Clinical manifestations

Presentation and course

IHS criteria. The third edition of the International Classification of Headache Disorders officially defines hypnic headache. It is described as “frequently recurring headache attacks developing only during sleep, causing wakening and lasting for up to 4 hours, without characteristic associated symptoms and not attributed to other pathology.” The formal criteria require headaches that 1) develop only during sleep and cause awakening, 2) occur greater than or equal to 10 days per month for greater than or equal to 3 months, 3) last greater than or equal to 15 minutes for up to 4 hours after wakening, and 4) have no associated cranial autonomic symptoms or restlessness. Additionally, the headache should not be accounted for by another ICHD-3 diagnosis. They further clarify that the headache typically starts after age 50 but may occur in younger patients, is not associated with specific stages of sleep, and may have accompanying migrainous features. MRI may show grey matter volume reduction in the hypothalamus. Important changes from the prior ICHD II classification include reduction of the number of required attacks per month from 15 to 10 and the possibility that migrainous features may be present.

Clinical observation reflects several nuances not fully enumerated in the official IHS publication. For example, patients with strictly unilateral headaches, pulsatile headaches, and more severe pain have been reported (21; 41; 19; 37; 52; 33). Although polysomnography is often recommended as part of the evaluation, hypnic headache does not occur strictly during either REM or non-REM sleep. In one small series, 16 out of 23 patients with hypnic headache also had a history of migraine, and in 11 of those 16 cases the migraines disappeared at or prior to the onset of hypnic headaches (64).

Timing. The headache occurs exclusively during sleep; it occurs during daytime hours only when associated with a nap. Some patients report awakening with the headache after a dream. Most patients (51%) experience the attacks between 2:00 A.M. and 4:00 A.M. (69). The headache occurs with striking circadian rhythmicity, with onset at the same time each night (62; 68; 16). This feature has earned it the title of “alarm-clock-headache syndrome.”

Age. Hypnic headache syndrome was originally described as a disorder confined to the elderly. In a larger review, however, the age of onset ranged from 36 to 83 years (21), with a mean of 58 years (69). The pooled mean age of onset of hypnic headache in a meta-analysis of 14 studies was 60.5 years (47). Subsequently, it was reported that hypnic headache can even begin during childhood or adolescence (28; 65; 11). A study of seven children with hypnic headache and a systematic review reported that headache features in children differed from those observed in adult patients (23). Children experienced throbbing and pulsating pain, whereas adults reported dull, pressure-like pain. Children also had lower frequency and shorter duration of attacks compared to adults. Perhaps a long delay in diagnosis is responsible for this mischaracterization; in one study the disorder was present for more than 10 years prior to diagnosis in 36% of patients (19).

Gender. Although five of Raskin’s original six patients were men, the disorder appears to occur equally, if not more commonly, in women. In a 25-patient sample from Brazil, 80% (20 out of 25) patients were women, and diagnosis took an average of 3 +/- 2.7 years from symptom onset (67). In 343 adult patients, 69% were females (69).

Duration. Hypnic headaches usually are short-lasting (30 to 120 min.) but can range from 15 minutes to 3 hours.

Frequency. Sixty four percent of patients usually experience one headache attack per night (69). Some may have recurrent bouts of two or more headaches throughout the night, causing repeated interruption of sleep (62; 55; 68). In over two-thirds of patients, the headaches occur with a high frequency of more than four nights per week; and in some the attacks occur nightly. There is some variability as patients with fewer weekly attacks have also been reported (41; 19).

Two temporal profiles of hypnic headache exist. Most sufferers have a chronic course, but some reports suggest that an episodic form, in which patients have a spontaneous or treatment-induced remissions, may exist (21; 44). A relapsing-remitting variant has also been reported at this point (61).

Intensity. Most headaches are of moderate intensity (61.5%), with up to 33% reaching the severe level (69). Only a small percentage of patients describe their pain as mild. Although headache frequency can vary, the intensity may remain relatively constant or may decrease over time. Pain intensity may be more severe in men (37).

