Hypnic Headache - the First German Cases Including Polysomnography

Abstract

We report the first four German cases of hypnic headache according to the criteria suggested in the literature. Furthermore, we present the results of polysomnography in two theses cases with hypnic headache occurring during REM sleep. In one case, hypnic headache was also associated with periodic limb movements. Although mean nocturnal oxygen saturation was decreased in another patient, nightly oxygen inhalation did not result in an improvement of the headache.

Introduction

Hypnic headache was first described by Raskin (1) and soon regarded as an idiopathic headache disorder. Since then, more than 50 cases from several countries have been reported in the literature and as congress contributions (for review see (2)). However, no German cases have been published to date. Furthermore, only few cases included polysomnography. In one case (3), hypnic headache occurred during REM sleep; in one case (4), it did not.

We describe four German patients with a hypnic headache syndrome according to the criteria of Goadsby and Lipton (2), which demand bilateral headache attacks without autonomic features awakening the patient from sleep and with a duration of 5–60 min. Structural lesions may not be present or do not explain the headache. Since several case reports with typical hypnic headache but with an attack duration of>60 min and unilateral headache have been published (5, 6), we allowed exceptions with respect to these criteria. We also allowed very mild autonomic features. In three of our cases, polysomnography was performed in order to exclude any other sleep abnormalities and to detect the sleep stages in which hypnic headache occurs.

Case reports

Case no. 1

A 66-year-old woman was admitted to our headache out-patient clinic with a 1-year history of headache attacks awakening her every night about 3 h after falling asleep (about 1 o'clock in the morning). The headache was always right-sided periorbital and not pulsating. The intensity was moderate (5 out of 10 on a visual analogue scale), and no autonomic symptoms such as lacrimation, rhinorrhoea, ptosis, reddening of the eye, photophobia, or phonophobia were noted. When taking acetylsalicylic acid the pain disappeared after about 30 min, without medication the pain stayed for>1 h. There was always only one single attack per night.

The patient suffered from diabetes mellitus Type 2 and was treated with oral anti-diabetic drugs. Twelve years ago she underwent kidney transplantation because of familial cystic kidney disease. The function of the transplanted kidney was nearly normal. No remarkable neurological diseases were reported in the patient history and in the family history. She never had headache before the onset of the nightly headache attacks.

The neurological examination showed mild signs of a polyneuropathy. A magnetic resonance imaging (MRI) scan of the brain with and without contrast, Doppler ultrasound of the brain-supplying arteries, and evoked potentials including blink reflex were normal, as were blood tests with the exception of mildly increased creatinine and urea.

We performed polysomnography in the patient and detected a poor sleep quality (24 arousals during 9 h sleep), decreased oxygenation (about 90–95% during the total night, twice for 3 min down to 85%), but no apnoeas. The headache attack occurred during the first REM sleep stage (about 3 h after sleep onset) and was not related to oxygen desaturation. We also detected periodic limb movements (PLM) of the legs during the total night without awakening. The patient was not aware of her nightly PLM. As a widow, she slept alone, she had no complaints suspicious for a restless legs syndrome. The hypnogram with the different sleep stages and the electromyography of the right anterior tibial muscle are presented in Fig. 1 (registration without any medication).

Figure 1

Hypnogram (top) and electromyography of the legs (bottom) in case 1 with the time on the absciisa. The onset of the headache attack is marked by an arrow. The bottom figure shows the activity of the right tibial anterior muscle. Analysis of the time periods with high muscle activity showed periodic limb movements with a duration of 3 s every 10 s. (S 1–4 denotes the different sleep stages.)

The patient was on 240 mg verapamile because of a mild arterial hypertension due to kidney disease. Because of many arousals during sleep, we added 75 mg amitriptyline. This medication improved her subjective sleep quality and decreased the intensity of pain, but did not stop the headache attacks. After an additional 80 mg of verapamile, the patient was nearly but not completely headache-free. She did not want to increase further the dosage of the drugs.

Case no. 2

A 58-year-old woman was admitted to our in-patient hospital because a giant cell arteritis was assumed by her general practitioner. She complained of dull headache attacks for 2 years occurring nearly every night and starting about 2 h after sleep onset. The pain was moderate and always left-sided, with a maximum around the ear and at the temple. No autonomic symptoms were reported. The duration of headache was 1–2 h, the patient did not want to take any acute medication. She had a history of neck pain and unilateral headache for several years fulfilling the criteria of cervicogenic headache by Sjaastad et al. (7), but not of migraine or any other idiopathic headache.

The history of the patient was unremarkable except for hysterectomy 10 years ago and Sjögren's syndrome diagnosed at the time of hysterectomy. No neurological diseases were reported in the family history. Neurological examination was unremarkable. In the psychiatric examination, a mild dysthymia according to the criteria of ICD 10 was noted. An MRI scan of the brain with and without contrast, EEG, evoked potentials, and Doppler ultrasound of the brain-supplying arteries were normal as well.

We performed a polysomnography which showed normal sleep stages and normal oxygenation without any signs of apnoeas or PLM. There was a physiological distribution of REM sleep. However, the patient had two very long periods of awakening and did not get a headache attack during this night. Unfortunately, she refused to undergo another polysomnography.

The patient was treated with amitriptyline 150 mg because of concomitant dysthymia. This medication improved her sleep and the headache, but did not stop the attacks. The patient did not want to try any other medication.

