Abstract
Background: Hypnic headache (HH) is a rare primary headache disorder that is characterized by exclusively sleep-related headache attacks. Because of its low prevalence, clinical features and therapeutic options are widely unknown or under discussion.
Methods: Twenty patients with HH were examined and interviewed using a standardized questionnaire in regard to their clinical characteristics and effective treatment regimens. Data were evaluated according to current International Headache Society (IHS) diagnostic criteria. Individual treatment history and effective treatment options were compared with expected efficacy based on current literature.
Results: In conflict to current IHS criteria, 15% of patients reported trigemino-autonomic symptoms. All patients showed distinct motor behavior during their headache attacks. In acute pain attacks caffeine was most effective. Lithium, topiramate, melatonin, amitriptyline and indomethacin were sometimes useful prophylactic treatment options but were often associated with side effects.
Conclusions: Our results underline the need for modification of the IHS diagnostic criteria of HH to better reflect the actual clinical characteristics of this headache. Caffeine should be considered as first-line acute therapy. Prophylactic medical treatment should be carefully evaluated in regard to side effects in this aged patient population, as this seems to be a major concern of patients apart from pure pain reduction.
Introduction
Hypnic headache (HH) is a rare primary headache disorder that mainly affects elderly people. It was first described by Raskin in 1988 (1), and 154 cases have been reported in the literature so far. HH is characterized by strictly sleep-dependent headache attacks which occur usually at the same time at night and awaken the patient. This clinical feature unequivocally distinguishes HH from other primary headache disorders and is the reason why HH is also referred to as “alarm clock headache” (2).
In 2004 HH was incorporated into the second edition of the International Classification of Headache Disorders (ICHD-II), in the subsection of “other primary headaches” (code 4.5) (2). The diagnostic criteria postulate a dull headache (A) that only develops during sleep and awakens the patient (B). At least two of the following characteristics should be fulfilled (C): (i) occurrence ≥15 times a month; (ii) duration ≥15 minutes after waking; and (iii) first occurrence after the age of 50 years. The headache should not be accompanied by autonomic symptoms and no more than one of nausea, photo- or phonophobia (D). A different underlying disorder must be ruled out before diagnosis (2). Despite these ICHD-II criteria, there is emerging evidence that further clinical features are characteristic for HH and should be considered for better recognition and diagnosis of this rare headache disease.
Because of very low prevalence of HH, larger controlled clinical studies on appropriate medical treatment options are still missing. Caffeine has been reported to be effective as acute as well as prophylactic treatment when taken before going to bed. Lithium, indomethacin, topiramate (TPM), flunarizine and amitriptyline seem to be effective treatment options as well (3–5). As the pathophysiology of HH is largely unknown, effectiveness of particular drugs might hint at possible pathomechanism.
The purpose of this study was to re-evaluate the diagnostic criteria of HH, assessing the largest German patient cohort of this rare headache disorder in comparison to literature. Moreover, we report response and termination rates as well as efficacy of several treatment options previously described to be effective in HH to enhance the therapeutic algorithm and illuminate the possible underlying pathomechanism.
Material and methods
Twenty patients suffering from HH were investigated. Nineteen of the 20 patients were diagnosed at our tertiary headache center over a period of 10 years. The diagnosis of HH was established or re-confirmed in a face-to-face interview by a headache-experienced neurologist between January and October 2009 and assessed according to the ICHD-II criteria (code 4.5).
During consultation a standardized questionnaire was administered to all patients. It captured information on patient demographics, age at diagnosis, disease duration until diagnosis, headache characteristics (duration, frequency per night and per month, localization, intensity on a verbal rating scale [VRS, no pain = 0; worst possible pain = 10]; associated symptoms including photophobia, phonophobia, nausea and vomiting; motor behavior; autonomic signs including rhinorrhea, nasal congestion, lacrimation, ptosis and facial sweating; aura symptoms; prodromal symptoms; headache triggers) other headache disorders and family history of headache. All patients were advised to keep a headache diary over the period of the following month. Additionally, information on prior and concomitant headache medication (drug, effectiveness, side effects, termination rates) was obtained from the patients. Effectiveness of acute treatment was determined by subjective reduction of pain intensity on VRS and/or reduction of attack duration of at least 50%. Prophylactic treatment was considered to be effective when patients reported a decrease of headache days of over 50% per month. All patients were contacted at least once via telephone within three months after consultation for evaluation of newly prescribed medication. Comorbidities as well as other medications were determined. Results were analysed descriptively. Neurological examination and cerebral MRI were unremarkable in all patients. The protocol of this study was reviewed and approved by the Medical Ethics Committee of the University Duisburg-Essen and written, informed consent was obtained from all participants.
