CH is a strictly unilateral headache that occurs in association with cranial autonomic features and, in most patients, has a striking circannual and circadian periodicity. It is an excruciating syndrome and is probably one of the most painful conditions known to mankind, with female patients describing each attack as being worse than childbirth.
The prevalence of CH is estimated to be 0.1%,[3
] although a recent study suggests that the prevalence of CH may be as high as two per 1000.[4
] The male:female ratio is 2.5:1.[5
] It can begin at any age, though the most common age of onset is the third or fourth decade of life.
Cluster attacks are strictly unilateral, although the headache may alternate sides. The pain is excruciatingly severe. It is located mainly around the orbital and temporal regions, although any part of the head can be affected. The headache usually lasts 45–90 min, but can range from 15 min to 3 h. It has an abrupt onset and cessation, and attacks are accompanied by cranial autonomic symptoms. Migrainous symptoms, such as nausea, vomiting, photophobia and phonophobia, are seen in significant proportions of cluster patients,[5
] and aura has also been reported.[7
] The vast majority of CH patients report restlessness or even aggressiveness during the attacks[5
] and, therefore, this feature has been incorporated into the ICHD-II diagnostic criteria.[8
] The condition can have a striking circadian rhythmicity, with some patients reporting that the attacks occur at the same time each day.
Alcohol, nitroglycerine, exercise and elevated environmental temperature are recognized precipitants of acute cluster attacks. Alcohol induces acute attacks, usually within 1 h of intake, in the vast majority of sufferers, contrasting with migraine sufferers who generally have headache some hours after alcohol intake. Alcohol triggers attacks during a cluster bout, but not in a remission.
CH is classified according to the duration of the bout. About 80–90% of the patients have episodic cluster headache (ECH), which is diagnosed when they experience recurrent bouts, each with a duration of more than a week and separated by remissions lasting more than 4 weeks. The bouts typically occur once or twice a year. The remaining 10–20% of the patients have chronic cluster headache (CCH), in which either no remission occurs within 1 year or the remissions last less than 1 month.[8
The major differential diagnostic considerations are the other TACs  and secondary causes of CH. The vast majority of CH patients have a primary headache syndrome, with symptomatic causes only being identified in a very small minority. However, the true prevalence of symptomatic causes of CH is unknown as there are no prospective population-based neuroimaging studies. A review of retrospective case reports published in the medical literature suggests that the TACs may be associated with pituitary tumors, although this most likely reflects a considerable element of publication bias.[9
] Similarly, an observational study of headache disorders in patients with pituitary tumors reported that CH occurred in 4% and SUNCT in 5%, but the study was conducted in a tertiary referral neurosurgical center and, therefore, does not give a meaningful indication of the prevalence of these headaches in patients with pituitary disorders.[10
] It remains unclear whether every TAC patient requires neuroimaging, although, if it is considered, then magnetic resonance imaging (MRI) is the preferred modality. Some authors suggest that all patients with TACs should have dedicated pituitary imaging. However, approximately one in 10 of the general population has an incidental pituitary microadenoma (<1 cm diameter) on routine MRI, and up to one in 500 will have a macroadenoma.[11
] This approach is therefore likely to identify a significant number of incidental lesions, which could then be erroneously considered to be the cause of the TAC syndrome. We suggest that all TAC patients should be carefully assessed for pituitary disease-related symptoms and that further investigations with MRI of the pituitary gland should be undertaken in patients with atypical features, abnormal examination or those resistant to the appropriate medical treatments.
Clinical features of the trigeminal autonomic cephalalgias
Subcutaneous sumatriptan 6 mg is the drug of choice as abortive treatment of a cluster attack.[12
] In CH, unlike in migraine, subcutaneous sumatriptan can be prescribed at a frequency of twice daily, on a long-term basis if necessary, without risk of tachyphylaxis or rebound.[13
0Inhalation of 100% oxygen, at 7–12 L/min, is rapidly effective in relieving pain in the majority of sufferers.[15
] It should be inhaled continuously for 15–30 min via a nonrebreathing facial mask. However, up to 25% of the patients note that oxygen simply delays the attack for minutes to hours rather than completely aborting it.[15
Sumatriptan nasal spray (20 mg) and zolmitriptan nasal spray (5 mg and 10 mg) are both more effective than placebo.[18
] Given the efficacy of both zolmitriptan 5 mg and 10 mg doses, it has been advised that 10 mg might be the optimal initial dose for those with very severe attacks occurring only once per day or every other day, while 5 mg should be the initial dose for those with more frequent attacks or poor tolerability.
