A total of 472 studies were initially identified from the electronic database search. After removing 243 duplicate studies, 229 studies remained for screening. During the screening process, 9 articles unrelated to headache conditions and 114 articles unrelated to HH were excluded. Additionally, 2 articles without available data were excluded, leaving 104 articles for eligibility assessment. Among these, 14 articles that did not meet the criteria of being case reports, case series, or observational studies were excluded. Furthermore, 81 articles described HH cases with an onset occurring after the age of 18 and 4 articles lacked full-text availability. Finally, a total of 5 articles met all the inclusion criteria and were included in the review (Figure 1).

From the 5 selected articles, 7 patients were identified who reported an onset of HH during childhood (14–18). The main characteristics and headache profiles of these patients have been summarized in Table 2. Patients #A–E have been reported in the review that specifically analyzed pediatric patients in 2015 (19). Patients #F and #G were described in separate case reports in adulthood (17, 18), but both experienced HH onset at the age of 15 with headache characteristics that are not reported to have changed over time. Four out of seven patients were females (57%). The mean age of onset for headache attacks was 10 ± 4.3 years (ranging from 3 to 15 years), and the age at publication of the case reports was 15.6 ± 13.6 years (ranging from 7 to 45 years). The diagnosis for patients #A–E was made on average 15.8 ± 25.0 months (ranging from 1 to 60 months) after the onset of the headache. The time interval from symptom onset to diagnosis was not reported for patients #F and #G. All patients experienced awakening due to pain during nocturnal sleep. Five patients (71.4%) woke up a few hours (ranging from 1 to 6 h) after falling asleep, while in two patients (#B and #D), awakening occurred during the early hours of the morning (between 2 am and 5 am). Except for patient #G (where data was not reported), the duration of the headache attacks exceeded 10–15 min, with some lasting up to 60 min and one case lasting 5 h. After experiencing relief from pain, the patients were able to go back to sleep. The mean frequency of attacks was 12 ± 2.8 days per month (ranging from 1 to 25 days per month), with the frequency for patient F being indeterminable due to the relapsing-remitting nature of the condition. Unilateral pain was reported in one case (14.3%), unilateral and bilateral pain in two cases (28.6%), and bilateral pain in 4 out of 7 patients (57.1%). The headache attacks were described as throbbing/pulsating in 3 out of 7 patients (42.8%) and as dull and stabbing in one case (14.3%). All patients reported pain ranging from moderate to severe intensity. None of the patients reported concomitant migraine features, except for occasional nausea in two patients (28.6%) and occasional photophobia in one patient (14.3%). Neurological examinations were unremarkable for all patients. Brain MRI and EEG were performed in 6 out of 7 cases, all of which showed normal results. Two patients had a positive family history of headache with one patient’s mother suffering from migraine with aura. Patient #D had thalassemia major and underwent bone marrow transplantation 6 months before the onset of HH. Their pharmacological therapy included cyclosporine, methylprednisolone, amoxicillin, fluconazole, and acyclovir (14). Patient #G reported sporadic heart palpitations, mitral valve prolapses, seasonal allergies, and obstructive sleep apnoea syndrome (OSAS) but had no other reported comorbidities. Six patients had follow-up clinical examinations for an average of 6.1 ± 4.8 months, ranging from 6 weeks to 12 months. Only two patients received acute-phase medication: patient #B found relief after taking acetaminophen approximately 30–60 min after the onset of the attack (15), while patient #F did not benefit from oral sumatriptan (100 mg), oral rizatriptan (5 mg), or oxygen inhalation. Most of the patients (57%) did not require medication, as each attack completely resolved in about 30 min. Two patients (#D and #E) received prophylactic treatment with melatonin (14). In the first patient (patient #D), an initial dose of 2 mg of melatonin at bedtime was administered. This resulted in a reduction in headache intensity (from severe to moderate) and frequency (from 25–30 attacks per month to 10–15 attacks per month). Consequently, the dosage was increased to 4 mg and over 6 months, the patient became asymptomatic (14). In the second patient (patient #E), an initial dose of 3 mg of melatonin at bedtime was given. The severity of headaches decreased from severe to moderate, and there was an immediate reduction in attack frequency from 12–15 attacks per month to 1–2 attacks per month. The patient remained headache-free during the subsequent 2 months (14). Patient #F did not report any effect in terms of frequency or severity of headache attacks after trial treatment with sodium valproate (500 mg twice daily), amitriptyline (75 mg once daily), duloxetine (60 mg twice daily), naproxen (500 mg twice daily), ibuprofen (600 mg three times daily), or propranolol (60 mg twice daily). However, an immediate response was observed when patient #F took 75 mg of indomethacin at bedtime. Indomethacin was successfully tapered off after 6 weeks, and there was no relapse during the following 3 months of follow-up (18). Patient #G, who also had co-morbid OSAS, reported remission of HH after using an oral appliance for mandibular advancement. However, the follow-up period for this patient lasted only 6 weeks (17). Overnight polysomnography was not conducted for any of the patients included in this review.

