Atypical Presentations of Cluster Headache

Abstract

Recently, cluster headache has been shown to occur with aura, suggesting that as more cluster patients are seen by headache specialists new forms of this well-defined primary headache syndrome will be identified. This study presents three atypical presentations of cluster headache: persistent or unremitting cluster, periodic cluster, and reflex or event-related cluster. Case reports are presented with an explanation as to why these headaches should be considered cluster headache.

Introduction

Recently, cluster headache has been shown to occur with aura (1), suggesting that as more cluster patients are seen by headache specialists new forms of this well-defined primary headache syndrome will be identified. This study presents three atypical presentations of cluster headache. Case reports are presented with an explanation as to why these headaches should be considered cluster headache.

Persistent or unremitting cluster

Cluster headaches are discrete short-lasting events. Headache duration is anywhere from 15 min to 3 h. In between individual attacks patients are either pain-free or experience a very mild twinge of pain on the side of the head where their cluster headaches arise. Two patients are presented who at all times had a severe unilateral headache which never ceased and was accompanied by pain exacerbation periods that mimicked typical International Headache Society (IHS) (2) diagnosed cluster headache attacks.

Patient 1

A 58-year-old woman presented with a 2-year complaint of a continuous left-sided headache. The pain was in a retroorbital location with some radiation to the temple. She had no pain-free time and her every moment pain intensity was 8 out of 10 (VAS scale). At least two times per day the pain would exacerbate to an excruciating level (10 out of 10). This exacerbation period would last 1–3 h and would typically occur in the late afternoon hours and 1–2 h after falling asleep. Her every moment pain had no associated symptoms. Her pain exacerbation periods were marked by photophobia, phonophobia, left eye ptosis, eyelid oedema and left eye lacrimation. She denied nausea, vomiting, conjunctival injection, or nasal congestion/rhinorrhoea. During the pain exacerbation periods she would become very agitated and unable to sit still because the pain would worsen. Once the exacerbation period ended her sense of agitation would cease. Past abortives included OTCs without relief; she was never on a preventive agent. She was never a smoker. Alcohol ingestion could trigger a painful exacerbation peak. Initially the patient was felt to have hemicrania continua (HC) because of the constant unremitting pain, although at a higher intensity than typically seen in HC. Indomethacin could not be utilized because of a past history of bleeding ulcers. High-dose rofecoxib was tried, 75 mg/day, without relief. Rofecoxib has shown some effect in HC in patients intolerant of indomethacin (3). During her second office visit she developed a painful exacerbation peak which was quickly alleviated with oxygen. Oxygen actually alleviated all of her pain, even her baseline pain, but the baseline pain returned within 1 h. The patient has since used oxygen multiple times and it always alleviates the pain exacerbation periods. A prednisone taper was administered but without pain relief. Verapamil was begun as a preventive (dosage raised to 720 mg/day) and significantly reduced her baseline pain intensity and also reduced her exacerbation pain intensity as well as the length of her exacerbation periods from 1–3 h to 15 min. High-dose valproic acid (3000 mg/day) was added to the verapamil and her headaches ceased. Brain magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) were normal.

Patient 2

A 74-year-old man presented with a 6-year complaint of a daily, severe (7 out of 10) left-sided headache in a retroorbital and temple location, which was always present. One to two times per day the pain would peak for 2–3 h and was extremely severe (10 out of 10). During the pain peaks he would develop left eye ptosis, conjunctival injection, lacrimation and nasal congestion. He would also become very agitated during the exacerbation periods, but this would cease when the pain went back to its baseline intensity. Only on occasions would the pain awaken him from sleep. He did have a past history of smoking but had stopped many years prior to headache onset. He failed numerous abortives including oxygen, and found relief only with daily sumatriptan injections, which he would use for the pain exacerbation periods. Sumatriptan would end his pain spikes but he would be left with the baseline pain. His age and history of hypertension were contraindications for sumatriptan, but he refused to stop the medication because he could not withstand the pain. Brain MRI and MRA were normal. An indomethacin trial for HC failed (up to 150 mg/day with no response). He was started on verapamil and at a dose of 640 mg/day he became pain-free. He then developed chest pressure and had to stop the verapamil. The headaches returned. On lithium carbonate 600 mg/day the headaches again alleviated.