Lidocaine solution, given as nasal drops (10% lidocaine solution) or a spray deep in the nostril on the painful side, has been reported to give mild to moderate relief in patients during a CH attack, although only a few patients obtain complete pain relief.[21
] Therefore, intranasal lidocaine serves as a useful adjunct to other abortive treatments, but is rarely adequate on its own.
Dihydroergotamine nasal spray
Dihydroergotamine (DHE) nasal spray 1 mg has been studied in a double-blind, placebo-controlled, crossover trial.[23
] There was no difference in the headache frequency or duration, but the pain intensity was significantly reduced with DHE compared with placebo. The dosage used (1 mg) was rather low; therefore, DHE nasal spray at a dose of 2 mg or 4 mg may be more effective than 1 mg, although this needs to be studied in a controlled fashion.
A novel inhaled formulation of DHE (MAP0004), which has a comparable time to peak concentration and area under the curve with intravenous DHE, has been shown to be effective in aborting migraine attacks in a recent phase 3 double-blind placebo-controlled trial.[24
] Because of the restricted choice of acute treatments for CH attacks, MAP0004 should be considered in future CH trials.
There can be a lag of several days to a few weeks before the efficacy of preventive treatments becomes apparent. Transitional treatments, which produce a rapid suppression of the attacks for a limited period of time or cannot be used for prolonged periods, can be used when waiting for the beneficial effect of a preventive treatment to become evident. Transitional treatments can be also be used in patients with ECH to treat relatively short bouts (≤1 month), without the need to start a preventive drug.
Several investigators have reported the beneficial effect of oral or parenteral corticosteroid regimens in the treatment of CH.[25
] The methodological quality of these studies is low, with uncontrolled and inconclusive studies being the norm. However, these studies have nearly uniformly reported positive treatment effects, and this is consistent with the clinical experiences of most physicians caring for CH patients.[27
] Caution has to be exercised in their use because of the potential for serious side-effects. Thus, a tapering course of prednisone or prednisolone for 3 weeks is prudent. Unfortunately, relapse almost invariably occurs as the dose is tapered. For this reason, steroids are used as an initial therapy in conjunction with preventives, until the latter are effective. We start patients on oral prednisone 1 mg/kg, to a maximum of 60 mg a day, for 5 days and thereafter decrease the dose by 10 mg every 3 days.
Greater occipital nerve block
A double-blind, placebo-controlled study of suboccipital injection with a mixture of rapid- and long-acting betamethasone have been performed in CH.[28
] The authors studied 16 ECH and 7 CCH patients. Eleven of 13 (85%) CH patients treated with betamethasone suboccipital injection became pain-free within 1 week compared with none of the 10 patients treated with placebo injection. This effect was maintained for at least 4 weeks in the majority of the patients. Given the relatively good evidence of efficacy, suboccipital steroid injection can be considered in the treatment of CH.[29
Repetitive intravenous DHE administered to inpatients over a period of 3 days was reported to be very useful in some cases of both ECH and CCH. In a study of 54 patients with intractable CH (23 episodic, 31 chronic), the open-label use of repetitive intravenous DHE rendered all patients headache free.[30
] At 12-month follow-up, 83% and 39% of the patients with ECH and CCH, respectively, remained free of headache. A retrospective analysis evaluated the efficacy and safety of intravenous DHE for the treatment of refractory CH in 70 patients,[31
] and showed a complete resolution of the pain at 1 month after treatment in 62% of the cases, partial improvement in 14% and failure in 24%. Side-effects were transient and well tolerated in most patients.
The preventive agents used include verapamil, lithium, topiramate, methysergide, gabapentin, melatonin and valproate. Verapamil is the first-line agent of choice. Second-line agents include lithium and topiramate, while methysergide is a reasonable choice for a third-line agent. provides an overview of the recommendations for CH preventive treatments by the European Federation of Neurological Societies
] and the American Academy of Neurology (AAN).[29
Preventive treatments of cluster headache
Verapamil is the preventative drug of choice in both episodic and chronic CH.[33
] Dosages commonly employed range from 240 mg to 960 mg in divided doses (two or three times a day). Verapamil can cause heart block by slowing conduction in the atrioventricular node. Observing for PR interval prolongation on ECG can monitor potential development of heart block. There is only one formal guideline in the literature for the titration of the verapamil dose.[35
] After performing a baseline ECG, patients are usually started on 80 mg tds and, thereafter, the total daily dose is increased in increments of 80 mg every 10-14 days. An ECG is performed prior to each increment. The dose is increased until the cluster attacks are suppressed, side-effects intervene or the maximum dose of 960 mg daily is achieved. ECG monitoring should be performed periodically in patients on long-term verapamil.
