are made fully analgesic. The overall results have been encouraging. Almost 75% of patients become free of cluster headache attacks. Recurrence of pain is possible, however, after a number of years. A repeat surgical procedure is considered in such patients. Corneal analgesia with resulting potential corneal infection or opacification and anesthesia dolorosa is an important sequelae. The benefit from surgery generally outweighs the potential sequelae, and, therefore, it is worth considering in patients with total resistance to medical treatment. Injections of percutaneous retrogasserian glycerol rhizolysis have resulted in an initial 83% improvement with a 39% recurrence (281); numbers are small, and the uncontrolled nature of the injection is unattractive.
Gamma knife radiosurgery has been suggested (101), but an open study indicates that such an approach is not useful in the long term (71) and can produce deafferentation pain (69). A more invasive approach is trigeminal root section. There are two substantial series in the literature (167; 160), and morbidity is considerable.
Based on functional imaging studies that identified the posterior hypothalamic grey as a target area of importance in cluster headache (228), deep brain stimulation has been used. Early reports indicate that this approach is effective in some patients (104; 194; 193), although again caution is required because one death has been reported (315). An alternative is occipital nerve stimulation, which is widely regarded as being helpful (32; 33; 210) and is without significant morbidity or mortality. It has a good long-term effect (212) and is to be preferred over deep-brain stimulation procedures. A completed clinical comparing 30% or 100% of the paresthesia threshold showed no difference, although participants in both arms saw reductions in median attack frequency (379). Long-term observations of three years or more suggest persistence of benefit (198).
Other treatments. There has been discussion on the use of hallucinogenic drugs to treat cluster headache. Perhaps one quarter of patients have used such treatments (301). Much work needs to be done to understand whether such an approach has a sustainable role or whether it is a distraction (226) given the comparable pharmacology between, for example, BOL-148, 2-bromo-lysergic acid diethylamide (164), methysergide, and ergotamine (23).
When reported during pregnancy, cluster headache responds to oxygen as it does outside pregnancy (117). Breast-feeding women can experience attacks that again respond to standard treatment (12; 365; 26). It seems likely, taken with the observation that high-dose corticosteroids are useful in controlling attacks, that cluster headache is largely ameliorated by the hormonal milieu of pregnancy. This is not the case for paroxysmal hemicrania.
There is no contraindication to regular general or local anesthesia in patients with cluster headache.
Peter J Goadsby MD PhD
Dr. Goadsby of King’s College London and the University of California, San Francisco, received consulting fees from Abbvie, Aeon Biopharma, Amgen, Biohaven Pharmaceuticals Inc, Electrocore LLC, Eli Lilly, eNeura, Epalex, GlaxoSmithKline, Impel Neuropharma, Lundbeck, Novartis, Pfizer, Prxis, Sanofi, Santara Therapeutics, Satsuma, and Teva Neurosciences and grants from Amgen, Celegene, and Eli Lilly.See Profile
Shuu-Jiun Wang MD
Dr. Wang of the Brain Research Center, National Yang-Ming University, and the Neurological Institute, Taipei Veterans General Hospital, has no relevant financial relationships to disclose.See Profile
Nearly 3,000 illustrations, including video clips of neurologic disorders.
Every article is reviewed by our esteemed Editorial Board for accuracy and currency.
Full spectrum of neurology in 1,200 comprehensive articles.
Neuro-Ophthalmology & Neuro-Otology
Aug. 17, 2022
Headache & Pain
Jul. 13, 2022
Headache & Pain
Jul. 11, 2022
Headache & Pain
Jul. 01, 2022
Jun. 11, 2022
Jun. 06, 2022
Jun. 04, 2022
Jun. 04, 2022