Abstract
Hypnic headache is a rare sleep-associated primary headache disorder. It was first described by Raskin in 1988 (1) and has been recognized as a primary headache disorder. In 2004, hypnic headache was included in group 4 (other primary headaches) of the 2nd edition of the International Classification of Headache Disorders (ICHD-II) (2). According to ICHD-II, subjects complaining of exclusively sleep-related dull headache can be diagnosed with hypnic headache if at least two of the following are present: occurrence > 15 times per month; duration of pain for ≥ 15 min after waking; age at onset after 50 years. In addition, absence of autonomic symptoms, presence of not more than one of nausea, photophobia or phonophobia and exclusion of another disorder are required for diagnosing hypnic headache.
The pathogenesis of hypnic headache remains unclear. Several possible mechanisms, including dysfunction of brainstem neural networks, have been discussed (3).
Polysomnographic studies have been performed in only 24 (22%) of the 109 reported cases of hypnic headache, and 15 patients (14%) had an attack during polysomnography (4). In 10 patients, the attacks showed a close temporal relation to rapid eye movement (REM) sleep (5–10). In five patients, the attacks had an onset during non-REM sleep (9, 11, 12). Although sleep-disordered breathing was found in seven of the 24 patients (29%) who underwent polysomnography, only one attack was definitely time-related to severe hypoxia (5).
Gil-Gouveia and Goadsby (13) have reviewed 101 patients and found that arterial hypertension was the most prevalent co-morbidity (n = 17). Ambulatory blood pressure monitoring was performed in 10 patients (7, 9, 10, 14–17). In four patients, mean blood pressure values were normal, but the authors did not comment on eventual blood pressure peaks (7, 10, 15). In four patients known to suffer from arterial hypertension, nocturnal ‘dipping’ of blood pressure was documented (9). Ghiotto and coworkers reported a patient with an isolated hypertensive episode not related to hypnic headache (14). In one female patient with known arterial hypertension, blood pressure measured by the patient during hypnic headache attacks was within her usual range (150/90 mmHg) (18).
Regarding the treatment of hypnic headache, several compounds have been found to be effective, but there have been no randomized controlled trials. Therapeutic options include lithium, indomethacin and caffeine. Single patients have been successfully treated with topiramate (17), botulinum toxin (19), pregabalin (20), and a combination of hypnotics (21).
In this case report we present a 54-year-old woman with onset of hypnic headache at the age of 47 years. This is the first Austrian case as well as the first case with serial polysomnography and blood pressure monitoring in treatment with indomethacin.
Case report
A 54-year-old female patient presented to our out-patient clinic in April 2007. The patient had a history of migraine with aura which usually occurred twice a month. At the age of 47 years the patient noticed an additional new type of headache. It occurred on at least 5 days per week always at night time, usually between 02.30 and 03.00 h, and woke her from sleep. The pain was dull in quality and localized bilaterally in the head. During the attacks the patient could not remain supine and walked around or sat down in a chair. The patient did not notice autonomic symptoms such as conjunctival injection, lacrimation or rhinorrhoea. Some attacks were associated with nausea, but photophobia and phonophobia did not occur. The attacks lasted between 15 min and 3 h after waking. During the 7 years since onset the patient had an average of two headache-free intervals per year, which lasted between 21 and 28 days. Two weeks after her first appointment the time of headache onset shifted from approximately 02.30 to 04.30 h (Fig. 1). Intriguingly, in this period the patient used to go to sleep at midnight instead of 22.00 h.
Headache diary documenting attacks of hypnic headache, nights with recording of polysomnography and blood pressure and treatment with indomethacin. Diamonds, hypnic headache; circled diamonds, recurrence of hypnic headache after the patient had not taken indomethacin the evening before; dark grey columns, recording nights; white columns, no intake of indomethacin; light grey columns, intake of indomethacin.