In the described patients, the headache exacerbation periods met International Headache Society (IHS) criteria for cluster headache. The cases should be considered an atypical presentation of cluster because of the daily, unremitting, unilateral, high-level intensity pain that never ceased, even between true cluster attacks. Cluster headache is marked by pain-free time in between attacks or a very mild intensity interictal baseline pain. Response to verapamil and lithium suggested the case patient's headaches were cluster in origin and the one patient's response to oxygen also intimated a cluster phenomenon.

A daily, constant, unilateral headache that is associated with pain exacerbation periods with migrainous or cluster-like features is a hallmark of hemicrania continua (4). In HC, the baseline pain is typically of mild to moderate intensity (4) and is rarely bothersome to the patient. In many instances a diagnosis of HC is missed by physicians because the patients focus only on their headache exacerbation periods and rarely if ever mention they have daily unremitting pain, because the pain is of low intensity. HC was in the differential diagnosis for both case patients. Indomethacin could not be initiated in patient 1 because of a past history of bleeding ulcers. She did, however, fail a trial of rofecoxib which has shown effect in HC (3). Her high-intensity baseline pain was atypical for HC and her response to verapamil and valproic acid suggested against a diagnosis of HC. Patient 2 had no response to indomethacin at a dose of 150 mg/day. Patients with HC may require doses of indomethacin> 150 mg but typically will have at least some partial benefit on this amount of indomethacin. Patient 2 responded to verapamil and lithium, suggesting against a diagnosis of HC.

Santonja et al. (5) reported a case of ‘continuous cluster’ similar to but not exactly like the above cases. A 43-year-old man presented with continuous left orbital pain with conjunctival injection of 3 years duration. On examination he had persistent ptosis and miosis. In this patient the pain was moderate to severe without aggravating factors. The headache was minimally responsive to indomethacin but he intermittently became pain-free after sumatriptan injections. Unlike the presented case patients, the Santonja et al. (5) patient had persistent autonomic symptoms and pain, suggesting a different disorder but a variation on the same theme.

Conceptually, ‘persistent or unremitting cluster’ could be caused by a hypothalamic generator that never turns off, constantly producing high-level pain with exacerbation periods of true cluster headache.

Periodic or intermittent cluster

The IHS criteria for cluster headache require at least one attack every other day. It is actually rare to see untreated patients who do not have at least one headache per day. A patient is presented who experienced on average one single cluster headache per year for many years and more recently one headache per month over the last year, with the remainder of the time being pain-free.

A 48-year-old woman with a history of Marfan's syndrome began to develop stereotypic headaches in her early twenties. Initially the headaches were very infrequent, occurring only one or two times per year (single individual headache) for about 20 years. Over the most recent year the headaches increased to one single headache per month and rarely three or four headaches in a month. Each headache would awaken her from sleep usually between 2 and 4 a.m. The headaches were 100% right sided in a retroorbital and temple distribution. The pain was extremely severe and would last on average between 2 h and 4 h. During the headache she would develop right eye ptosis and lacrimation as well as right nasal congestion. She would become very agitated and would press her head up against a wall or take a belt and wrap it around her head and tighten it to try to alleviate her pain. She denied any nausea, vomiting, photophobia or phonophobia. She never developed a headache during the daytime. There was no strict pattern as to when the headaches would occur. Her examination was normal except for a decreased palpebral fissure of the right eye. No true ptosis or miosis was noted.