Results
Patient characteristics and headache profile of 20 patients with hypnic headache
SD, standard deviation; VRS, verbal rating scale.
Demographic data
Of the 20 patients diagnosed with HH, 65% were women (N = 13). Mean patient age at the time of consultation was 67.5 years (range: 44–81). One patient was younger than 50 years (44 years). Mean age at onset of HH was 58.9 years (range: 29–74), with two patients aged <50 years (10%). Interval from disease onset until diagnosis ranged between one and 18 years (mean 5.9 years). All but one patient were diagnosed in a specialized headache center. One patient was diagnosed with HH by his general practitioner.
Headache characteristics
Mild pain intensity (VRS 1–4) was reported by 15% of patients (N = 3), moderate (VRS 5–7) by 30% (N = 6) and severe (VRS 8–10) by 55% (N = 11). All of the patients described a dull headache character, two patients (10%) reported additional throbbing/pulsating symptoms and one patient (5%) observed additional stabbing pain features. The pain was bilateral in 14 patients (70%) (occipital in six patients, frontotemporal in four patients and holocephal in four patients). Six patients (30%) reported a hemicranial headache (three left-sided, two right-sided, one alternating). The mean duration of untreated headache attacks was 113 minutes (range: 15–240), with five patients (25%) reporting headache lasting longer than 180 minutes (two patients 200 minutes, three patients 240 minutes). One patient had, despite his regular attacks of 120 minutes, rare HH attacks lasting up to 360 minutes. Frequency of attacks was 23.8 per month (range: 10–31). In 95% of the patients (N = 19), the time of pain occurrence was between 2 and 4
Associated clinical features
Twenty percent (N = 4) of patients had headache-associated symptoms (10% photo- or phonophobia, 10% nausea), whereas no patient expressed more than one symptom. In three patients mild concomitant autonomic symptoms (one nasal congestion, one lacrimation, one rhinorrhea) were documented. Nasal congestion and rhinorrhea were localized ipsilateral to the one-sided headache, and lacrimation was bilateral in a patient with holocephal localized headache. Autonomic symptoms were less prominent compared with cluster headache (CH). None of these patients fulfilled the ICHD-II criteria for trigeminal autonomic cephalalgias (TACs), including CH and chronic paroxysmal hemicrania (CPH).
None of the patients stayed supine when awakening with headache. Thirty-five percent (N = 7) reported mild motor behavior, 40% (N = 8) moderate and 25% (N = 5) severe motor behavior. After getting up, two patients read/watched TV/listened to radio, four ate something, three drank something (other than coffee), seven drank coffee, seven walked around, two took a hot shower, and three cooled their head. However, none of them reached the level of agitation/restlessness usually observed in CH patients. All patients denied a headache aggravation by physical activity.
Twenty-five percent (N = 5) of patients reported prodromal symptoms. One patient felt increased tiredness the evening before headache occurrence, two patients reported head pressure, one felt diffuse malaise. One patient reported change in weather as a trigger for headache. No further triggers or aura were described by anyone.
HH diagnosis according to ICHD-II criteria
Seventeen patients (85%) fulfilled the criteria of HH according to the IHS-classification (A–E), and 14 patients met all ICHD-II diagnostic criteria (code 4.5). Two patients were younger than 50 years at time of onset of HH and therefore just met criteria C1 (occurs ≥15 times per month) and C2 (last ≥15 minutes after awakening) but not C3 (first occurs after the age of 50 years). One patient had fewer than 15 attacks per month (10 attacks) and, therefore, did not fulfill criteria C1. Three patients contradicted the ICHD-II criteria because of accompanying trigeminal autonomic symptoms (criterion D).
Comorbidity of other headache disorders
Seventy percent of patients (N = 14) reported an additional headache disorder besides HH. Forty percent (N = 8) had a history of migraine (two with aura; six without aura), although just six patients were still suffering from migraine attacks at time of investigation. Two patients did not have migraine attacks for more than five years. Thirty percent (N = 6) had episodic tension type headache. No patient had TACs. On average, those patients with a second headache disorder had 2.5 additional headache days a month apart from their HH. All patients could clearly distinguish between both headache types.
Comorbidity with other diseases
Five patients were suffering from arterial hypertension but all were well controlled with antihypertensive medication (five beta blockers, one additional ramipril). One reported chronic back pain, one barret esophagus, two hypothyreosis, one psoriasis, three rheumatoid arthritis, one tinnitus, one chronic obstructive pulmonary disease (COPD), one depression, one osteoporosis and one renal insufficiency.
