Some symptoms of migraine appear to be related to the hypothalamus [18, 19]. The hypothalamus is related to the circadian cycle, but a circadian cycle in migraines is controversial [24, 26]. Therefore, this study aimed to use WeChat as a survey tool to explore the correlation between hypothalamus-related symptoms and migraine attacks and examine the circadian cycle of migraines. The results indicate that HFM and LFM display different migraine attack temporal patterns. Hypothalamus-related non-headache symptoms were present in all phases of migraine attacks. The rates of non-headache symptoms were significantly different in different phases.
Many diseases have specific circadian rhythms. For example, acute myocardial infarction, sudden cardiac death, and ischemic stroke mainly occur in the morning [16]. A study suggested that migraine attacks have circadian rhythms, with the incidence in the morning being higher than in other periods [17]. Another study showed many attacks in the morning and around noon, suggesting a double-peak phenomenon [24]. On the other hand, it has been reported that most migraine patients have no specific circadian rhythms [26]. The present study suggests that migraine does not show a circadian rhythm when considering all participants. But when classified as LFM and HFM, migraine attacks displayed specific patterns. In LFM, the migraine attacks were more frequent from 06:00 in the morning and reached a peak around 08:00–09:00 to decrease gradually in the afternoon, with few attacks between 18:00 and 05:59 the next morning. On the other hand, the trend of the HFM group was opposite to that of the LFM group. The attack rate was lower in the morning, gradually increasing in the afternoon, and the attack rate was higher at night.
Changes in migraine circadian rhythm are closely associated with triggers. Hypersomnia, sleep deprivation, hard light, noise, unpleasant odor, and stress are common triggers of migraines [27]. Between 06:00 and 11:59, in addition to hypersomnia and sleep deprivation, most triggers are present. Especially in Shandong Province, work usually starts at 08:00, so the work pressure is high, which may be one reason for the high incidence of paroxysmal migraine attacks. The incidence of LFM attack is decreased significantly after 18:00, which may be related to emotional relaxation at the end of the workday. Low attack frequency at night and before dawn may be related to sleep. Indeed, sleep is a common strategy to relieve migraines [28].
On the other hand, the frequency of attacks in HFM was relatively high at night and before dawn and might be related to sleep disorders. Most (68%-84%) patients with CM have sleep disorders and report insomnia every day or almost every day [29], while the incidence of insomnia in patients with PM is only 38% [29]. In addition, the incidence of daytime somnolence in migraine, especially CM, was significantly increased [30]. Daytime somnolence in CM might be the reason for the low attack frequency in the morning.
The suprachiasmatic nucleus of the hypothalamus is a key structure that regulates the human circadian rhythm [31]. Thus, the hypothalamus plays an important role in the changes of the circadian rhythm of migraines [2]. The hypothalamus might be involved in the chronic progression of migraine [5], and hypothalamic dysfunction might also cause the change in the circadian rhythm of migraine [32]. Therefore, the interaction between the external environment (i.e., headache triggers) and the hypothalamic circadian rhythm might be the reason for the formation and change of the circadian rhythm in migraine. The differences in circadian rhythms between the LFM and HFM groups were significant, so it was speculated that the change in circadian rhythm might be a sign of migraine chronicity.
Migraine is often accompanied by a variety of non-headache symptoms before and after the onset that also affects the patients’ daily work and quality of life [33]. Hypothalamus-related non-headache symptoms are common in migraine patients, even in children [34]. The present study suggests that autonomic symptoms were obvious in the prodromal and headache phases, especially in the headache phase, mainly manifesting as nasal obstruction, runny nose, photophobia, palpebral edema, and conjunctival congestion. In the postdrome phase, the incidence was significantly reduced. Some non-specific symptoms such as fatigue, inattention, and neck stiffness continued to be higher. Although these symptoms are not as obvious as headaches, they also affect patients’ work and study and often led to disability. Non-headache symptoms of migraine mainly occur in the prodromal and postdrome phases [35], as observed in the present study. Autonomic symptoms such as nasal obstruction and runny nose mainly occur in the prodromal and headache phases of migraine, and the incidence is lower in the postdrome phase. It suggests that the nuclei related to the autonomic nerve of the hypothalamus are mainly activated in the prodromal and headache phases.
The pathogenesis of hypothalamus-related non-headache symptoms remains unclear, and some might be related to hypothalamic neurosecretory function. Neuropeptide Y (NPY) is a neurotransmitter secreted by the hypothalamus and is involved in food intake and appetite regulation, pain, and circadian rhythm [36]. NPY can cause blood glucose fluctuation through appetite regulation, leading to hypothalamic dysfunction, ultimately resulting in the non-headache symptoms of migraine [37]. Orexin (or hypocretins) is another hypothalamic neuropeptide. The secretion of orexin has an obvious circadian rhythm. The secretion increases during wakefulness and decreases during sleep [38]. It is considered to play a central role in the smooth transition between sleep and wakefulness. Orexin can directly stimulate monoaminergic and cholinergic hypothalamic brainstem networks in the locus coeruleus to promote wakefulness [3, 39, 40]. Reduced orexin can lead to sleep regulation disorders such as narcolepsy [41]. Hypothalamic orexin fibers have been confirmed to project directly and indirectly through the gray matters around the midbrain to the trigeminal cervical complex in the brainstem and regulate the conduction of trigeminal nociceptive signals [42]. These neuroendocrine functions might lead to symptoms such as fatigue, emotional instability, and inattention. A generally accepted hypothesis is that symptoms such as yawning and nausea are associated with dopamine, and dopamine receptor antagonists can reverse these symptoms [43], which might be related to that dopamine can act on the active site of the hypothalamic-pituitary axis [44]. Many hypothalamus-mediated migraine-related symptoms might be triggered by a series of persistent nociceptive information from the meninges. These pain signals reach the limbic system and hypothalamus through the spinal trigeminal nucleus, leading to non-headache symptoms [45, 46]. Neurons and nuclei in the hypothalamus receive numerous direct nociceptive inputs through the trigeminohypothalamic tract (THT) [46]. These hypothalamic neurons or nuclei have the functions of regulating body temperature, food and water intake, sleep, and circadian rhythm [47], and are closely associated with autonomic nerves, endocrine, and homeostasis, which result in migraine-related symptoms (such as food desire and loss) such as loss of appetite, fatigue, paresthesia, easy depression, irascibility, and other symptoms [48].
An innovation of the present study was using the social app WeChat to explore the circadian rhythm of migraine attacks and non-headache symptoms related to the hypothalamus. The present study used the LFM and HFM classification instead of paroxysmal migraine and chronic migraine and revealed specific migraine patterns.
Still, the present study has some limitations. First, the present study enrolled 162 patients with 2875 attacks, but the sample size is still small, especially patients with MA and menstrual migraine, so that no related subgroup analysis could be conducted. Second, due to higher requirements on patients’ educational level, older or illiterate patients were not enrolled, which may cause a selection bias.