The headache characteristics reported by the patient met the ICHD-II criteria for hypnic headache (22), and normal findings of the clinical neurological examination as well as of further investigations confirmed the diagnosis. Magnetic resonance imaging (MRI) did not reveal any intracranial lesion. Routine blood laboratory tests were normal, except for mild hypercholesterolaemia and hypertriglyceridaemia. Obstructive sleep apnoea syndrome as a cause for the nocturnal headaches was excluded using a portable device for monitoring blood oxygen saturation during sleep (Poly-MESAM; ResMed Austria Medizintechnik GmbH, Vienna, Austria).
For analysing the relation of headache onset to sleep stages and arterial blood pressure, ambulatory polysomnography (Alpha Trace–Langzeit-EEG-Recorder; B.E.S.T. Medical Systems, Dr Grossegger & Drbal GmbH, Vienna, Austria) and simultaneous half-hour blood pressure monitoring (MobiloGraph; GepaMed Medizintechnik GmbH, Stolberg, Germany), starting at 14.00 h and stopping at 07.00 h, were performed. The patient was told to record any episodes of headache during that night. In addition, she kept a standardized evening and morning protocol (23) (according to the guidelines of the German Society for Sleep Medicine) for recording subjective and objective sleep parameters.
In the night of the first measurement, the patient awoke with a dull, bilateral headache at 04.33 h, but went back to sleep shortly thereafter. The polysomnographic recordings showed a moderately fragmented sleep pattern and a slight reduction of REM sleep time and indicated that the headache occurred during the second REM sleep phase of that night, followed by a short arousal as reported by the patient in her headache diary (Fig. 2). Blood pressure measured at 04.30 h was 161/97 mmHg, that measured 30 min before at 04.00 h was 135/73 mmHg and that measured 30 min afterwards at 05.00 h was 150/88 mmHg. There was one additional hypertensive peak at 14.30 h (159/100 mmHg). Details of the nocturnal blood pressure are shown in Fig. 2.
Ambulatory polysomnographic recording and half-hour blood pressure measurements before treatment with indomethacin. The arrow indicates the onset of headache at 04.33 h during the second REM sleep phase, followed by a short arousal and associated with an increase in blood pressure. W, wakefulness; REM, rapid eye movement sleep; S1–S4, slow-wave sleep stages 1–4.
The day after these measurements the patient was started on indomethacin 75 mg b.i.d. (day 1). According to the headache diary the patient had a total of eight headache attacks (occurring at 02.30, 04.30, 05.30 or 07.30 h) from day 1 to day 13 of treatment with indomethacin. From day 14 on she was free of hypnic headache (Fig. 1).
On day 21 the patient tapered the daily dose of indomethacin to 75 mg due to dizziness after the ingestion of the morning tablet. She remained completely free of hypnic headache for a total of 68 days until she forgot to take indomethacin (day 83). The following night hypnic headache reappeared at 04.30 h (Fig. 1). She did not take indomethacin on days 88, 101 or 109 and suffered from hypnic headache during each of the subsequent nights (Fig. 1).
The patient herself discontinued the treatment with indomethacin on day 114, explaining this decision later by the fact that she had undergone surgery because of an abscess in her mandible on day 99, that she had taken amoxicillin and clavulanic acid for 10 days and no longer wanted to use medication. Interestingly, hypnic headache recurred once on day 115, but not thereafter (Fig. 1).
Polysomnography was repeated on day 126 after the patient had taken a single evening dose of 75 mg indomethacin (Fig. 3). The total amount of time spent in REM sleep was twice as high as before she had started indomethacin, and the total time spent in slow-wave sleep was more than three times higher (Fig. 4a). According to the patient's evening and morning protocol, subjective sleep parameters (Fig. 4b) and sleep efficiency (Fig. 4a) showed a clinically relevant improvement.
Ambulatory polysomnographic recording after administration of an evening dose of 75 mg indomethacin.
Objective (a) and subjective (b) sleep parameters. White columns represent the first recording, grey columns the second and black columns the third recording. The hatched columns in (b) represent standardized values.