Therapy of hypnic headache
Therapeutic options, treatment and discontinuation rates due to side effects of acute and prophylactic medication are displayed in Figure 1. Medication includes prior and current treatment. No patient had medication overuse headache.
Treatment options, response and discontinuation rates in hypnic headache. (A) Acute treatment. Effectiveness of acute treatment is determined by subjective reduction of pain intensity on a verbal rating scale and/or reduction of attack duration of at least 50%. (B) Prophylactic treatment. Treatment is considered to be effective when patients reported a decrease of headache days of 50% per month.
Acute treatment
The most commonly used acute medication was caffeine (at least one cup of strong black coffee after headache onset). It was effective in 11 of 14 patients. Caffeine-containing analgesics were tried in nine patients and effective in 56% (N = 5). Four patients took triptans (two rizatriptan, one sumatriptan, one zolmitriptan), which were effective in all patients. One of these patients reported headache-associated trigeminal autonomic symptoms. One observed severe motor behavior during the headache attacks but no agitation. Both patients did not fulfill the ICHD-II criteria for TAC.
Nonsteroidal anti-inflammatory drugs (NSAIDs) and acetaminophen/paracetamol were taken by 11 or 12 patients, respectively, but were not effective in most of the patients (NSAIDs 19%, acetaminophen/paracetamol 8%). One hundred percent oxygen was used in four patients but was effective in only one patient. Metamizol and opiates were rarely applied (one patient each) and were not effective.
Prophylactic treatment
The most commonly used prophylactic treatment was lithium. The dose of lithium was adapted to plasma level (therapeutic aim: 0.5–1.0 mmol/l). Lithium was taken by 11 patients and was effective in four patients. Three of them had no further headache attacks under therapy. Significant pain reduction was on average observed within two to four weeks of therapy. Lithium therapy was discontinued by 55% (N = 6) because of side effects, including sleepiness, tremor, psoriasis and worsening of headache attacks.
Eight patients tried caffeine as prophylactic treatment, which was effective in 50% (N = 4). Patients drank at least one cup of strong black coffee before going to bed. Two patients (25%) were free of headache under therapy. Pain suppression was observed almost immediately after starting therapy. Two patients stopped the vespertine intake because of nervousness and problems falling asleep. Most of the patients did not want to try this treatment option because they were afraid of sleeplessness.
Topiramate (TPM) was taken by five patients (daily dose 25–100 mg) and effective in two patients. Despite efficacy complete suppression of headache attacks could be achieved just in one patient (daily dose 75 mg). The medication had an impact within the first four weeks of therapy. Three patients stopped therapy because of increased sleepiness, vertigo and pronounced headache intensity.
Indomethacin (daily dose between 25–75 mg) was effective in one of three patients. This patient was completely free of headache with a dose of 25 mg in the evening. The headache subsided within the first week of therapy. Two patients stopped medication because of gastrointestinal side effects.
Three patients took melatonin (daily dose between 3–5 mg), which was effective in one patient, who reported moderate responsiveness, achieving reduction of headache frequency and intensity by half. Time until onset of efficacy could not be determined. One patient stopped intake of melatonin because of aggravation of symptoms.
Amitriptyline (daily dose 10–75 mg) was used by five patients and was effective in three patients. All three patients reported a moderate response to therapy, which resulted in a more than 50% reduction of headache frequency within four weeks of therapy. None of them was free of headache under therapy. Three patients discontinued medication, because of dry mouth, personality change and cardiac arrhythmias.
Discussion
We report the largest case series on patients with HH in Germany so far. Eighty-five percent (N = 17) fulfilled current ICHD-II criteria of HH, 70% (N = 14) even met all criteria (code 4.5 A–E) (2). Three of our patients contradict the current diagnostic criteria because of associated trigeminal autonomic symptoms. None of these patients met the diagnostic criteria of a TAC, mainly due to longer duration of the headache attacks and bilateral localization of the pain. These results are in line with previous data observing autonomic symptoms in up to 14% of patients with HH (3,4). Cranial autonomic symptoms are quite common in TACs, such as CH, where it is reported in up to 95% of the patients, but also more than half of all migraineurs experience autonomic symptoms on occasion (6). Allowing at least one autonomic symptom in the diagnostic criteria would probably better meet the clinical picture of HH patients and would increase diagnostic sensitivity.
Our clinical data confirm a higher prevalence of HH in women compared with men, which was previously reported in most of the former meta-analysis and case series (3,4). In contrast to other studies but in line with the current IHS –classification, all of our patients described their HH to be of dull quality. Other studies reported on throbbing/pulsating, sharp/stabbing, or explosive quality of pain in an appreciable portion of patients (3–5).