When comparing the data of the 7 patients herein described with the latest review of 343 adult patients (6), several characteristics appear to overlap. Both pediatric and adult patients experienced pain exclusively during nocturnal sleep or diurnal naps, resulting in awakening for all cases. The condition predominantly affects females, with a prevalence of 69% in adults and 57% in children, consistent with recent studies (4, 6, 8–10). Bilateral pain is more frequently observed in children (85%) compared to adults (55%), and both populations reported moderate to severe pain intensity (100% in children, 94% in adults). Neither age group exhibited migraine-associated features (57.1% in children, 62.6% in adults) or autonomic symptoms (83.3% in children, 92.4% in adults). However, several differences between the two age groups have been observed. Pediatric patients tend to have a shorter diagnosis latency (15.8 ± 25.0 months, range 1 to 60 months) compared to adults (7.6 ± 14.2 years, range 0.1–39 years). This may be attributed to the heightened concern among caregivers, leading to earlier medical consultations. The frequency of attacks is lower in children (12 ± 2.8 days/month, range 1–25 days/month) compared to adults (21.9 ± 7.6 days/month, range 3–31 days/month). In pediatric patients, attacks occur a few hours after falling asleep in 71.4% of cases, while in adults, most episodes occur between 2 and 4 a.m. (51.1%). Attack duration differs significantly, with adults experiencing longer episodes (93.6 ± 65.3 min, range 10–600 min) compared to children, whose episodes typically last up to a maximum of 60 min, except for one patient who reported pain persisting for up to 5 h. Pain quality is predominantly described as dull or pressure-like in adults (74.4%), whereas throbbing or pulsating character is more frequently reported by children (42.8%). This finding highlight that the differences in the main features of HH between adults and children outweigh the overlapping characteristics. Considering that 5.5% of adult patients had fewer than 10 episodes per month, which is the cut off considered in the diagnostic criteria, the episodic form of HH has been proposed (6), which could align with the characteristics observed in children. The clinical presentation of primary headache disorders, such as migraine, can also differ between childhood and adulthood (20). The ICHD-2 acknowledged some differences in migraine characteristics between the two age groups (11). For example, migraine pain is commonly unilateral in adults but often holocranic in children (21). Differences in the evolution of migraine pain have also been observed (21). The reasons behind these divergent clinical manifestations of certain headache disorders between pediatric and adult populations are not yet fully understood. It has been postulated that differences in myelination degree and synaptic maturation may contribute (21). However, these discrepancies are not adequately accounted for in the latest version of the diagnostic criteria, which may impede the identification of affected children.

One of the reasons why children did not meet the criteria of the ICHD-2 was the requirement of unilateral location (22). Additionally, children did not fully satisfy criterion 3 of the ICHD-2 for diagnosing HH, which requires the headache to occur after the age of 50 years. With the revision of the ICHD-2 (11), significant changes were made to criterion “C”; the occurrence of headache was reduced from 15 to 10 days per month, and the requirement of “first occurs after the age of 50 years” was eliminated (refer to Table 1) (12). According to the ICHD-3, HH can occur at any age, including pediatric cases, although such cases are extremely rare. Furthermore, the ICHD-3 no longer states that the patient must not have more than one migraine feature (such as nausea, vomiting, and photo/phono-phobia). The criteria for “probable HH” in the ICHD-3 are even more inclusive (12). In this review, all seven cases described were diagnosed according to the ICHD-2, and four children did not fully meet the diagnostic criteria (see Table 3). In fact, they exhibited only one of the characteristics of criterion C, which is that the headache lasts for ≥15 min after waking up. As for the ICHD-3, only 2 children did not meet criterion C. However, all patients met the ICHD-3 criteria for probable HH. This data highlights the increased sensitivity of ICHD-3 in diagnosing HH in childhood compared to ICHD-2. Despite the diagnostic criteria for HH being the same for both adult and pediatric patients, certain features remain inconsistent, particularly the frequency of attacks. This study underscores the need for a separate diagnostic criterion for HH in children, emphasizing the lower frequency of attacks in this population.

