Abstract
Introduction
Hypnic headache (HH) is a rare disorder characterized by frequent nocturnal attacks of a generalized headache beginning late in life. Its pathogenesis is largely unknown. It is believed to be associated with a disturbance of the supra-chiasmatic nucleus of the hypothalamus and reduced melatonin secretion (1–3).
We report a case fulfilling proposed diagnostic criteria for hypnic headache (4). Despite an excellent response and long-term remission with lithium therapy, she suffered a relapse following a flight to the South-west across three time zones, supporting the hypothesis of a chronobiological disorder.
Case report
A 68-year-old housewife had suffered from strictly nocturnal headache attacks for the past 2 years. The frequency of attacks ranged from one per night to three per week. The headache always awakened her from sleep between 03.00 and 04.00. The pain was bilateral, felt in the frontal and occipital regions, sometimes predominating on the right side, and was of moderate intensity. There were neither oculo-sympathetic nor focal neurological symptoms. She also denied nasal obstruction, rhinorrhea, facial oedema or erythema. She occasionally became nauseated but never vomited. When the pain was more intense she took either paracetamol (500 mg oral) or ergotamine (0.75 mg, oral), which were usually effective. The headache lasted for c. 40–45 min. Before the first consultation the headache attacks became more frequent and prolonged (lasting for up to 2 hours), not responding to the usual therapy. During the attacks she stayed out of bed, calmly waiting for it to subside. She had no headache during the day.
She had a past history of migraine attacks without aura, hiatal hernia and a Caesarean section. One of her daughters suffered from migraine.
Both her general physical and neurological examinations were normal and her blood pressure was 130/80 mmHg. There was no pain or tenderness over the posterior cervical region. A cranial CT scan was normal and an X-ray study of the cervical spine showed degenerative changes with a narrowing of the C5-C6 disk space.
She was prescribed lithium carbonate 400 mg as a single bedtime dose, with a rapid and remarkable improvement. During the following month she only experienced two mild nocturnal headache episodes that were easily controlled with paracetamol. She discontinued lithium carbonate 3 months later, and was well and free of pain for the next 10 months.
At that time she flew from Portugal to Brazil (GMT 03.00), three time zones to the west. During that 10-day trip she experienced one headache every night, awakening her at 03.00–04.00 am (local time). The pain was identical to the one just described. She tried to control it with an iced Coca Cola can on her head until she obtained some relief with her usual analgesics. There was no pain during the day. The headache disappeared spontaneously by travelling back home 10 days later. There was no need to take any further prophylactic treatment.
Three months later she suffered another relapse, experiencing one headache every night for 15 consecutive days. Each attack lasted about 1 h, and had the previously described features except for nausea. During this episode she also reported a painful point over the occiput while in pain (but not outside the attack). She was depressed because her husband had been recently diagnosed with cancer. She was prescribed lithium (400 mg), dothiepin (75 mg) and alprazolan (0.25 mg twice a day), with rapid disappearance of her nocturnal headache attacks.
Discussion
This 68-year-old woman had frequent short lasting, strictly nocturnal, headache attacks for 2 years. The pain was bilateral, lasted for up to 2 h and had no associated cranial autonomic symptoms. There was an excellent response to lithium on two occasions. There was no underlying organic disease. These symptoms fulfil the proposed diagnostic criteria for hypnic headache syndrome (4).
This clinical history overlaps many published descriptions of hypnic headache (1–18), except for two aspects: the remitting–relapsing course and the precipitation of the attacks by travelling (across time zones), disappearing with her return back home.
The natural history of hypnic headache is not well known. A review of 53 (1–18) cases, published between 1988 and 2001, suggests that this syndrome tends to be rather chronic from the outset. The majority of patients have a long history of symptoms, 4.9 years in average but ranging to a maximum of 30 years (3), before diagnosis is made. However, we do not know if the frequency of attacks is high since the onset or if it increases over time. In our own patient there were periods of infrequent attacks but we could not identify long (> 1 month) headache-free intervals before diagnosis was made. The exact frequency of the attacks at onset was not recalled.