On day 171, polysomnography with simultaneous half-hour blood pressure measurements was performed free of medication. Polysomnography showed two arousals from REM sleep that were not associated with hypnic headache, and the patient did not report any other headache during this night (Fig. 5). The sleep profile was similar to that before treatment with indomethacin, and the amount of slow-wave sleep and REM sleep had decreased compared with the second recording (Fig. 4a). The improvement in subjective and objective sleep parameters recorded in the evening and morning protocol remained stable, however (Fig. 4a,b) Half-hour blood pressure monitoring revealed a physiological nocturnal drop in arterial pressure (Fig. 5). Two hypertensive peaks around 18.30 h (176/101 mmHg, data not shown) and 20.00 h (180/ 96 mmHg, data not shown) were not associated with headache attacks.
Ambulatory polysomnographic recording and half-hour blood pressure measurements after discontinuation of indomethacin.
The patient remained free of any medication; in particular, she did not take indomethacin, other analgesics or antihypertensive drugs. She remained free of hypnic headache until her last visit on day 201, i.e. 87 days after discontinuation of indomethacin, and reported that she had once again an idea of what ‘good sleep’ is about after 7 years of frequent nocturnal headache.
Discussion
Hypnic headache is a rare condition, mostly affecting the elderly and generally considered a benign disorder. We could rule out an intracranial mass, sleep apnoea syndrome or nocturnal hypertensive peaks in our patient by brain MRI, monitoring of blood oxygen saturation, routine laboratory examinations and blood pressure monitoring.
Hypnic headache is considered as a disorder of the circadian rhythm (12) and has been reported to be related to REM (7, 9), as well as non-REM (9, 12) sleep. In our patient, polysomnography demonstrated onset of hypnic headache during the second REM sleep phase. Intriguingly, we also found a diminished amount of time spent in REM sleep, prolonged sleep latency and reduced sleep efficiency. The patient reported reduced subjective sleep and awakening quality.
Capuano and colleagues performed three consecutive polysomnographic recordings in a single patient with hypnic headache before and during treatment with amitriptyline, which completely abolished the nocturnal headache (8) and led to an increase of REM sleep from 7.0% to 19.1%.
Reports on the effect of non-steroidal anti-inflammatory drugs (NSAIDs) on sleep are controversial. NSAIDs have been shown to disrupt normal sleep in humans (24). Treatment with indomethacin led to a significantly elevated REM sleep time in rats (25). Our polysomnographic data show a marked increase of REM as well as slow-wave sleep after administration of 75 mg indomethacin.
Hypertensive peaks may be related to the occurrence of headache as well as to REM sleep (22). In a patient suffering from chronic paroxysmal hemicrania, concurrence between blood pressure rise and pain attacks has been observed (26).
In our patient, blood pressure monitoring indicated mild hypertension, and there were two hypertensive peaks in each of the two measurements. However, it is impossible to draw any conclusions regarding the relation of blood pressure and headache in our patient.
Among the cases reported in the literature we have found six patients with hypnic headache who had undergone ambulatory blood pressure monitoring as well as polysomnography (5, 7, 9). However, in none of these patients were the recordings performed simultaneously.
Indomethacin has been previously reported to be beneficial in the treatment of hypnic headache. In contrast to the patient presented here, pain was localized predominantly unilaterally in those responders (5, 10, 15, 27–30). During follow-up ranging from 12 weeks (29) to 13 months (15), only three of nine patients remained hypnic headache-free after discontinuation of treatment (5, 15, 29). Several mechanisms underlying the response to indomethacin have been postulated. One could speculate that the increase of REM sleep related to the intake of indomethacin in our patient may also play a role.
Our patient went into full remission of her bilateral hypnic headache during treatment with indomethacin. Accidental interruption of treatment led to a recurrence of hypnic headache during four nights. Surprisingly, permanent discontinuation of indomethacin caused only one single recurrence of hypnic headache in the first night, followed by persistent remission during a follow-up period of 12 weeks. Although speculative, this single recurrence (and the recurrences before) might be explained by rebound headache.
In conclusion, this first Austrian case of hypnic headache is also the first case with serial polysomnography and blood pressure monitoring in treatment with indomethacin. Our findings agree with previous reports demonstrating a relation between hypnic headache and REM sleep and confirm the efficacy of indomethacin. Indomethacin may have an effect on the sleep pattern and may induce sustained remission in some patients with hypnic headache.