CGRP and PACAP-38 play an important role in diagnosing pediatric migraine

Migraine is a neurovascular disorder that is common and multifactorial. It is a major cause of disability globally, but the medical literature and studies on the subject have been scarce [1]. Especially in children and adolescents, the diagnosis and treatments of migraine are largely ignored. Migraine is a common headache disorder in children and adolescents [2]. Headache disorders were one of the top ten causes of global DALYs (disability-adjusted life-years) for adolescents in 2019. Headache disorders were the second leading cause for individuals aged 10–24 years, which was just lower than road injuries [3]. A study of the prevalence of migraine in Taiwan suggested that the prevalence was 4.8%, 7.1%, and 8.4% for individuals aged 13, 14, and 15 years, respectively [4]. Many studies have also suggested that the prevalence of migraine increases with increasing age [4, 5] and is consistently higher for girls than boys [4‐6]. Migraine can cause absence from school and impaired performance in homework and activities, which result in significant disability in the lives of children and adolescents.

Currently, the diagnosis of migraine follows the third edition of the International Classification of Headache Disorders (ICHD-3), but there are many differences in symptoms between adults and children. Although unilateral location is frequently described in adult patients, the typical headaches in children and adolescents tend to be more bilateral [7]. Another study suggested that younger children have a shorter history of this disease, a reduced frequency of attacks, and a shorter duration of episodes than older children [8]. The diagnosis of migraine is more dependent on the description of the clinical manifestation after normal imaging and laboratory examinations have ruled out other secondary headaches. The difficulty of diagnosing migraine in children is undoubtedly increased by the fact that the child’s memory and description of the symptoms of the headache attack are unclear.

Migraine is defined as a strong genetic component and involves the activation of trigeminal pain pathways. Currently, cortical spreading depression (CSD) pathogenesis is highly valued [9, 10] and has been reported to activate trigeminal and parasympathetic pathways [10]. The activation of the trigeminal ganglion causes the release of these vasoactive peptides, particularly calcitonin gene-related peptide (CGRP) and pituitary adenylate cyclase-activating polypeptide-38 (PACAP-38), from sensory nerves terminals and thus these vasoactive peptides are implicated in pain pathways [11, 12]. Many studies have reported that the plasma levels of CGRP and PACAP-38 in migraine patients are higher than those in healthy people [13‐17]. Another clinical trial involves infusing PACAP-38 into healthy participants which also confirmed that the elevation of plasma PACAP-38 can cause migraine attacks [18]. Furthermore, CGRP receptor antagonists and monoclonal antibodies toward CGRP and CGRP receptors showed a positive relief of acute and chronic migraine in clinical trials [19‐24].

To date, literature has provided contradictory findings on the association between plasma CGRP, PACAP-38 levels, and pediatric migraine. V Gallai et al. found that CGRP levels in pediatric migraine patients with or without aura were significantly increased during attacks compared with the interictal period [25]. However, Fatma Hanci et al. found that PACAP-38 levels in both ictal and interictal plasma were higher in migraine children without aura than in controls, but CGRP remained unchanged [26]. It remains to be explored whether PACAP-38 and CGRP work in pediatric migraine as a result of the contradictory findings. In addition, there is little literature on whether plasma CGRP and PACAP-38 levels can improve the diagnostic accuracy of migraine in children.

This study prospectively investigated plasma levels of CGRP and PACAP-38 in children with migraine. We compared the plasma CGRP and PACAP-38 levels in patients with age-matched healthy children and analyzed the results according to different subgroups of patients. We also explored the correlation between vasoactive peptides and clinical characteristics of pediatric migraine. Furthermore, we also analyzed the diagnostic value of CGRP, PACAP-38, and the combination of CGRP and PACAP-38 in pediatric migraine.

Our findings indicated that plasma CGRP and PACAP-38 levels in children with migraine were significantly higher than those in control group. Similarly, Fan PC et al. have reported pediatric patients with migraine had higher plasma CGRP levels than non-headache children [28]. We found that plasma CGRP and PACAP-38 levels were higher in children with migraine than in control group during both attacks and non-attacks, and the plasma CGRP and PACAP-38 levels were elevated in the migraine with and without aura group than in control group, same results as for some adults with migraine [29, 30]. Hanci F et al. found higher plasma PACAP-38 levels, but not CGRP levels in children with migraine without aura than in healthy children during both attacks and non-attacks [26]. P-C Fan et al. described that CGRP levels were higher in the ictal phase than in the controls, but no difference was found between the interictal phase and healthy children [28]. These findings indirectly suggest that CGRP and PACAP-38 play an important role in the pathogenesis of migraine in children. The mechanisms of CGRP mediating pediatric migraine mainly are involved in the vasodilation theory [31, 32], neurogenic inflammation [33, 34], peripheral and central sensitization [35, 36], and cortical diffusion inhibition diffusion (CSD) [37‐39], and nitric oxide generation. However, CGRP alone cannot explain all clinical manifestations of migraine and may have a cooperative action with other peptides in inducing migraine attacks, such as PACAP [40]. Interestingly, PACAP has several similar actions with CGRP in inducing migraine attacks, and PACAP also triggers CGRP release in the TNC, which researchers speculate is a potential epistatic modulator of CGRP in migraine [40], suggesting that CGRP and PACAP might work together to induce migraine attacks, which is supported at a molecular level by the fact that CGRP and PACAP receptors share a RAMP1 (receptor activity modifying protein 1) subunit [14].