Character. The most common sensation is a dull pain in 75%, occurring in greater than one-half of patients, and up to a third describe the pain as throbbing or pulsating 69). Rarely is it considered sharp or stabbing.

Location. In 50% to 75% of the cases, the pain is bilateral and diffuse (69; 47); however, frontotemporal and unilateral variants with rare posterior predominance have also been described.

Associated symptoms. Individuals may have a previous or coexisting headache history, but hypnic headaches are infrequently associated with migrainous features (eg, photophobia, phonophobia, nausea, or vomiting) and are rarely accompanied by autonomic features (eg, lacrimation, rhinorrhea, or ptosis). Thus, one may distinguish this disorder from cluster headache and migraines that may also occur nocturnally. Literature suggests that patients may engage in typical motor activities (getting up, drinking a cup of coffee, reading, or watching television). These activities also distinguish hypnic headache from migraine, in which patients prefer bedrest, and cluster headache, during which patients are often restless or agitated (16; 21; 44; 19).

Secondary causes. Hypnic headache is a primary headache disorder; however, in several cases presenting as hypnic headache, another secondary cause has been identified including pontine stroke (48), posterior fossa meningioma (58), nonfunctioning pituitary macroadenoma (24), growth-hormone-secreting pituitary tumor (73), symptomatic nocturnal hypertension (09), postcoiling and clipping of an intracranial aneurysm (01), and cerebellar hemangioblastoma (53). One case report describes a patient with apparent hypnic headache and basilar artery dolichoectasia who responded well to verapamil (50). Although there is an association with obstructive sleep apnea (detected in 83% and 73% of cases in two series), there is no temporal relation between oxygen desaturations and headache. Arterial hypertension is also frequently reported in hypnic headache patients, but that relationship also remains unclear (42).

Prognosis and complications

Few prospective studies of the disorder have been conducted. It has been reported that up to 17% of patients with hypnic headache have achieved spontaneous relief (21). Most sufferers, in contrast, seem to have a chronic, unrelenting condition. In one study, 71% of patients continued to have headaches; of those, headache severity decreased in four patients (33%), remained unchanged in seven patients (58%), and increased in one patient. Headache frequency was increased in four patients, decreased in three, and remained unchanged in the others. Spontaneous relief had occurred in only one patient (after 3 years). In contrast, one review describes that nearly 40% of patients reported headache freedom with treatment (42).

An episodic variant of hypnic headache has been reported (44; 61). In these cases, remissions may occur spontaneously or following a course of treatment. In one series, up to 53% of patients entered an episodic course following treatment (41).

Clinical vignette

A 64-year-old man presented with a 7-year history of headaches (26). Before that, he had been headache free and had never had headaches with alcohol consumption. He had begun to have daily headaches that were bioccipital and constant, and the headaches occurred only after he fell asleep. On several occasions, he had fallen asleep during the day and had developed headaches then as well. When present, the headaches were moderately severe, and activity had no effect. Typically, he got up and made a cup of coffee; after drinking the coffee, he went back to sleep. The entire attack typically lasted 90 minutes. He had no nausea, photophobia, phonophobia, or osmophobia. He had no cranial autonomic features. He took no regular medications and had no other medical problems and no family history of headache. On examination, he was well, he weighed 78 kg (172 lb), and his blood pressure was 125/75 mm Hg. On cranial nerve examination, the visual fields were full, and the fundi were normal. His eye movements were normal, and no facial asymmetry was present. He had no trigeminal sensory loss. Limb examination was normal. Gait and coordination were also normal. MRI of the brain with gadolinium was normal. He had overnight blood pressure measurements taken, which were normal. He agreed to an in-hospital trial of single-blinded caffeine 250 mg infusion versus placebo. On the evening of the placebo, he had a headache, and on the caffeine evening, he had no headache and slept through the night. He was diagnosed with hypnic headache and advised to take a strong cup of coffee nightly; at follow-up 8 years later, he remained headache free except on nights when he had skipped the coffee.

