Paroxysmal hemicrania in children and adolescents: A systematic review

We aimed to report the accessible demographic, clinical, and radiological characteristics of reported pediatric paroxysmal hemicrania (PH).


INTRODUC TI ON
Trigeminal autonomic cephalalgias, including cluster headaches, paroxysmal hemicrania (PH), and short-lasting unilateral neuralgiform headache with conjunctival injection and tearing/cranial autonomic features, are approximately 100 times less prevalent than migraine. 1 They are defined by severe unilateral short episodes of pain. They share rhinorrhea, nasal congestion, lacrimation, and conjunctival injection as ipsilateral cranial autonomic symptoms. They do, however, differ in duration, frequency, and type of treatments. 2 PH is classified as chronic paroxysmal hemicrania (CPH) and episodic paroxysmal hemicrania (EPH) in the International Classification of Headache Disorders, 3rd edition (ICHD-3). 3 In EPH, attacks occur in periods lasting 7 days to 1 year, separated by pain-free periods lasting 1 month or longer. In CPH, attacks occur for more than 1 year without remission or with remissions lasting less than 1 month. 4,5 Reports from clinical reviews and expert opinions often note that PH typically begins in adulthood. [6][7][8] However, reports have indicated a several-year delay in diagnosis, which could be attributed to atypical headache features in children and adolescents (migraine, for example, can have a shorter duration in pediatric patients). 3 While PH in adults is a well-known diagnosis among neurologists, current awareness and knowledge of PH in children are not satisfying.
Even though it has been a long time since the first presentation of PH in a child was reported, 9 it is unknown whether PH in children follows the same patterns as in adults. 10 There is a lack of awareness of PH in pediatricians, 11 as well as difficulty in taking children's history of attacks and pain characteristics.
This study aims to describe the broad age ranges of PH in children and adolescents and to report the accessible demographic, clinical, and radiological characteristics of reported pediatric PH, as well as their response to treatment.

Study selection
The protocol of this study was registered with PROSPERO (ID: CRD42022306190). We followed the latest version of the PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analyses) guidelines. 12 We searched for studies up to January 2022 without time limitation in the following databases: PubMed, Web of Science, and Scopus. A literature search was conducted for studies investigating PH among children or adolescents under 18 years of age using combinations of the following keywords: "paroxysmal hemicrania" AND ("infants" OR "children" OR "adolescents") and their equivalent keywords and phrases based on Medical Subject Headings of the US National Library of Medicine (Table S1 in supporting information). The relevance of these terms was determined based on some previous reviews. 2,[13][14][15] Our search strategy in different databases is available in Table S2 in supporting information. Moreover, references from relevant articles were evaluated for potentially missed studies. After excluding duplicate publications, two researchers (MB and VM) independently screened titles and abstracts of the retrieved studies. Likewise, studies that remained of interest were then screened based on their full text by MB and VM.
In the screening and eligibility steps, disagreements were settled by discussion between the two researchers. In all cases, a consensus was reached.

Inclusion and exclusion criteria
Studies included in the present systematic review met the following criteria: (1) all types of observational or interventional studies, (2) published in English-language, peer-reviewed journals, Meeting abstracts and similar articles that may have less strict peer review, as well as articles published prior to 1988, were excluded. 16 Although we did not limit the date in our primary search, we chose 1988 as the starting date for screening because the first edition of the ICHD was published that year. 16 In addition, because of a wide range of publication dates from 1989 to 2021 based on our primary search, the ICHD version varied between 1 to 3. However, we discussed all symptoms based on ICHD-3 criteria. Furthermore, studies were screened for duplicate data based on authors, publication year, participant numbers, and characteristics. When possible duplicate data were found, the authors were contacted to clarify whether the data sets were independent.
indomethacin and showed complete responses to treatment, while some needed combination treatment of indomethacin with other medications.

Conclusion:
Although pediatric-onset PH has similar features to adult-onset PH, there are some challenges with ICHD criteria for younger children that limit the ability to confidently assign a diagnosis. Moreover, owing to concomitant migrainous features, PH may be confused with migraine in children and adolescents.

K E Y W O R D S
adolescent, child, headache, paroxysmal hemicrania

Data extraction
Two authors (MB and VM) independently extracted data elements from each article. Included studies were entered into a prepiloted form for assessment of study quality and evidence synthesis. The items extracted from the included studies were first author; publication year; country; PH class; sex; age; age of onset; frequency, duration, site, severity, and quality of pains; triggers; symptoms accompanied by headaches, including conjunctival injection, lacrimation, nasal congestion, rhinorrhea, eyelid edema, forehead/facial sweating, miosis and ptosis, nausea, photophobia, and phonophobia, as well as a sense of restlessness/agitation; response to treatment; laboratory investigations; imaging; comorbidity; and family history.