The outcome after diagnosis is variable. Thirty-seven out of 53 published cases have some follow-up information beyond the initial therapeutic response. These can be grouped in three main patterns (Table 1).
Hypnic headache: course after diagnosis
n=number of cases; NI = no information.
The majority, 43.2% (16/37), of patients had a good outcome but required chronic therapy. The most commonly prescribed prophylactic drug, in this group, was lithium (6/14 patients). These patients responded to therapy but relapsed whenever medication was discontinued, suggesting some dependency from treatment. Two patients just required symptomatic treatment for acute attacks with caffeine combined with an analgesic or a triptan (3).
One-third of patients (11/37, 29.7%) had incomplete or no therapeutic control. The most commonly prescribed medication was lithium (9/11 patients). Patients either responded initially but relapsed with ongoing therapy, or they stopped treatment for some reason (side-effects were the main reason in five patients). Others refused to take therapy or did not respond to it (3/11).
Twenty-seven per cent of patients (10/37) had an excellent outcome. Some of them (4/10) apparently became cured after a short therapeutic trial with either lithium (1), indomethacin (18) or flunarizine (9). Two resolved completely after longer periods of treatment with lithium and two others remitted spontaneously.
In 18 patients there is either no follow-up information (eight cases) (3) or just a brief mention of an initial response to therapy (10 cases, 7 of whom were taking lithium) but no further evaluation.
Our patient fitted better the third group. She responded well to lithium and continued free of headache for 6 months after stopping therapy. However, the headache recurred twice. In one occasion it responded to medical therapy, on the other it remitted spontaneously. The characteristics of the attacks did not significantly change in the three active periods, except for the disappearance of nausea. This case shows that hypnic headache can have active periods, like cluster headache, with very frequent attacks, and long inactive phases where patients are well without medication. Although relapses have been described in this syndrome they were related to interruption of medication or interpreted as a lack of therapeutic response. More prospective studies of HH patients are necessary to evaluate how frequent this relapsing pattern can be.
Headache recurrence, in our patient, was temporally associated with two possible precipitants: travelling and depression. Long distance travel brings with it several environmental modifications: a time lag, a seasonal change (with a marked alteration of daylight exposure and environmental temperature), and changes of lifestyle. In this case, the time zone difference was about 3 h (GMT 03.00) and there was a change from summer to wintertime. There were no significant changes in altitude or diet. The active period began on her arrival in Brazil and finished, without any specific prophylactic treatment, after her return to Europe. While abroad the headache attacks recurred every night at her usual hour.
Trigger factors have not been well studied in HH. Although this headache has been related to sleep, little is known about the effects of sleep changes on the headache pattern. Factors that may trigger the attacks are not evident from most reports, except for two patients who were depressed at the time HH was diagnosed (3) and one patient who had suffered a minor head trauma (3). Namely, we could find no mention of the effects of sleep changes, nor of events that may modify sleep (sleep deprivation, prolonged sleep, changes in sleep habits, such as the ones associated with retirement, etc.) that might act as triggers to HH in predisposed individuals. Contrary to cluster headache, the effect of alcohol intake seems to be none in HH (1). The association between HH attacks and sleep phases was not consistent in different studies (14, 15).
HH has been conceptualized as a chronobiological disorder (1, 2). This hypothesis has been based on several types of evidence: (a) the alarm clock regularity of the attacks, awakening the patient from sleep every night at the same hour; (b) the rapid response to short therapeutic trials with lithium carbonate, a useful drug in disorders with a circadian periodicity; and (c) the shift forward of HH attack timing with lithium (2, 5). The present case brings another element to this hypothesis, i.e. the possibility of a periodic pattern and the susceptibility of HH to changes of circadian rhythms.