Abstract
Hypnic headache (HH) or ‘alarm clock’ headache is a rare idiopathic headache included in the current classification of the International Headache Society (1). Clinical and therapeutic features suggest a possible involvement of chronobiological mechanisms, while the few studies that used polysomnography (PSG) observed a frequent association of the attacks with rapid eye movement (REM) sleep. We describe a case of non-rapid eye movement (NREM) HH, with circadian clustering of the events, studied using a daily headache diary, to assess the course of the HH episodes over a long period and performing repeated pre- and post-treatment PSG.
Case report
A 40-year-old man presented with a 18-month history of dull headaches during nocturnal sleep and diurnal naps. The attacks were moderate to severe in intensity, usually localized in the left periorbital and frontotemporal region, awakened the patient from sleep and lasted about 1 h. The pain decreased spontaneously and there was no need for drugs to control it. HH episodes recurred with an average of eight attacks per month, once, or at most twice a day. No associated symptoms were referred, in particular no nausea, vomiting, photo- or phonophobia, or signs of autonomic activation. The patient did not associate the onset of HH with particular sleep situations (i.e. hallucinations or dreams), but generally his nocturnal rest was not refreshing, with difficulty in maintenance of sleep and mild excessive daytime sleepiness.
The patient's medical history was unremarkable. In particular, he did not suffer from migraine, arterial hypertension or mood disturbances and he did not report snoring or airflow arrest during sleep. His family history was negative for headache. He was not taking any medication at the time of evaluation. Physical and neurological examinations were normal. Blood cell count, erythrocyte sedimentation rate and standard biochemistry showed no abnormalities. Brain magnetic resonance imaging gave normal findings.
The patient filled in a 5-month daily headache diary which showed a peculiar distribution of the HH episodes both in nocturnal sleep and during diurnal nap (see Table 1). Overnight PSGs were recorded during three consecutive nights. Only during the second night did the patient present two episodes of HH arising from NREM sleep, as shown in Figs 1 and 2. During the three nights, there was no evidence of sleep apnoea. In fact, the patient showed an apnoea–hypopnoea index of <5 on each night, with an average nocturnal SaO2 of 95.5%. Furthermore, the two attacks showed no close temporal relationship with the occurrence of respiratory events. We did not record any limb movements in periodic sequence.
Hypnogram of the second night of polysomnographic recording. On the abscissa the sleep stages [rapid eye movement (REM); S1, stage 1; S2, stage 2; S3, stage 3; S4, stage 4), arousal and movement time (MT) are indicated. Black arrows mark the onset of the hypnic headache attacks.
Shift from stage 4 non-rapid eye movement sleep to wakefulness induced by a hypnic headache attack.
A 5-month daily headache diary
▪, The 24-h course of the 22 hypnic headache attacks.
Treatment was started with amitriptyline at 30 mg at bedtime. During the next 9 months he experienced no HH episodes, with a significant amelioration of sleep parameters evaluated by PSG.
Discussion
Since 1988 several cases of HH have been reported in the literature. Unfortunately, most of them have lacked a scrupulous sleep study, necessary to improve our knowledge of this idiopathic headache. We report here a case of HH meticulously studied with both clinical and PSG approach.
Both the regularity of attacks and the positive response to lithium, known to have a therapeutic value in other chronobiological disorders such as cluster headache and affective disorders, suggest a primary involvement of suprachiasmatic nuclei, the ‘biological clock’, in HH. Thus, we decided to evaluate attacks' course using a simple but useful tool, a daily headache diary. Over a period of 5 months we observed a regular schedule of HH attacks (at 17.00 h and at 03.00 h). Our data support the possibility that HH is a chronobiological disturbance.
The three overnight pretreatment PSGs showed the sleep alterations caused by HH attacks. In fact, during the second recording the patient remained awake for a long time after the attack. HH attacks disappeared after drug treatment and sleep parameters changed, with an important reduction of the wake after sleep onset and the number of awakenings and with a clear improvement of the quality of sleep. Regarding sleep architecture, unlike previous studies (2–4) we did not observe a reduction of REM sleep in the night with HH attacks; rather, there was a relevant shift of REM sleep onset, probably due to the abrupt interruption of sleep caused by the first HH episode, which started about 90 min after sleep onset replacing the first period of REM sleep.
Several authors have speculated that HH represents a REM-related phenomenon, adducing evidence that paradoxical sleep is characterized by an absence of activity in peculiar brainstem regions involved in the antinociceptive system, such as the dorsal raphe and the locus coeruleus (5, 6). Our data did not confirm this pathophysiological hypothesis. This is the third report of NREM HH described in literature. Previously only Manni et al. (4) and Arjona et al. (7) have documented attacks arising from NREM sleep.
Comparing the first and second recordings (the night before and the one with HH attacks) with the third (the night after HH attacks) and fourth recordings (after drug treatment), we noted a progressive increment in the patient's arousability (increase of the arousal index and the number of limb movements). Capuano et al. have reported similar results (2), suggesting a possible role of hypoarousal as a pathogenic mechanism of headache.
Dodick reported the case of one patient, affected by grade 5 snoring, severe obstructive apnoea syndrome and low levels of SaO2, who showed HH attacks occurring during REM sleep. After treatment with nasal continuous positive airway pressure and supplemental oxygen his clinical status improved. Consequently, Dodick suggested an important role of sleep apnoea and oxygen desaturation in the genesis of his HH (8). Notably, events of airflow decrease or arrest during sleep are associated with headache (9), but clinical aspects of headache in obstructive sleep apnea syndrome (OSAS) seem different from those reported by patients complaining of HH. Recently, Alberti et al. (10) reported that in OSAS patients headache shows a tension-type pattern and occurs on the morning after the awakening in 74% of cases. Differently, HH arises from nocturnal or diurnal sleep and awakes the patient. Its characteristics are more similar to cluster headache, except for cranial autonomic symptoms. Furthermore, several cases of HH, including ours, did not confirm any temporal association between attacks and respiratory events (3, 4, 11). Thus, we believe that HH pathophysiology is not dependent on sleep-disordered breathing and that sleep apnoea and oxygen desaturation, in the case observed by Dodick, could be only a precipitant factor.
Evers et al. (3) treated three of their four patients affected by HH episodes with tricyclic antidepressants, with poor results. To our knowledge, only Capuano et al. (2) have reported a good response to amitriptyline in a patient with this specific idiopathic headache. Similarly to them, we observed a fast and complete response to amitriptyline in our patient.
Considering HH as a circadian disorder, we believe that the use of a long-term daily headache diary and of PSG may be useful in the clinical diagnosis.