The patient's headache was considered most consistent with cluster headache, but the frequency was not at all typical of cluster (only a single cluster headache per year for many years and now on average one headache per month). She was placed on melatonin 9 mg at bedtime and has been headache-free for 5 months. Because of the change in headache frequency and a history of Marfan's, a brain MRI/MRA was completed as well as an MRA of the cervical carotid arteries. The study revealed two small saccular aneurysms arising from the distal cervical right internal carotid artery suggesting the sequela of a prior dissection; the remainder of the study was normal. As the dissection was on the same side as the headaches it could have been the initial cause of the headaches or the reason the headache frequency recently changed. How, though, a carotid dissection could produce an episodic headache syndrome is difficult to explain.

The differential diagnosis for nocturnal or sleep-induced headaches includes cluster, migraine and hypnic headache. The patient had a unilateral severe headache of short duration with associated autonomic symptoms and restlessness, which were all consistent with a diagnosis of cluster headache. The frequency of the headaches (much less than one every 2 days) was the only thing that excluded the headaches from IHS diagnosed episodic cluster (3.1.2). The duration of the headaches and the fact that they only occurred with sleep suggested against migraine as a diagnosis. Hypnic headaches only occur during sleep, are typically 15–180 min in duration and can be bilateral or unilateral pain (6). Unlike the presented patient's headaches, hypnic sufferers do not develop autonomic symptoms in most instances and they normally suffer with at least half the nights per month with headache. Overall the described patient has very infrequent cluster (periodic or intermittent cluster), possibly secondary to a past carotid dissection. This case patient did not meet criteria for IHS diagnosed episodic cluster but could be considered IHS 3.1.1 cluster headache periodicity undetermined.

Reflex or event-related cluster

Valsalva-induced cluster

Cluster headaches typically occur spontaneously with no needed triggers, although alcohol and nitroglycerin, for example, can induce a cluster headache when a patient is in cycle. A patient is presented whose episodic cluster headaches only manifested following valsalva manoeuvres such as coughing or sneezing and never occurred spontaneously. When out of cycle the same valsalva triggers would not cause a cluster headache. Unlike typical cluster, these headaches were completely abolished with daily indomethacin.

An 86-year-old man was referred for a 10-year history of severe, cough-induced headaches. He had three to four headaches daily, with each headache typically lasting 4 h, although they could be as short as 30 min. The headaches were all stereotyped: 100% right-sided in a forehead–temple distribution with some radiation to the right ear and occipital-nuchal area. The pain was described as sharp and unbearable and was accompanied by autonomic manifestations such as unilateral right eye tearing, conjunctival injection, and nasal rhinorrhoea. There was no evidence of ptosis, nausea, vomiting, photophobia or phonophobia. Rest or lying down exacerbated the symptoms and the patient found himself constantly moving about during an attack. Curiously, the headaches never occurred spontaneously but only after the patient would cough, sneeze, or bend over. Although every valsalva manoeuvre did not elicit a headache, every headache was triggered by the patient performing one of the above actions. The headaches occurred in cycles, each lasting for several months. In between cycles, he would be headache-free. During his remission periods he could cough, sneeze, or bend over without any induction of a headache.

General and neurological examinations were normal for age, with a normal fundoscopic examination. Imaging studies included an MRI of the brain, with and without gadolinium, the results of which were normal for age. Although an MRA study of the extracranial carotid arteries revealed a possible 75% stenosis of the right carotid artery, subsequent Doppler studies failed to reveal a significant stenosis (16–49%). MRA of the intracranial circulation was normal.

Because of the valsalva inducing character of the headaches, indomethacin was started at 25 mg orally each day with another 25 mg given on an as needed basis. Within several days of starting indomethacin the patient stopped developing headaches with valsalva manoeuvres. As long as he maintained his indomethacin treatment (25 mg/day) he was headache-free. One time the patient developed gastrointestinal disturbances that were thought to be indomethacin-related. When he stopped indomethacin, the valsalva-induced headaches returned within 1 week. Once indomethacin was restarted, the headaches ceased.