Because migraine attacks with aura symptoms and migraine in the ictal period can cause more discomfort for patients, we paid more attention to groups with aura symptoms and groups in the ictal period. The findings indicated that in the migraine with aura group, CGRP levels in the ictal phase were higher than those in the interictal phase, suggesting that CGRP may be involved in the formation of pain in children with migraine [36, 41] since pain is perceived by patients during migraine attacks. CGRP can sensitize peripheral nociceptors and enhance central nervous system sensory input to heighten pain perception. CGRP is expressed on almost 50% of human trigeminal ganglion (TG) neurons, most of them on unmyelinated nociceptive C-fibers. CGRP receptors are expressed on myelinated A-fibers. When trigeminal neurons are activated, CGRP is released from C-fibers, and act on receptors of A-fibers to stimulate second-order trigeminal nucleus caudalis (TNC) neurons and then relay to the thalamus, ultimately causing central sensitization [42‐44]. In migraine in the ictal phase group, we found that plasma CGRP levels in the migraine with aura group were significantly higher than those in the migraine without aura group. Gallai V et al. have reported increased CGRP levels in pediatric migraine with aura compared with migraine without aura in the ictal period, indicating that CGRP may induce aura symptoms of migraine in children [37‐39]. Several findings support that CGRP can influence the formation of aura symptoms by mediating the mechanism of CSD. CSD is related to aura symptoms in migraine, which may interact with CGRP. During CSD, more endogenous CGRP could be released, which was revealed by using rat cortical brain slices [37]. CGRP receptor antagonists weakened pial dilation induced by CSD [38, 39] and the magnitude of the CSD effect in vitro [45].

Multiple linear regression analysis showed that plasma CGRP levels correlated to nausea, which is consistent with the existence of CGRP in the enteric nervous system, involved in regulating gastrointestinal motility and secretions [14]. But there were no correlations between CGRP plasma levels and other gastrointestinal symptoms, such as vomiting and abdominal pain. We think lots of factors can take responsibility for gastrointestinal symptoms. For example, some patients have dyspepsia or an irregular diet, such as spicy food, eating too much, and eating food before sleeping, which also can cause gastrointestinal symptoms. These correlations between CGRP, abdominal pain, and vomiting may be influenced by the inability to remove confounding factors due to the limited clinical information available.

Multiple logistic regression showed that CGRP and PACAP-38 are independent risk factors associated with the diagnosis of pediatric migraine. ROC analysis suggested that CGRP and PACAP-38 are of great value in diagnosing migraine in children because of their high AUC, sensitivity, and specificity. Fan PC et al. have reported similar sensitivity, specificity, and positive likelihood ratio of CGRP in diagnosing pediatric migraine [28]. Considering the similar plasma level changes of CGRP and PACAP-38 in our studies and the similar effects in the pathogenesis of migraine discussed in previous studies, we considered the combined diagnostic value of CGRP and PACAP-38. The findings indicated that the combination of CGRP and PACAP-38 had the greatest AUC and specificity compared to CGRP and PACAP-38, suggesting its superior value in diagnosing pediatric migraine and differentiating pediatric migraine from non-migraine. To the best of our knowledge, this is the first study to analyze the diagnostic value of PACAP-38 in pediatric migraine and the first to combine CGRP with PACAP-38 as a combination marker to diagnose migraine in children.

Furthermore, we found that the duration of attacks of some children with migraine was less than 2 hours, which cannot be explained by the selection bias that would be more likely to select children with longer pain episodes. Battistella et al. and Raieli V et al. have reported the shorter duration of headaches in younger children, which can be by active mechanisms (e.g., sleeping), rather than physiological pain mechanisms [8, 46]. Sleeping can alleviate headache attacks, and poor sleep quality might increase the risk of headache attacks. Preschool children, especially those under 6 years old, sleep longer and they are usually more likely to sleep during the day [46]. Therefore, the headache attacks are easily stopped and cause a shorter duration per attack.

And almost half of the children with migraine have aura symptoms and most of them have concomitant symptoms. The VAS score showed that 79% (66) of patients have moderate to severe migraine. Migraine seriously affects the healthy quality of life of children, worthy of attention of the clinicians.

Our study found significant elevation and almost the same changes in CGRP and PACAP-38 levels in pediatric migraine, which suggested that CGRP and PACAP-38 play an integral role in pediatric migraine and may work together in the pathogenesis of migraine in children. Multiple logistic regression showed that CGRP and PACAP-38 are independent risk factors for pediatric migraine. Moreover, there was a high AUC, the best sensitivity, and a higher OR of CGRP and PACAP-38, suggesting that CGRP and PACAP-38 are good indicators to diagnose pediatric migraine, and the combination of CGRP and PACAP-38 has a great value in diagnosing pediatric migraine and distinguishing migraine and non-migraine headache in terms of the largest AUC and specificity. Higher CGRP levels were found in the ictal phase group than in the interictal phase and in the aura group than in the without aura group, which indicated that CGRP may take part in the formation mechanism of pain and aura. Therefore, in the future, more prospective clinical studies including more patients can explore the sensitivity and specificity of CGRP and PACAP-38 in diagnosing migraines. Furthermore, more basic studies should pay more attention to the association of molecular pathways between CGRP and PACAP-38.

Children younger than 6 years have an unclear memory and description of the symptoms of the migraine attacks. For example, they cannot describe the site or the severity of migraine attacks clearly. If necessary, the description of the guardians of the younger children should be referred. The information we collect (frequency of migraine, presence or absence of aura, etc.) depends on recall of the patients and/or parents. Therefore, recall bias may occur. In collecting the data, we gave the patient sufficient time to recall and combine it with their guardian’s description in order to try to minimize the possibility of bias. In addition, selection bias was inevitable and all researchers preferred to select the more severe and typical cases. With regards this, information collection, data analysis and interpretation of results are the responsibility of different researchers and they were prohibited from discussing the clinical information collected and trial data with each other.