Cluster Headache

Synonyms: migrainous neuralgia; histamine headache; 'alarm clock headache'; ciliary neuralgia; hemicrania neuralgiformis chronica; Horton's headache; petrosal neuralgia; Bing's erythroprosopalgia; suicide headache

Recognised over 100 years ago, this condition is different from migraine - clinically, aetiologically and genetically. It is a disorder producing severe unilateral pain, localised in or around the eye and accompanied by ipsilateral autonomic features. It is quite rare, often misdiagnosed and frequently poorly managed.

Although short-lasting, the extremely painful nature of the headache causes patients very great distress and has earned the name 'suicide headache'. It is described as one of the most painful conditions known to man. The clustering of attacks means that there is significant interruption to life and work. The underlying mechanism of headache is poorly understood and prophylactic treatments are therefore empirical.

Episodic cluster headaches (CHs)

These are CHs occurring in periods lasting from seven days to one year, separated by pain-free periods lasting a month or longer. Cluster periods usually last between two weeks and three months.

Chronic CHs

These are defined as CHs occurring for one year without remissions or with short-lived remissions of less than a month. Chronic CH may arise de novo or develop from episodic CH.

Epidemiology

• A meta-analysis in 2008 found widely differing prevalence figures resulting from different research methodology. They concluded that the pooled data showed: 
• About 1 in 1,000 people suffer from cluster headache.
• It is a disease of long, sometimes lifelong duration.
• More men than women are affected, particularly in chronic CH. The overall sex ratio was 4.3 (male-to-female).
• The male-to-female ratio was higher in chronic CH (15.0) compared with episodic CH (3.8).
• The ratio of episodic vs chronic CH was 6.0.
• The analysis revealed a relatively stable lifetime prevalence.
• The condition usually begins between the ages of 20 and 40 years but can start at any age.
• Progression from episodic to chronic CH is associated with late onset, occurrence of sporadic attacks, higher frequency of cluster periods and shorter periods of remission.
• Head injury, cigarette smoking and alcohol use are also associated with a worse prognosis.

Pathophysiology

It is still not known why some individuals experience CH or why they exhibit the characteristic periodicity. However, progress is being made and a number of theories are plausible and may help progress with new treatments. Briefly:

• A small number of family and twin studies have shown that genetic factors are important. First-degree relatives of CH patients are more likely to have CH than the general population. About 1 in 20 of those with CH have another family member who also has CH. A significant association between the HCRTR2 gene and the disease has been reported.
• Positron emission tomography during cluster attacks has revealed that the ipsilateral hypothalamic grey matter is the area most activated.
• CH is generally considered a neurovascular headache comprising hypothalamic dysfunction associated with vascular changes. It is thought that hypothalamic dysfunction causes dilatation of cranial vessels mediated by trigeminoparasympathetic reflexes.
• Altered melatonin levels in CH have been found. This finding suggests a mechanism for CH periodicity and some useful therapeutic interventions. A small double-blind placebo-controlled trial shows melatonin is effective in CH prevention.
• It is possible that biological oscillators within the hypothalamus mediate seasonal changes which are linked to photoperiodic changes.
• These theories offer interesting explanations of the periodicity of CH and suggest further novel therapeutic options.
• It has been suggested that the relationship of melatonin to the light/dark cycle makes light therapy a possible therapeutic option.

Clinical features

It is important to distinguish between cluster attacks (individual episodes of pain) and cluster bouts (the time over which recurrent attacks occur). It is also important to distinguish between episodic CH and chronic CH.

Pattern of occurence

• Headaches typically occur in bouts which last 6-12 weeks, once a year or two years, often at the same time each year.
• The headache typically occurs at night, 1-2 hours after falling asleep, although this is not always the case.
• A circadian pattern is accompanied in 85% of patients (episodic CH) with remissions of 1-4 years.
• About 10-15% of patients have CH chronically with no remission. Often the interval between bouts is the same and there is a tendency for the interval to lengthen with age.
• 10% of those with episodic CH go on to develop chronic CH.
• About a third of those with chronic CH become episodic.

Nature of symptoms

• The pain comes on rapidly (without aura) over about 10 minutes.
• The pain maintains an intensity, is excruciating, sharp and penetrating (not pulsatile as with migraine).
• The pain is centred around or behind the eye, temple or forehead, although the neck and other parts of the head can be involved.
• Pain is unilateral and mostly stays on the affected side with each attack.
• It typically lasts from 45-90 minutes (range is 15 minutes to 3 hours).
• Attacks of pain occur once- or twice-daily (occasionally more often, even up to 8 times daily).
• Wakened patients may beat their heads against the wall in distress.
• Associated autonomic features of ipsilateral lacrimation, rhinorrhoea, nasal congestion, eyelid swelling, facial sweating or flushing and conjunctival injection and a partial Horner's syndrome with miosis and ptosis may be present: two or more in the presence of the extremely severe periocular headache will secure the diagnosis.
• Nausea may accompany the pain, but is much less of a feature than with migraine.
• Sufferers, unlike with migraine, cannot keep still and are described typically as restless.
• Patients pace around, occasionally banging their heads on walls and furniture.

