Clinical manifestations

Presentation and course

The hallmark of abdominal migraine is recurrent abdominal pain, which is frequently defined as recurrent, discrete episodes of abdominal pain with clear-cut, symptom-free intervals. It is a common disorder that has been estimated repeatedly over the last several decades to affect 10% to 20% of all school-aged children, with a slight increase in females after 9 years of age (73). About 5% to 10% of children with recurrent abdominal pain have an underlying organic process that contributes to their pain. In an effort to improve the understanding of this syndrome, a panel of pediatric gastroenterologists who are experts on functional gastrointestinal disorders previously proposed classifying children with recurrent abdominal pain into five subtypes: functional dyspepsia, irritable bowel syndrome, functional abdominal pain, abdominal migraine, and aerophagia (70).

By using similar definitions as those published by the ICHD-II (35), the mean age of onset of abdominal migraine is reported to be 7 years, with peaks at 5 and 10 years, similar to childhood migraine (01). It is characterized by recurrent, acute-onset, incapacitating, non-colicky, midline abdominal pain that lasts for hours and is accompanied by pallor and anorexia. Vomiting may be an accompanying feature but is often less severe in comparison with cyclic vomiting (53). There is usually a history of migraine headache, either in the child or the family, but headache may be minimal or absent during the attacks. Episodes of abdominal pain are paroxysmal and separated by symptom-free intervals of weeks to months. In the same population-based study (01), abdominal pain in children with abdominal migraine was usually described as “just sore” or “dull” (60%) but could be colicky (22%). It was periumbilical in most (78%) but could be felt diffusely in 16%. The pain was occasionally preceded by nonspecific prodromal symptoms, such as behavior or mood changes or anorexia (14%). Attacks of abdominal pain occurred on a mean of 14 times per year (range: 2 to 100 times), with each attack lasting for a mean of 17 hours (range: 1 to 72 hours). Children with abdominal migraine lost a mean of 3.7 school days (range: 0 to 36 days) per year due to abdominal pain. Abdominal migraine was recognized again as an abdominal pain-related functional gastrointestinal disorder by the Rome III Child/Adolescent Committee with its criteria based on scientific evidence whenever possible (69). Functional impairment, which has always been considered a major issue in migraine, was for the first time being included in the definition of the criteria. A survey done in a Japanese pediatric clinic of 3611 cases from May 2010 to April 2015 showed a low prevalence of recurrent abdominal pain (1.47%) as well as abdominal migraine (0.19%) (38). A survey was completed in emergency departments in Italy and France on children and adolescents (ages 6 to 17) (50). Compared with controls, patients with migraine but not tension-type headache were associated with a higher chance of having functional dyspepsia (odds ratio (OR)=10.76), abdominal migraine (OR=5.87), and irritable bowel syndrome (OR=3.47).

Abdominal migraine is usually considered a benign childhood condition, although rare patients develop severe complications. Using their own criteria of abdominal migraine, Dignan and colleagues found that abdominal migraine had resolved in 61% of patients, and 70% of patients were either current (52%) or previous (18%) sufferers from International Headache Society (IHS)-classified migraine at follow-up at the mean age of 17 years (25). It has been strongly suggested that the diagnosis of abdominal migraine can be regarded as proven only when detailed inquiry and follow-up have revealed that the patient has suffered from migraine headache (42). Nevertheless, this point of view has not been considered a necessary criterion in the ICHD-III (36) or Rome III pediatric criteria (69) for abdominal migraine.