The intricate relationship between sleep and headache has captivated human curiosity for centuries, yet it remains poorly understood. Sleep is associated with certain headache syndromes, and headaches can disrupt sleep hygiene, while sleep disturbances can trigger or exacerbate the frequency and severity of headache attacks and their comorbidities (23–26). Similar to other headache types, the exact pathophysiological mechanisms of HH in childhood have not yet been fully elucidated. It has been proposed that HH may be a chronobiological disorder (2). Many patients consistently experience headache attacks at the same time during the night, commonly known as “alarm-clock headache.” Hypothalamic dysfunction and dysregulation of serotonin and melatonin have also been hypothesized as contributing factors to the disease (10, 27). The suprachiasmatic nucleus (SCN) of the hypothalamus, which plays a crucial role in the endogenous circadian rhythm, is considered a potential region involved in the onset of HH (28). The number of cells in the SCN significantly decreases with age (10, 29). The SCN projects efferent and receives afferent signals from the brainstem periaqueductal grey, locus coeruleus, and raphe nuclei, which are important brainstem structures involved in pain modulation and sleep (10, 29–31). With advancing age, the activity of the hypothalamic-pineal axis and SCN decreases, resulting in reduced melatonin production, which can be completely absent after the age of 60 years (10, 29, 32, 33). Dysregulation of melatonin synthesis is a well-known documented factor in headache development (2, 31, 34, 35). These age-related changes may explain the significantly higher incidence of HH attacks in adult patients. The hypothesis of posterior hypothalamic involvement in the pathophysiology of HH is supported by findings of decreased grey matter volume in this brain region in HH patients compared to age-and gender-matched healthy controls (36). Headaches associated with sleep can occur during sleep, after sleep, and during different sleep stages (37). It was initially suggested that HH was associated with rapid eye movement (REM) sleep, suggesting it was a REM-sleep disorder (38–41). However, subsequent studies contradicted this notion and demonstrated that HH attacks predominantly occurred during non-rapid eye movement (NREM) sleep (42–46). Further analysis did not reveal a clear REM or NREM subtype of HH, as both types of headache attacks could be observed in the same patient on the same night (42). Instead, HH attacks may be linked to changes in sleep microstructure. In a single patient treated with amitriptyline, a slight increase in cyclic alternating pattern (CAP) rate, from 40 to 46%, was observed despite the improvement of nocturnal symptoms and overall NREM sleep parameters (47). The authors concluded that studying the macro and microstructure of sleep could be valuable in elucidating the pathological mechanisms of HH, although no further studies have pursued this line of research. Additionally, similar to other headache forms, mood fluctuations, changes in daylight exposure, environmental temperature, and lifestyle factors may trigger the recurrence of HH (48).

The main differential diagnosis for HH encompasses secondary headaches and other primary headache disorders (12, 49). In the case of secondary headaches, it is crucial to rule out other potential conditions that can occur during sleep and lead to patient awakening, with specific attention to hypoglycemia (50), medication overuse (12, 51, 52), and sleep apnoea (53). Overnight polysomnography may be necessary in selected cases to exclude OSAS and other sleep disorders. However, the presence of OSAS does not necessarily exclude an HH diagnosis (12, 17). Additionally, intracranial vascular pathologies (54–56) and brain tumours must be considered and ruled out (57–60). Brain MRI is required to investigate any structural abnormalities, such as tumours and it may also reveal a reduction in grey matter volume in the hypothalamus (36). Nocturnal hypertension has also been identified as a potential cause of HH (61–63), and 24 h blood pressure monitoring is recommended in patients with cardiovascular risk factors to assess nocturnal hypertension. Primary headache disorders that tend to occur during sleep, such as migraine, cluster headaches, paroxysmal hemicrania, cervicogenic headache, and short-lasting, unilateral, neuralgiform headache attacks with conjunctival injection and tearing (SUNCT), should also be investigated before reaching an HH diagnosis (12, 32). After that other primary headaches and any secondary or organic causes of pain-induced awakenings are ruled out, the diagnosis of HH is primarily based on clinical assessment and adherence to the ICHD-3 diagnostic criteria.

Recommendations for the acute and prophylactic treatment of HH attacks in children are scarce and primarily based on individual case reports, case series, and literature reviews (6, 14–19). In adults, prophylactic treatment has shown better outcomes (6). Two patients included in this review (patients #D and #E) benefited from melatonin therapy and achieved prompt relief from headaches. These results may provide empirical evidence of the involvement of melatonin and the circadian rhythm system in HH pathogenesis. However, these findings differ from those reported in the adult population, where the response rate to melatonin is approximately 50% (6, 64, 65). Melatonin, a chronobiotic hormone primarily secreted by the pineal gland during the night, has demonstrated therapeutic efficacy in certain forms of headache (66, 67). Patient #F showed improvement with indomethacin treatment. Indomethacin has better blood–brain barrier penetration compared to other nonsteroidal anti-inflammatory drugs, such as naproxen and ibuprofen, and specifically inhibits nitric oxide-induced dural vasodilation (68). This may explain the greater benefit of indomethacin in HH and other headache types (6, 68–70). In patient #G, mandibular advancement with an oral appliance for OSAS resulted in complete control of HH symptoms; however, only a 6 week follow-up period is available for this patient. Although OSAS has been frequently associated with HH, prevalence studies have not confirmed a higher prevalence of OSAS in HH patients compared to unaffected individuals (41). Previous reports on the use of positive airway pressure in HH patients have yielded inconclusive results (10, 39, 40, 70). In one case series, polysomnography recordings in HH patients did not reveal any episodes of OSAS (39). One patient showed improvement with continuous positive airway pressure therapy and supplemental oxygen, while three patients did not (40). Currently, treatment recommendations for HH in children are based solely on anecdotal case reports and uncontrolled trials in small case series, emphasizing the need for prospective trials comparing different prophylactic treatment strategies against placebo. Since HH attacks in pediatric patients generally last less than 60 min, episodes often go untreated. Only patient #B experienced benefit from acetaminophen, while patient #F tested various medications (oral sumatriptan, oral rizatriptan, and oxygen inhalation) without encouraging results. Similar to experiences with adult patients, lithium and caffeine may prove effective in the pediatric population, but larger placebo-controlled studies are required to validate this hypothesis (6).