Triggers

• Alcohol is a potent precipitant of attacks. However, in episodic CH, normal alcohol consumption can be resumed once the cluster period is over.
• Histamine and nitroglycerine are also provokers of attacks in chronic CH and during cluster periods in episodic CH.
• For some patients, heat, exercise and solvents can precipitate attacks.
• Disruption to sleep patterns (for example, by shift work, jet lag, etc) can also exacerbate or trigger CHs.

Diagnosis

The diagnosis is made from the history. The International Headache Society (IHS) guidelines have suggested the following diagnostic criteria:

• At least five attacks fulfilling the criteria below.
• Severe, or very severe, unilateral orbital, supraorbital and/or temporal pain lasting 15-180 minutes if untreated.
• Headache accompanied by at least one of:
• Ipsilateral conjunctival injection and/or lacrimation.
• Ipsilateral nasal congestion and/or rhinorrhoea.
• Ipsilateral eyelid oedema.
• Ipsilateral forehead and facial sweating.
• Ipsilateral miosis and/or ptosis.
• A sense of restlessness or agitation.
• Attacks occur from one every other day to eight times daily.
• Not attributable to another disorder.

Differential diagnosis

This could include a longer list of causes of headache but those most similar to CH in the IHS guidelines are:

• Paroxysmal hemicrania.
• Short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT).
• Probable diagnoses (of CH, SUNCT and paroxysmal hemicrania) where insufficient attacks have occurred or diagnostic criteria are not fulfilled.

Investigations

The diagnosis is usually made from history and neurological examination. As with any primary headache, some patients may need imaging, and the red flags of headache indicating the need to search for a secondary cause are:

• Change in pattern of headache.
• New headache at age over 50.
• Onset of seizures.
• Headache with systemic illness.
• Personality change.
• Symptoms suggestive of raised intracranial pressure (morning headache, headache with coughing, sneezing, straining).
• Acute onset of the worst headache ever (possible intracranial aneurysm).

Management

It should be remembered that:

• Some patients will have found medical interventions unhelpful or hard to tolerate.
• Patients may not have had the benefit of up-to-date review of treatment.
• Patients may be depressed or despondent about the condition.
• Patients experiencing their first attacks will be greatly distressed and will need reassurance.
• Drug treatment is always necessary for effective control.
• The realistic aim of treatment is suppression of the attack and reduction in frequency and severity of attacks. There is no likely prospect, at present, of curative medical treatment.
• Prophylaxis should begin as soon as possible after the start of a new cluster period, as there is evidence that it is most effective at this point.
• Failure of one drug does not predict failure of another.
• All treatments are potentially toxic and shared risk/benefit evaluation is an essential part of management.
• Partial relief presents a dilemma, as using two drugs together is potentially of greater toxicity that one alone, but it may be hard for patients to sacrifice what has been gained in order to try another treatment which may be more or less effective.
• When benefit is not seen within one week of reaching the maximum dose of a drug then it should be changed or supplemented.
• Acute therapy may be used in addition to prophylactic therapy if breakthrough attacks continue.

It is therefore important to:

• Establish rapport, based on knowledge and understanding of the condition generally and the patient's particular experience.
• Anticipate that polypharmacy is likely to be required.
• Be prepared to follow up closely to monitor efficacy of treatments, side-effects, etc.
• Consider drawing up a programme of measures, perhaps with a patient-held record book or diary.

General advice

• Be prepared for attacks. Patients should be encouraged to have both acute and preventative treatments available. This may involve completion of Home Oxygen if oxygen is to be used.
• Stop smoking, as this can increase the risk of chronic CH developing.
• Abstain completely from alcohol during periods of CH and in chronic CH.
• There is no evidence that abstinence from smoking during attacks affects CH.[14] 
• Maintain a regular sleep routine and good sleep hygiene (avoiding tea, coffee, etc).

Acute attack

Sumatriptan - by subcutaneous injection - and oxygen are likely to be the mainstay of treatment for most patients:

• Sumatriptan - 6 mg subcutaneously is effective. The best evidence for benefit is from sumatriptan subcutaneously. However, for regular attacks over several weeks, there is a danger of overdosage. Other acute therapies like oxygen should be used in addition to preventative measures.
• Sumatriptan nasal spray - works less well for most patients (use when the subcutaneous route is unacceptable). It is not licensed for use in CH.
• Oxygen - 100% oxygen, given for 15 minutes up to five times per day, is safe and effective in 80% of cases. It is given by a tight-fitting mask (not a standard mask or nasal cannulae) at 12 L/minute.[3] It is particularly useful for night attacks. The most convenient form is the 460 L CD oxygen cylinder (lasts about 40 minutes at high flow rates) with concentration venturis. HOOF forms will need to be completed for oxygen supply. Useful guidance for patients is available on the OUCH (UK) website.