Biological basis

Etiology and pathogenesis

The etiology and pathogenesis of this headache syndrome is unclear. The unique features of the syndrome include the exclusive relationship of the headaches, with sleep and the tendency for the headaches to begin in mid to late life. These features imply that this headache disorder may be related to the physiology of sleep (51; 06; 07; 12; 28), and some have proposed a neurodegenerative component, but this has not been proven. Polysomnography in patients with hypnic headache has shown that attacks can occur in both REM and NREM sleep (45; 18; 41; 66; 35). Other documented factors included sleep architecture characterized by low efficiency, an increase in sleep latency, and a markedly decreased percentage of REM sleep (03; 14; 22; 60).

To date, there has been only one report of a patient in whom correction of sleep-disordered breathing with CPAP and nocturnal oxygen supplementation has improved hypnic headache (14). Other studies have found no correlation between the two conditions (14; 60; 44; 45; 41).

Though compelling evidence is lacking, several authors have hypothesized about the possible mechanism of hypnic headache, based on the presumed mechanisms of effective therapies. These theories include causative changes in cerebral blood flow and sympathetic outflow, increased intracranial pressure, and age-related changes in sleep architecture potentially related to alterations in the suprachiasmatic nucleus of the hypothalamus. To investigate this hypothesis, researchers used voxel-based morphometry in 14 patients with hypnic headache and 14 sex-matched controls. Patients with hypnic headache had decreased gray matter volume in the posterior hypothalamus, posterior anterior cingulate, operculum, and frontal and temporal lobes (32). In an earlier study, subjects with hypnic headache were reported to have a decrease in N2 latency and an increased area under the curve of nociceptive blink reflex (31). However, a study of 15 patients with hypnic headache and 15 age- and sex-matched controls found no significant differences in the nociceptive blink reflex and trigeminal pain-related evoked potentials between the two groups (30). There is also a case report of transient hypnic headache accompanying discontinuation of long-term lithium therapy, as well as one describing relief from hypnic headaches following discontinuation of an angiotensin-converting enzyme inhibitor with little to no data in regard to hypnic headache as a manifestation of medication overuse (33). Melatonin concentration changes over 20 hours seem to play no role in the pathogenesis of hypnic headache (54). One hypothesis assumes a potential role of the glymphatic system in hypnic headache (72).

Epidemiology

The hypnic headache syndrome is rare. The prevalence has been estimated at 0.07% based on a Mayo Clinic series of 19 patients seen over a 9-year period (16). A prevalence study in Iceland in 1398 adults reported a prevalence of 0.22% (20). Reports in headache subspecialty centers suggest a slightly higher prevalence of approximately 0.3% (44; 41; 19). A systematic review of 14 studies calculated the relative frequency of hypnic headache as 0.21% (95% CI: 0.13–0.35) (47). However, an accurate estimate of the incidence and prevalence of this disorder is difficult to ascertain. Because of its scarcity, the disorder is likely underrecognized and underreported, making an estimate of its true incidence and prevalence difficult (43).

Prevention

Because the etiology and pathogenesis of this disorder remain to be elucidated, those at risk cannot be identified. Therefore, preventive strategies are nonexistent, but prophylactic treatments aimed at suppressing headache attacks are feasible.

Differential diagnosis

The differential diagnosis of hypnic headache includes other episodic short-lasting headache disorders that have a predilection to occur during nocturnal sleep (58). The most reliable distinguishing features are outlined in Table 1. The importance of distinguishing these headache syndromes lies in their differential response to therapy.