All of our patients reported more or less pronounced motor behavior during the headache attacks. No one was able to stay supine when awakening with pain. This clinical feature was not as pronounced as frequently seen in CH, where a sense of restlessness and agitation is included in the ICHD-II diagnostic criteria (code 3.1 C6) (2). Motor behavior was also reported in previous descriptions of HH (7,8). Donnet and Lantéri-Minet mentioned that all of their 22 HH patients felt compelled to get up during headache attacks and do distinct activities like reading, watching television, eating, drinking or urinating (4). Evers and Goadsby reported in a meta-analysis of 71 HH patients distinct physical activity during the headache attacks in almost all of them (3). Acting out motor behavior that can be easily distinguished from CH-associated agitation or restlessness seems to be an important criteria to differentiate HH from other primary headache disorders, such as tension-type headache or migraine (2). It should be added as a diagnostic feature to the ICHD-II criteria what has already been proposed by Donnet and Lantéri-Minet (3).
One difference that should be addressed in comparison to former literature is the longer duration of headache attacks that was observed in some of our patients, in one even up to 360 minutes. ICDH-II criteria do not limit the duration of the attacks and merely require duration longer than 15 minutes (2). However, in most of the previously reported HH patients, headache duration was not longer than 180 minutes (3–5). Longer duration up to ten hours was mentioned in just a few studies (9,10). All of our patients with longer headache duration reported that the pain subsided slowly and a persistent dull feeling remained until late morning, so that an exact end of the headache attack was difficult to determine. It should therefore be considered that discrepancy of duration might, at least in some cases, be due to different definitions of the end of an attack.
Twenty-five percent of our patients (N = 5) reported prodromal symptoms during the day and evening. These symptoms were of unspecific nature (aggravated tiredness, head pressure, malaise). Hypothalamus-related symptoms such as ravenousness and brain stem–associated symptoms such as pronounced yawning that are often observed in migraineurs are missing (11,12). Prodromal symptoms in HH are less compelling than in migraine especially due to the high attack frequency in HH. Headache attacks occur in most of our patients almost daily, so that consistent prodromal symptoms are difficult to determine.
Remarkably, 70% of the patients (N = 14) suffered from another type of headache, either migraine (40%; N = 8) or tension-type headache (TTH) (30%, N = 6). The comorbidity of another primary headache disorder is commonly observed in HH. Most notably, high prevalence of prevailing migraine in up to 40% of the HH patients was observed in other studies as well (3,4). To what extent there might be a causal relationship behind this observation can only be hypothesized at the moment. A pre-existing headache disorder might be a risk factor for developing HH. Deeper understanding of the pathophysiology of HH is needed to answer this question.
Larger studies on treatment of HH patients are still lacking. Due to the low prevalence of this disease, most cases are reported as single treatment reports or smaller open case series, deeply limiting the informative value of reported data. In regard to our observations, caffeine as well as caffeine-containing analgesics seem to be promising acute treatment options in HH. Previous studies done merely report moderate efficacy of caffeine (3). We observed the best efficacy when patients drank a cup of coffee when awakening with headache. A headache preventive effect was observed when patients drank a cup of coffee before going to sleep. Despite its good efficacy many patients stopped intake of caffeine or did not even start therapy because they were concerned about sleep disturbances. The intrinsic acute analgesic properties of caffeine for headache relief have been reported in several other headache disorders, including TTH and migraine (13–15). However, its efficacy is mainly of modest magnitude in these headache disorders and not comparable with the dramatic reduction of headache observed in some HH patients. Pharmacologically, caffeine is a competitive antagonist at the adenosine receptor (A1, A2A and A2B). Adenosine reduces cortical hyperexcitability and is a significant cerebral vasodilatator. Caffeine, therefore, is a potent cerebral vasoconstrictor (16), which suggests an involvement of the vascular system and/or alteration of the excitability level in the pathophysiology of HH.
NSAIDs as well as acetaminophen/paracetamol, metamizol and opioids did not show satisfying response rates, which is in line with previous reports (3).
Efficacy of triptans as observed in four of our patients is not a common observation in HH and just reported in a few cases (3,17). Interestingly, three of our patients who responded to triptans suffered from a unilateral headache. Two patients reported pronounced motor activity during the headache attacks. One patient observed infrequent nasal congestion associated with the headache. These clinical features suggest a cluster-like subtype of HH, although none of these patients fulfilled the ICHD-II criteria for CH or a different TAC. Cluster-like clinical features might predict therapeutic response to triptans that presumably cannot be detected in other subgroups of HH. A major problem, especially in prophylactic treatment of HH, is the bad tolerability of medication in general in this aged patient population. Therefore, existing data on prophylactic treatment in HH show variable results.