Other possible treatments for acute attacks include:

• Ergotamine - of limited use in view of its poor oral absorption and side-effect profile. Before the advent of sumatriptan it was used more often.
• Anti-inflammatories - such as indometacin which can be used.
• Metoclopramide - may be useful as an adjunct to acute treatments.
• Lidocaine - 1 ml of 4% lidocaine, which can be given intranasally to the affected side.

Prophylaxis

Verapamil

• This should be considered for prophylaxis of CH. It is the first-line choice for both episodic and chronic CH.
• It is started at doses of 40 mg twice-daily, building up to as much as 960 mg daily.
• Side-effects of constipation and flushing may limit use in some.
• ECG monitoring (for AV block) is required at doses over 120 mg daily and fortnightly ECG monitoring is required with successive dose increases (because of the risk of dysrhythmias).
• Some experts believe that standard-release verapamil formulations are more effective in CH than the modified-release versions.
• If unfamiliar with this use of verapamil, or if the patient does not respond to treatment, the GP should seek specialist advice.

Prednisolone

• This may be preferred, as unlike other treatments, it is started at full dose. A starting dose of 60-100 mg, once-daily for 2-5 days is recommended.[14] If this treatment works it usually does so very quickly.
• The treatment's effectiveness is thought to operate through an effect on trigeminal activation and on altered melatonin metabolism related to hypothalamic dysfunction.
• Treatment dose must be swiftly reduced after 2-5 days, in 10 mg increments every 2-3 days, so that treatment is discontinued in 2-3 weeks.
• Relapse may occur as the dose is reduced - second and even third courses may be used with caution.
• Prednisolone may be a helpful add-on acute treatment for breakthrough attacks when prophylaxis has been commenced.
• Short courses of this nature do not seriously risk suppression of endogenous steroid production.

Lithium

• This should be considered if verapamil is not effective.
• It is most useful for chronic CH in doses of 600-900 mg daily.
• It is less effective in episodic CH, where doses of 800-1600 mg daily are needed, and serum concentrations may reach the range 1.0-1.4 mmol/L. Tolerance may develop.
• Serum concentrations must be regularly monitored, as must renal, thyroid and cardiac function, and symptoms of lithium toxicity (polydipsia, polyuria, nausea, diarrhoea) may mean its abandonment. Non-steroidal anti-inflammatory drugs (NSAIDs) should not be taken concomitantly.

Melatonin

• There is evidence of dysfunction in melatonin metabolism in acute CH.
• Melatonin is a possible addition to the list of prophylactic agents for nocturnal attacks as benefit has been seen in some small studies.
• It is taken at night at doses of 7.5-10 mg but, with currently available formulations, administration has to be timed to half an hour before the anticipated attack (difficult). Slow-release preparations may be more effective, but this is speculative until more evidence is forthcoming.

Topiramate
This may be useful but only under specialist supervision and only in chronic CH.

Sodium valproate
This has also been used for prophylaxis, usually in chronic CH.

Ergotamine

• This can be used two hours prior to a predicted attack: 2 mg orally or 1 mg rectally.
• For unknown reasons CH patients appear relatively resistant to the toxic side-effects of ergotamine which limit its use in migraine. However, it is a potent vasoconstrictor and is contra-indicated in hypertension, vascular disease and patients with multiple risk factors for vascular disease.

Methysergide
This drug has largely fallen out of use due to the small but serious risk of retroperitoneal fibrosis. Under specialist care up to 6 mg in divided doses can be used for CH: it is most useful in the episodic form (for fewer than six months at a time).

Nifedipine
This can also be useful in the chronic form.

Deep brain stimulation

• This may have a place in intractable chronic CH. Limited reports suggest complete relief and apparently few side-effects.
• The ipsilateral posterior hypothalamus is targeted for electrical stimulation.

Surgery
Trigeminal nerve blockade gives temporary relief if medical therapy is not helping.

More invasive procedures
These are used as a last resort. They involve chemical or physical ablation to parts of the trigeminal nerve. These can be effective but are likely to be reserved for refractory chronic CH. Cases require referral for consideration of neuromodulation and more invasive treatments.

Alternative therapies
Therapies such as acupuncture have anecdotally been very helpful to some patients. Unfortunately, the evidence of benefit is poor. A Bandolier review concluded that there was no evidence of benefit and better trials are needed.