Table 1. Distinguishing Features of Hypnic Headache and Other Headache Syndromes

Diagnostic workup

New-onset nocturnal headache in the elderly should always be a cause for concern, and secondary causes need to be excluded. The differential diagnosis includes giant cell arteritis, cerebral tumors, increased intracranial pressure, nocturnal hypertension, obstructive sleep apnea, other breathing disorders, depression, drug withdrawal, medication overuse, and psychiatric disorders with sleep disturbances. MRI of the brain will help exclude structural intracranial pathology, and if there is a suspicion of obstructive sleep apnea or another primary sleep disturbance, overnight polysomnography should be part of the evaluation. Goadsby recommends 24-hour ambulatory blood pressure monitoring as well, which is reasonable, as patients with nocturnal hypertension have been shown to respond to blood pressure modulation (25). Routine blood work, including a complete blood count, sedimentation rate, C-reactive protein, and chemistry profile, should exclude inflammatory and metabolic conditions that may give rise to headaches in this age group.

Management

In Raskin’s original monograph of this syndrome, all six patients described had a positive therapeutic response to lithium (62). This observation has been corroborated by several authors (55; 27; 49). The effective dose ranges between 150 and 600 mg daily in single or divided doses to a serum level of 0.5 to 1.0 mmol/L (42). The side effect profile of lithium can be prohibitive, especially in this age group (68; 16). Lithium therapy’s effectiveness has been the best studied, with 35 patients receiving the therapy, 26 of whom received benefit (22). The effects of indomethacin have also been reported, with moderate-to-good results in approximately two-thirds of cases (36; 14; 10; 08; 17). Indomethacin may be particularly beneficial in patients with hemicranial hypnic headaches or in those patients with associated autonomic features. Both indomethacin and lithium must be used with caution in the elderly population and often may be discontinued due to side effects.

Caffeine (either a cup of coffee prior to bedtime or a tablet containing 40 to 60 mg of caffeine) has been shown to have beneficial effects in preventing these nocturnal headaches (16; 34; 74). Surprisingly, this treatment seems to be very well tolerated and does not interfere with patients’ ability to fall asleep. In practice, many clinicians begin with caffeine as a first-line agent. Treatment nonresponders would then be offered the second-line therapies: sustained-release indomethacin 25 to 150 mg at bedtime or lithium 150 to 600 mg daily (40; 42).

There are limited data regarding the effectiveness of other treatments including topiramate (25 to 100 mg/day) (29; 04; 38), oxetorone (60 to 100 mg/day) (19), and sodium ferulate 150 to 300 mg/day) (59); however, there is a suggestion of benefit (39). Other medications felt to be ineffective but with conflicting evidence include tricyclic antidepressants, flunarizine, beta-blockers, verapamil, NSAIDs, prednisolone, other anticonvulsants, acetazolamide, melatonin, benzodiazepines, pizotifen, and methysergide (39; 71). There are reports of onabotulinum toxin for effective management of hypnic headache (46). One case report documented the success of an occipital nerve stimulator in treating chronic, refractory hypnic headache (70), and another patient was treated with greater occipital nerve blocks (63). Yet another case report described amelioration of hypnic headache using a mandibular advancement oral appliance (05). Lack of response has been reported with NSAIDs, oxygen inhalation, acetaminophen, and opioids (57).

Notably, many of the treatments above are from small case series or single reports with a low level of evidence, and many may have conflicting reports (33). In summary, lithium, caffeine, indomethacin, and melatonin were considered effective in 78% of 343 patients (69).

Although many patients experience headaches that are felt to be self-limited, some patients do require acute treatment options. The best efficacy has been demonstrated in caffeine and caffeine-containing analgesics with conflicting evidence in regard to triptans. Nonsteroidal anti-inflammatory (NSAID) agents, acetaminophen, flunarizine, verapamil, topiramate, beta-blockers, acetazolamide, metamizole, opioids, nimesulide, ergotamine, and high-flow oxygen are ineffective or have only very limited evidence of efficacy (42; 02).

Special considerations

Pregnancy

No data are available on the natural history of this headache disorder during pregnancy. The average age of onset is approximately 60 years of age, which is well beyond typical childbearing age.