Case no. 3

A 67-year-old man was admitted to our headache out-patient clinic because of a 10-year history of dull headache attacks of moderate intensity occurring every night, always about 1 h after falling asleep. This resulted in an attack frequency of about two to four attacks every night during the last 2 years. The duration of the headache attacks was about 60 min. The headache was diffuse and bilateral with mild lacrimation on both eyes. Acetylsalicylic acid or paracetamol did not improve the headache.

The patient had a history of arterial hypertension which was treated with enalapril. An MRI scan of the brain showed signs of a vascular leucencephalopathy but no infarction and no contrast enhancement. The neurological examination was normal, as were an EEG and nerve conduction velocity studies. In the patient's and in his family's history, no further idiopathic headache was known.

We performed a polysomnography in three nights showing that the headache always started during a REM sleep stage. A typical night (recorded without any medication) is shown in Fig. 2. The nightly oxygenation in this patient was decreased (about 90–93%, 4 min in total <88%) but did not suggest apnoea syndrome. No PLM could be observed. The arterial blood pressure was normal during the total night (between 100/70 and 140/90 mmHg). In another night, we applied oxygen supply (inhalation of 100% with 2 l/min via the nose). However, this did not result in an improvement of headache intensity or in a reduction of attack frequency.

Figure 2

Hypnogram of case 3 with the time on the abscissa. The onset of headache attacks is marked by an arrow. (S 1–4 denotes the different sleep stages.)

The patient was treated with different tricyclic antidepressants which all showed no benefit. We also tried caffeine and flunarizine, which was not effective either. A small dose of lithium resulted in an increase of arterial blood pressure and was discontinued. Finally, 240 mg verapamile (time release formulation) per day improved the headache markedly with only rare attacks of about 15 min duration.

Case no. 4

A 67-year-old woman was admitted to our headache out-patient clinic with the diagnosis of cluster headache. She complained of severe nightly non-pulsating headache attacks always right-sided periorbital and temporal but without any autonomic symptoms such as nausea or vomiting, photophobia, phonophobia, lacrimation, rhinorrhoea, or reddening of the eye. The attacks always occurred between 2 and 4 o'clock for 2 h, always one attack per night for 10 years at the time of admission.

The neurological examination, an MRI scan of the brain with and without contrast, and evoked potentials were normal. Unfortunately, the patient refused to have a polysomnography. However, she stated that she awoke with the headache attack often during a dream. The history of the patient was unremarkable except for a hysterectomy about 20 years ago.

The attacks decreased in intensity after taking acetylsalicylic acid. Subcutaneous sumatriptan and oxygen inhalation did not have any influence on the pain. We prescribed verapamile with a daily dose of 240 mg. This resulted in an improvement of her headache frequency with only one attack per week.

Discussion

We report four patients fulfilling the diagnostic criteria for hypnic headache as suggested by Goadsby and Lipton (2), but with longer duration of attacks than 60 min and with sometimes unilateral headache, since cases with unilateral hypnic headache and longer attacks have been published (5, 6). In one case, mild lacrimation was allowed. No patient had similar headaches not stemming from sleep or during daytime naps. The clinical and demographic features of our patients are in concordance with the several other cases reported in the literature (1, 5). These reported cases represent about 0.1% of all headache patients seen in our headache out-patient clinic during a 4-year period.

In three patients, we were able to perform polysomnography which showed onset of headache only during REM sleep stages in two patients. In the remaining patient, onset of headache was reported to go along with dreaming. The association between onset of hypnic headache attack and the first REM sleep stage in the night has already been reported in another case (3). A temporal association between dreaming and onset of hypnic headache attacks also supports the hypothesis that hypnic headache is linked with REM sleep, as is reported in other cases (1, 5, 8, 9). We furthermore observed desaturation of oxygen down to about 85% in two of our cases. Dodick (3) also observed desaturation in hypnic headache, but with a much higher decrease down to about 70%; in his patient, CPAP was effective in resolving hypnic headache. Only one case in the literature with polysomnography, however, did not show an association of hypnic headache with REM sleep; in this case, no relevant desaturation of oxygen could be noted during the night (4). Tricyclic antidepressants were tried in three of our patients and found to be ineffective. These drugs are able to suppress REM sleep (10), which is a support of the hypothesis that hypnic headache is associated with REM sleep but that deprivation of REM sleep is not an appropriate treatment.

However, it is not proven by these cases that hypnic headache is obligatorily and directly caused by altered REM sleep. It may also be that during REM sleep a threshold is lowered for attacks to occur but caused by other reasons. During REM sleep, the dorsal raphe and the locus coerulei nuclei are not active (11). These brain areas are essential parts of the human antinociceptive system, and it might be that the link between hypnic headache and REM sleep can be found in this (physiological) transient dysfunction.

In one patient, we tried oxygen inhalation during an additional night; the reason for this was a remarkable, although not dramatic, decrease of nightly oxygenation (about 90%). This did not lead to an improvement of the headache. This finding suggests that mild hypoxaemia in the night is not a main cause of hypnic headache. Further studies will be required to conclude this with certainty.

Patient 1 had PLM during the night. This is a phenomenon often observed in patients with kidney disease and increased creatinine or urea, but never before reported in patients with hypnic headache. Therefore, we do not believe that PLM is specifically associated with hypnic headache. However, we recommend performing polysomnography in patients with hypnic headache, including electromyography of the limbs to gain more data on this question.

In conclusion, these four cases can be added to the sample of patients with hypnic headache reported to date in the literature. Furthermore, they give evidence that hypnic headache is an idiopathic headache disorder associated with REM sleep.