Lithium is an effective agent for CH prophylaxis, although the response is less-robust in ECH than in CCH.[33
] Most patients will benefit from dosages between 600 mg and 1200 mg daily or at plasma concentrations comprised between 0.8 mEq/L and 1.0 mEq/L. Renal and thyroid function tests are performed prior and during treatment in view of the long-term risk of hypothyroidism and nephrogenic diabetes insipidus.
Five open-label studies have reported the efficacy of topiramate in the preventive treatment of CH.[37
] The dose of topiramate used in these studies ranged from 25 mg to 250 mg daily. The side-effect profile of this agent, including cognitive slowing and depression, often limits its use.
Methysergide has long been used for the treatment of CH.[25
] It is an ideal choice in patients with short cluster bouts, which last less than 4–5 months. Doses up to 12 mg daily can be used, if tolerated. Prolonged treatment has been associated with fibrotic reactions (retroperitoneal, pulmonary, pleural and cardiac), although these are rare.[22
] We advise a 1-month holiday every 6 months of therapy and check for evidence of pulmonary, cardiac, renal or abdominal pathology yearly if repetitive courses of treatment are required over a prolonged period.
Other preventive treatments
In a double-blind, placebo-controlled trial of melatonin 10 mg, five of 10 subjects randomized to melatonin were rendered pain free within 5 days, while none of the 10 subjects taking placebo derived any benefit.[43
] Recently, Peres and Rozen[44
] reported two CCH patients inadequately managed on verapamil 640 mg daily who were rendered pain free with add-on therapy with melatonin 9 mg daily. The authors concluded that melatonin could be a useful adjunctive treatment for CH prophylaxis.
Gabapentin was tried at the dose of 900 mg/day in an open-label fashion in eight ECH and four CCH patients.[45
] All patients were rendered pain free within 8 days of initiating therapy. Patients with ECH discontinued gabapentin after 60 days of treatment without recurrence of the attacks. The four CCH patients remained pain free at follow-up of 4 months. This astonishingly high response rate needs to be reproduced in controlled trials.
Surgical options are the measures of last resort in medically intractable patients, and should only be considered when the pharmacological options have been exploited to the fullest.[49
] Historically, destructive procedures, like trigeminal sensory rhizotomy and radiofrequency trigeminal ganglio-rhizolysis, have been tried in CH.[50
] However, they are associated with considerable morbidity and therefore have been largely abandoned. Neurostimulation therapies that entail peripheral or central nervous system targets are emerging as very promising approaches.
Occipital nerve stimulation
Studies conducted in small cohorts of medically intractable CCH patients have shown occipital nerve stimulation (ONS) to be a promising therapy.[52
] Magis and colleagues[52
] treated eight patients with medically intractable CCH using unilateral ONS. After a mean follow-up of 15 months, two patients were pain free, three patients had a 90% reduction in attack frequency while two patients had improvement of around 40%. Interruption of ONS was followed within days by recurrence and increase of attacks in all improved patients. Burns and colleagues[53
] treated 14 patients with medically intractable CCH using bilateral ONS. At a median follow-up of 17.5 months, 10 of the 14 patients reported improvement that was sufficiently meaningful for them. Subjective self-reporting of improvement was 90% or more in three patients, 40–60% in three patients and 20–30% in four patients. Benefit from stimulation was not immediate, with maximal effect noted after several months.
Hypothalamic region deep brain stimulation
Based on the finding of ipsilateral posterior inferior hypothalamic activation in CH, various centers have treated intractable chronic CH patients by electrode implantation and stimulation of this region.[54
] Leone and colleagues[57
] have recently reviewed the results of hypothalamic region deep brain stimulation (DBS) in 38 patients; 23 patients (61%) were rendered pain free or almost pain free. In most patients, the headaches recurred when the stimulation was stopped. A French multicenter, randomized, double-blind, crossover study enrolled 12 patients, with only 11 undergoing surgery. Results were not encouraging in the blinded cross-over phase.[58
] However, the crossover assignment was very short (only 1 month) and, in the open phase, six of 11 patients were considered responders.
CCH is a devastating illness and, given the low morbidity of ONS and the relatively consistent outcomes, one might argue that this modality should be explored before DBS, which is associated with a small risk of morbidity and mortality.