Statistical measures
The mean of the continuous variables, including the age of onset and age of diagnosis, was calculated from the included data. Also, the sex ratio was calculated by dividing the number of boys by the number of girls. SPSS version 24 (SPSS Inc.) was used for calculations.
Studies that missed the corresponding measures were excluded from the parameter calculation.

Quality assessment
For quality assessment, two independent reviewers (MB and VM) assessed the methodological quality of the included studies through the Joanna Briggs Institute's (JBI's) critical appraisal checklist, which is a reliable and valid quality index for the appraisal of case reports and case series. 18,19 Discrepancies were resolved through discussion and the involvement of a third reviewer (MSD) where necessary.
JBI's critical appraisal checklists were structured into fixed sets of questions for case reports and case series of 8 and 10 questions, respectively, focusing on different aspects of bias in the study's design, conduct, and analysis. Bias was assessed as a judgment with yes, no, unclear, or not applicable answers for each question. Prior to critical appraisal evaluation, decisions about the scoring system and the cutoff for inclusion of a study in the review were determined in advance and agreed on by all participating reviewers. All reviewers agreed that studies with three or more negative responses were assumed to be of low methodological quality and should be excluded from the final results.

RE SULTS
A total of 182 records were identified and reduced to 116 after removing duplicates. After screening and the application of the eligibility criteria, 22 articles 9,10,20-39 met the inclusion criteria. Figure 1 presents the flow diagram of the included and excluded articles in our review according to PRISMA. 40 Owing to the lack of detailed case presentation in some cases, we could not ascertain the diagnosis of PH by the latest version of ICHD according to the provided information (Table 1). Instead, we decided to rely on the authors' claims about the diagnosis mentioned in each article. However, we made sure that the authors made the diagnosis based on the latest version of ICHD according to the publication date.  To the best of our knowledge, the first pediatric case of PH was reported by Kudrow in 1989. 9 Among previous studies in the pediatric population, CPH, EPH, and unspecified PH were reported in 23, 7, and 5 patients, respectively. PH has been described among children and adolescents in various parts of the world, but the races were not specified in most of the studies. With the exception of one study, the sex was reported in all cases. 9 In our review of all pediatric PH cases, we found a boy-to-girl ratio of 1.125:1 (18 boys vs. 16 girls). The youngest age of onset was reported by de Almeida et al. 23 They reported a 10-year-old girl who had complained of left unilateral headache attacks since she was 1 years old. After excluding the cases with unknown ages of onset or diagnosis, 10,20,33,38 the mean age of onset among reported cases in children and adolescents was 6.5 years, while the mean age of diagnosis was 8.2 years. [Correction added on 22 August 2022, after first online publication: In the preceding sentence, 6.3 and 7.9 years were changed to 6.5 and 8.2 years, respectively and the references were changed from 9,20,33,34 to 10,20,33,38]. The shortest course of pediatric PH was reported by Ishii et al. 27 in an 11-year-old boy from Japan, who underwent treatment with indomethacin within 2 days of his first referral. On the other hand, the longest course was reported in a 10-year-old girl who experienced attacks since she was a 1-year-old infant. 23 The most frequent attacks 20 were attributed to a 14-year-old girl with 32-48 attacks per day and the least frequent attacks 23 were described in a 10-year-old girl with one attack per week. The duration of attacks varied between 2 minutes and 40 minutes. 20,23 The pain was typically unilateral, but could alternate sides between attacks 38 and was only rarely bilateral. 32 Left-sided pain was approximately twice as common as right-sided headaches (11 left sided vs. 6 right sided). The characteristics of the pain in PH were usually severe in intensity (led to crying/screaming, awakening the child at night, and wishing to die) in almost all cases and have been described by patients as sudden, sharp, stabbing, shooting, throbbing, and excruciating.
Attacks of PH invariably occur in association with ipsilateral cranial autonomic features in all cases except six children, which we noted above. 10 [Corrections added on 22 August 2022, after first online publication: The misspelling of the author names "Muritz" and "Borious" was corrected to "Mauritz" and "Borius" and the text in the columns "Site" and "Quality" was shifted to the appropriate columns. Further details were added throughout Table 1  we found several cases with family history positive for migraine 25,[32][33][34][35]37 and unspecified headaches, 28,30,33,34 breath-holding, 35 and bipolar disorder. 22 Almost all patients benefited from indomethacin and showed complete response to treatment, 21,22,[24][25][26][27]29,30,32,33,[35][36][37]39  One study suggested using a headache diary as both an idea for diagnosis and an initial nonpharmacological therapeutic intervention in children with PH. 28 In addition, a recent study described a child treated by occipital nerve stimulation for CPH. To the best of our knowledge, this is the first child reported to undergo this new treatment option, which is used for patients who are not responding to other treatments. 39

