Abstract
Introduction
The term hypnic headache syndrome was first used by Raskin to indicate a sleep-related, diffuse, pounding headache, often associated with nausea but without other autonomic symptoms, that persisted for 30–60 min and occurred at a constant time during the night in six patients, all aged over 65 years (1).
To the best of our knowledge, a total of 48 cases, three of them in Italy, have been reported in the literature since the first description (1–15).
In 1997 the diagnostic criteria for hypnic headache syndrome were proposed (7) (Table 1) and it has been suggested to add this syndrome to the International Headache Society Classification System in the section ‘Miscellaneous headaches unassociated with structural lesion’ (16).
Suggested criteria for hypnic headache
Patients suffering from this form of headache generally respond favourably to a bedtime dose of 300–600 mg of lithium carbonate (1–3, 5). Because in the age group affected by hypnic headache lithium may have marked side-effects, alternative therapeutic strategies are needed.
We report two new cases of hypnic headache syndrome in whom the administration of prednisone induced rapid and long-lasting resolution of headaches.
Case reports
Patient 1
A 68-year-old man had a 6-month history of severe, diffuse, dull pressure headaches that awakened him almost every night, between 02.00 and 03.00, sometimes during a nightmare or a dream. The headache attacks had a duration of 30–60 min but at times they lasted up to 2 h. The pain was so severe that he was forced to get up and walk back and forth relentlessly, moving untidily in tears to seek relief. Nimesulid and metamizole sodium were not effective. Nausea or other autonomic symptoms were absent. In the past he rarely suffered from mild to moderate, exclusively diurnal, attacks of tension-type headache. His general medical and neurological examinations were unremarkable. All his laboratory tests, including the complete blood count and the erythrocyte sedimentation rate, were in the normal range, as well as his EEG and cranial CT. He had been treated with a morning dose of 25 mg of prednisone for 15 days and then with 12.5 mg for 15 additional days. The nocturnal attacks of headache rapidly decreased in frequency and intensity. Complete pain resolution was achieved after 1 week of treatment. At 1-month, 3-month and 18-month follow-ups he was free of nocturnal headaches but complained of occasional daytime mild or moderate episodic tension-type headaches.
Patient 2
An 80-year-old woman, who had never suffered from headaches, for 6 months was almost nightly awoken at a consistent time (04.30–05.00) by a severe, non-pulsating, fronto-parietal, bilateral headache of 30–45 min duration, which generally forced her to get up and walk to seek relief. No autonomic symptoms were reported and the pain did not respond to the administration of naproxen. An arterial hypertension had been diagnosed 5 years prior to presentation and the patient had been successfully treated with clonidine. Suspecting that the nocturnal headaches could be caused by high blood pressure, she often measured her blood pressure during the headache attacks, but the values remained in her usual range (150/90 mmHg). No general medical or neurological abnormalities were evident on examination. Standard blood chemistries, complete blood count and erythrocyte sedimentation rate were normal, as were EEG and CT of the head. A morning dose of 25 mg of prednisone was given for 15 days and 12.5 mg for another 15 days. After 6 days of treatment a complete resolution of nocturnal headaches was observed. At 1-month, 6-month and 18-month follow-ups she did not report any type of headache.
The clinical features of the cases are presented in Table 2.
Characteristics of cases
Discussion
A new-onset headache in the elderly, as in patient 2, or a marked change in the features of an already experienced headache, as in patient 1, need a search for organic disease. In the reported cases the normality of general medical and neurological examinations, laboratory tests, EEG and cranial CT permitted the exclusion of secondary symptomatic headaches.
The hypnic headache syndrome shares some features with cluster headache, i.e. short duration, predictability of time of onset, clustering of attacks and, in our patients as well as in other reported cases, severe pain with agitated behaviour, but the characteristics of pain, more frequently diffuse and bilateral rather than focal and lateralized, the later age at onset and the constant and complete absence of symptoms of autonomic dysfunction, differentiate hypnic from cluster headache.
In the presently reported patients, the high frequency of headaches per month, the occurrence at a consistent time at night, the bilateral localization of pain and the lack of association with autonomic features agree with the diagnostic criteria of hypnic headache proposed by Goadsby and Lipton (7). One of the suggested diagnostic criteria was the attack duration of 5–60 min (7). In patient 1 the episodes of hypnic headache lasted for as long as 120 min. As similar attack duration is frequently reported in the literature (4–6, 8, 9, 11, 12) and some authors describe even longer episodes (5, 8), a modification to this point seems advisable, so we propose an extension in the length of the headaches to 5–180 min.
In the present cases the pain was diffuse and bilateral but the review of the literature revealed a remarkable proportion of patients with unilateral pain (5, 6, 8, 9, 12, 15). Therefore, as the location can either be generalized or unilateral, it would be sensible to mention location in the description of hypnic headache but not to establish it as a diagnostic criterion.
Raskin (1) suggested that hypnic headache may be a chronobiological disorder and speculated that, as the mammalian biological pacemaker is serotonergically modulated, lithium, which enhances the serotonergic neurotransmission, can exert an effect on circadian rhythms and consequently prevent attacks of hypnic headache. It has been suggested that hypnic headache attacks may be associated with rapid eye movement (REM) sleep (1, 2, 8). In this case, the capacity of lithium to suppress REM sleep may explain its effectiveness in this disorder. Recently, however, the temporal relationship of REM sleep and headache was not confirmed in one case in which polysomnographic study showed arousal at stage 3 slow wave sleep caused by a headache episode (13). Indeed, the majority of authors used a bedtime dose of lithium carbonate (300–600 mg) with generally good response (1–5, 7–10, 12) but some patients were unresponsive or had partial or transient response (4, 13). Intolerable side-effects caused the discontinuation of lithium in a significant number of patients (2, 4, 8, 11). The administration of a daily dose of 25–150 mg of indometacin was effective in nine patients (9, 14, 15). Two patients had a good response with 5 mg of flunarizine at bedtime (6). A dramatic improvement with 1200 mg of gabapentin per day was reported in a patient for whom a 600-mg bedtime dose of lithium did not relieve the headaches (13). Caffeine 60 mg or a cup of coffee at bedtime resulted in a successful prevention of nocturnal headaches (8, 10).
Steroids are fast-acting, very effective preventive agents in cluster headache (17). The reason for this efficacy is speculative: as lithium, they may act modulating the normal serotonergic tone and, in addiction, they may exert a stimulating effect on the opioid system (18).
In both our patients a daily dose of 25 mg prednisone for 15 days followed by 12.5 mg for 15 additional days, in association with 150 mg ranitidine per day, suppressed the attacks of nocturnal headache. This treatment was well tolerated and without side-effects. The therapeutic response was rapid (within 1 week) and the successful prophylaxis was long-lasting (18 months).
In conclusion, prednisone was effective in preventing hypnic headaches and could be considered as a valid alternative therapy in patients who did not respond to or did not tolerate lithium.