Lithium (plasma level of 0.5–1 mmol/l) was tried most frequently in our patients but was successful in only 36%. This observation contradicts prior reports that showed response rates of up to 88% (3,5). One explanation might be that more than half of our patients who took lithium had to discontinue it because of side effects, so that a genuine response rate is difficult to determine. Lithium is a successful treatment option in other headache disorders showing chronobiological features, such as CH, indicating once again an interesting overlap in these two disorders. The exact mode of action of lithium in chronobiological disorders is still controversial. Lithium leads to an increase of serum melatonin level (18–20), which might be important in the pathophysiology of HH. Additionally, lithium causes down-regulation of serotonin receptors and an increase of serotonin release (21), suggesting an involvement of the serotonin metabolism in HH, which might also explain the good response to triptans in some cases.
Melatonin (daily dose 3–5 mg) was taken by three patients and was effective in one (33%). Dodick et al. previously reported that melatonin might be effective in some cases (7). The consistent nocturnal onset time of the headache attacks suggests an involvement of melatonin metabolism in the pathophysiology of HH. Melatonin is a neurohormone synthesized particularly during the night in the pineal gland. Its incretion is strongly regulated by the hypothalamus, notably the suprachiasmatic nucleus (SCN), which has a key role in circadian rhythmicity. Studies on melatonin incretion rhythm in HH are lacking, but disrupted melatonin levels have been observed in CH (22), possibly indicating an altered melatonin incretion in HH as well. Efficacy of melatonin as prophylactic treatment in patients with CH has been reported (23). The exact pathophysiological mechanism of melatonin in this regard is still unknown, but analgesic effects, including potentiation of GABA-ergic inhibition as well as modulation of cellular calcium influx (23) and modulation of 5HT2 receptors, was suggested (24).
Indomethacin (daily dose 25–75 mg) was taken by three patients and was effective in one patient, in whom complete resolution of attacks could be achieved. Its effectiveness has been reported especially in typical indomethacin-sensitive headache disorders, notably paroxysmal hemicrania and hemicrania continua (25)—connecting HH with TACs once again. Indomethacin was reported in several case studies as showing moderate-to-good efficacy in about two-thirds of the patients (3). Dodick et al. noted that indomethacin would be preferentially helpful in patients with unilateral HH (7). Our only patient who was successfully treated with indomethacin had a bilateral headache, contradicting this hypothesis. Indomethacin is thought to act by lowering cerebrospinal fluid (CSF) pressure (26). The clinical feature of alleviation of the headache by getting up out of bed supports the pathophysiological hypothesis of increased CSF pressure in the pathophysiology of HH (27).
Amitriptyline was beneficial in a few cases, but most of the patients did not tolerate treatment because of side effects. Although showing quite a good response rate, the efficacy of amitriptyline remains questionable because treatment did not achieve complete absence of HH in any patient. The efficacy in this one patient was merely rated moderate. In contrast, complete suppression of headache attacks was observed with lithium as well as indomethacin treatment in at least a few cases. Our results are in line with data from the largest review on HH, which included 71 patients and reports a response rate of amitriptyline of one out of 18 treated patients (3).
TPM was effective in two of five patients. Again, a high termination rate due to side effects was observed. Two cases of successful treatment with TPM have been reported in the literature so far (28,29). TPM is an antiepileptic drug (AED) that is also beneficial in the treatment of migraine and CH. It was suggested that TPM modulates the abnormal activation of a proposed hypothalamic pacemaker that is believed to be the pathophysiological correlate of the striking chronobiological features of HH (27,29).
In summary, our clinical data suggest that HH might be related to TACs as well as migraine in regard to pronounced overlap of typical clinical features and response to specific treatment. A common pathophysiological pathway, at least partial, remains speculative. Trigeminal autonomic symptoms as well as physical activity during headache are important clinical features that should be incorporated into the IHS classification of HH. As far as treatment efficacy is concerned, our data indicate that caffeine should be considered as the first-line substance for acute as well as prophylactic treatment. Triptans were often efficient and should be tried in the acute attack, especially in patients with pronounced additional cluster-like symptoms. Lithium, topiramate, melatonin, amitriptyline and indomethacin might be effective, but outcome is variable. Possible side effects and drug interactions should be considered before starting medication, especially in elderly patients.
Disclosure
The authors report no conflicts of interest.