DISCUSS ION
To the best of our knowledge, PH was first described in two adult patients by Sjaastad  Thereafter, to the best of our knowledge, the first pediatric patients with EPH were reported by Blankenburg et al. 21

in 2009.
PH is a rare condition. The incidence and prevalence of PH in children and adolescents, as well as adults, are unknown. Although there are no data available, the observed frequency in pediatric clinics seems comparable to what can be seen in adult settings. 1 The prevalence of PH was estimated to be 1% to 3% of the prevalence of cluster headaches. 6 Given that the prevalence of cluster headaches 47 is approximately 1 in 500, the prevalence of PH would be approximately 1 in 25,000. We identified that PH has been described among the pediatric population in three continents (excluding Africa, Australia, and Antarctica). However, epidemiologic data are scarce, and many cases of PH are probably still overlooked.
The current literature on pediatric PH suggests that PH can start early in life, with the youngest documented case at 1 year old. We In our review, we found a boy-to-girl ratio of 1.125:1. This result is in contrast to some recent retrospective studies in adults that showed a female predominance. 6,7,48 Two other studies in the adult population reported an approximately equal prevalence in women and men. 8,49 Comprehensive and well-designed epidemiologic studies are needed to find the true extent of the sex ratio in PH.
In these 38 cases that we systematically reviewed, there was no evidence of a family history positive for PH whereas we found family history positive for breath-holding, bipolar disorder, migraine, and This study was not included in our review because it did not fulfill the inclusion criteria. There is also a report on hemicrania continua, which belongs to the classification of trigeminal autonomic cephalalgias, in a family (a 44-year-old mother presented with hemicrania continua for 32 years and her 23-year-old daughter with hemicrania continua for 10 years). 51 Genetic studies on PH are hampered by small sample sizes.
The clinical features of PH based on different versions of ICHD criteria are given in Table 2.
In the ICHD criteria, 4,16,17 it was noted that attacks have a frequency of more than five per day for more than half the time, although periods with lower frequencies may occur. In several cases, the frequency of PH attacks was reported to be lower than five attacks per day. 10 in which cranial autonomic features and restlessness occur at a lower rate in pediatric patients. 52 We also suggest that these characteristics might be under-or over-diagnosed, particularly in younger children.
Furthermore, the safety and efficacy of indomethacin in pediatric pa- limits the applicability of an "indomethacin trial" as a criterion of PH.
Concerns about liver or pancreas problems, stomach bleeding, and kidney damage should be addressed as a result of long-term usage of indomethacin as a prophylactic medication. In addition, indomethacin might worsen asthma in children 54,55 and is contraindicated in patients with a history of asthma. 53 As asthma was a common comorbidity in children with PH, 27 Several limitations should be considered for this study. First, we considered articles that made the diagnosis of PH using the ICHD criteria, even if all the diagnostic criteria items were not mentioned in the article. We could not ascertain the diagnosis of PH by the latest version of ICHD because not all cases included all the necessary criteria of ICHD-3 in their presentations. Second, given the design of the included studies, which were merely case reports or case series, the findings must be interpreted with caution. This limitation is mostly related to the concern that a disproportionate number of atypical cases of PH are more likely to be reported in case reports and case series, which can lead to bias.
Moreover, case reports and case series are uncontrolled study designs that are susceptible to bias, so this information may not be generalizable to the larger population. 56 Third, an incomplete evaluation of atypical and undifferentiated PH-like headaches in pediatrics might occur as our search criteria may not have been mentioned in the title or abstract.
Additionally, it is reasonable to think that the full range of atypical or undifferentiated PH-like headaches are not reported in the literature and require a methodology other than a systematic review (such as an observational study of atypical PH). Finally, during the screening step, we decided to consider the age of diagnosis, not the age of onset, to decrease the heterogeneity of the included studies. During our primary search, we realized that most case reports or case series focused on the clinical characteristics of patients during adulthood. We assumed that even if the age of onset was in childhood in these cases, the symptoms might change eventually, as some evidence declared that the nature of paroxysmal headaches changes over time, 35,36 or even changes to another type of headache. 33 Thus, we tried to focus on the studies on child and adolescent populations.

CON CLUS ION
In summary, the current literature strongly suggests that PH can

ACK N OWLED G M ENT
We are very grateful to Dr. Negin Badihian for her kind assistance with the systematic review search.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no competing interests.

PROTO CO L R EG I S TR ATI O N
This study was registered as a systematic review in PROSPERO (CRD42022306190).