There are very few if any documented cases of cluster headache that occur solely after a particular trigger and never spontaneously. The case patient does have cluster headache (short-lasting, severe, unilateral headaches with autonomic symptoms which follow a circadian pattern), but it cannot be considered typical cluster as they occurred only after valsalva manoeuvres, never spontaneously, and were prevented by indomethacin, an agent not known to be an effective cluster preventive. Concerning the patient's event-related headaches, it is of note that when he was out of a cluster cycle he could perform the same valsalva manoeuvres and never trigger a headache. Indomethacin was chosen as treatment because of its success in other valsalva-triggered headaches, cough and exertional headache. Indomethacin's unique ability, compared with other non-steroidal anti-inflammatory drugs (NSAIDs), to lower CSF pressure may be why it is able to prevent valsalva-triggered headaches.

Valsalva-induced cluster headache should not be confused with benign cough headache (BCH). Although BCH is valsalva-triggered, it is characterized as bilateral pain with a duration of less than 1 min and is not associated with autonomic features. Perini et al. (7) documented a man with a unilateral headache that was induced by coughing, was refractory to dihydroergotamine, and relieved by the NSAID, nimesulide. The author termed this ‘benign cough “cluster” headache’. However, given the very short duration of the attacks (only 1–10 min) and the absence of any autonomic symptoms, this may have been something other than cluster headache.

In epilepsy, there is a unique form of ‘reflex seizures’ in which a specific sensory input is required to trigger an epileptic response and without this input the patient will never develop a seizure (8). The seizure precipitant is unique to that individual and typically only one specific trigger is recognized per patient. Known reflex stimuli include sensory modalities such as visual (see below: cluster triggered by viewing television) or a function modality such as walking or reading. Valsalva-induced cluster maybe a form of ‘reflex or event-related cluster’ or cluster headaches that only occur after a specific trigger and never spontaneously.

Cluster triggered by viewing television

Recognized precipitators of cluster headache attacks include alcohol ingestion, nitroglycerin, exercise, a warm environment and histamine. Two patients presented with a distinct complaint of cluster headaches being triggered by watching television. Each patient had chronic cluster (37-year-old male, 46-year-old female) and upon presentation stated that they had completely stopped watching television because this event would always trigger one of their attacks. Both patients related that they also experienced spontaneous cluster headaches and some attacks were precipitated by alcohol, but television viewing was a definite consistent trigger. Regardless of what time of day they viewed television, and what day of the week, an attack would always ensue usually within 30 min. If television viewing only occurred during the evening then the patients may have falsely recognized TV viewing as a trigger, as spontaneous cluster attacks normally occur during late-day hours; however, this was not the case. Both patients also related that they knew they were improving on their cluster preventives when they could watch television without developing a headache. Neither of the patients had visited a cinema during their cluster cycle, so one cannot comment on whether television alone was the precipitator of their headaches or if other viewing modalities could trigger attacks. Neither patient stated that working on a computer would precipitate a cluster headache. No other visual triggers were noted by either patient.

Why television viewing would precipitate cluster attacks is unknown. Television as a trigger for neurological events has been documented in the literature. Televised images are the most common stimulus for ‘reflex photosensitive seizures’. Televisions that emit higher levels of long-wavelength red light appear to be the ones that can induce visually triggered seizures (9). Possibly a higher level emission of long wavelength red light may in some way activate the hypothalamus of cluster sufferers, thus inducing cluster attacks. This could be studied in the laboratory utilizing cathode ray tubes that emit higher levels of long-wavelength red light and a position emission tomography scan looking for hypothalamic activation. A large number of cluster sufferers have been found to be photophobic during individual attacks and more light-sensitive during cluster periods vs. times of remission (10). Television viewing as a trigger for cluster may be more prevalent than just these two reported patients, but as this type of precipitating manoeuvre is not normally inquired about during a headache history, there is no way to determine this at present.

Conclusion

The primary headache syndromes in their typical forms have been recognized for centuries. As we have had more exposure to headache patients these syndromes have become less homogeneous and different subtypes have been recognized. Migraine has many forms and cluster headache may also have various presentations. It is important to identify the described syndromes as cluster-like, as successful treatment depends on this recognition.