Cyclic vomiting is one of “the childhood periodic syndromes that are commonly precursors of migraine” (“migraine equivalent”). It is also a self-limiting episodic condition of childhood, with periods of complete normality between episodes. This disorder was not included as a childhood periodic syndrome in the ICHD-I (45). The clinical features of this syndrome resemble those found in association with migraine headaches, and multiple threads of research over the last years have suggested that cyclic vomiting is a condition related to migraine. It is occasionally used interchangeably with abdominal migraine (21). Clinical comparison can hardly distinguish these two syndromes. The mean age at onset is 5 years, and it usually demonstrates predominance in females (66; 52). Affected children typically experience eight attacks per year, and the mean duration of each attack ranges from hours (02; 66; 52) to days (26). Patients with cyclic vomiting syndrome are frequently comorbid with anxiety or depression (49). One study showed that children with cyclic vomiting had a lower health-related quality of life than those with irritable bowel syndrome, and both children and their parents reported a lower quality of life than healthy controls (80). Generally speaking, in abdominal migraine, the predominant symptom is abdominal pain, whereas in cyclic vomiting, the predominant symptom is vomiting (78). The updated ICHD-III criteria for cyclic vomiting syndrome (code 1.6.1.1) are as follows (36):

Similarly, the Rome III pediatric committee also proposed diagnostic criteria for cyclic vomiting syndrome, either for infants and toddlers or children and adolescents. The diagnostic criteria include both of the following (69):

The major difference between the two criteria for cyclic vomiting syndrome was similar to that of abdominal migraine, ie, the number of episodes needed to suffice the diagnosis: two for the Rome III criteria, and five for the ICHD-III criteria.

There are several reasons for the limited recognition of abdominal migraine among professionals. Experts in migraine tend to be neurologists who, by the nature of their specialty, do not usually see children with recurrent abdominal pain. In addition, abdominal pain is not a common feature of adult migraine (72). A study in the United States found that among a group of 458 children with chronic, idiopathic, and recurrent abdominal pain, abdominal migraine represented about 4% to 15% based on either ICHD-II or Rome III criteria (20). Unfortunately, none of them received such a diagnosis. Therefore, the pediatricians or pediatric gastroenterologists are much more familiar with this syndrome. Referral to these specialties is warranted before the diagnosis of abdominal migraine is made.

Prognosis and complications

Abdominal migraine is usually considered a benign childhood condition. A significant portion of patients with abdominal migraine have migraine or develop migraine when they grow up. The frequency of IHS migraine occurrence in patients with abdominal migraine was much higher than in controls (20%) (25). A single case report demonstrated the development of late-onset ophthalmoplegic migraine in a patient with previous childhood abdominal migraine, but whether this presentation was coincidental or had plausible pathophysiological basis deserves further exploration. Unlike cyclic vomiting, patients with abdominal migraine rarely have complications. However, to prevent misdiagnosis, it is crucial to search for underlying organic lesions before making the diagnosis. Furthermore, it has been reported that a substantial proportion of children with cyclic vomiting syndrome continued to experience other somatic symptoms such as headache (42%) and abdominal pain (37%) with or without resolution of vomiting (29). A remarkable proportion of adult patients with cyclic vomiting syndrome are reported to have depression (78%) and anxiety (84%) (61). A biopsychosocial approach is imperative to alleviate the burden of the illness.

Clinical vignette

A 7-year-old boy was referred by the family physician, who wondered whether the child's abdominal pain could be due to migraine. The boy had improved with pizotifen and, as a result, was eating better and had less abdominal pain.

The child indicated that his umbilicus was the source of the pain, but both the mother and the child were vague about the duration of the pain–sometimes lasting minutes, sometimes longer. There was no associated vomiting, and headache was never present with the abdominal pain.

The mother had migraine without aura.

A general medical examination showed a thin youth without abnormalities of the nervous system.

The neurologist advised carefully recording pain frequency, duration, and source. The patient’s mother was also advised that a referral could be made, if necessary, to a pediatrician or a gastroenterologist.

Biological basis

Etiology and pathogenesis

The exact cause of abdominal migraine is unknown. Those patients with established migraine—interspersed with abdominal components of the attacks—highlight our ignorance of the site of origin and mechanism of migraine. Vascular, biochemical, and neural mechanisms have been offered as causes of migraine; the conclusions favor a neurologic etiology (10). Cyclic vomiting linked with ictal generalized epileptiform discharges on electroencephalogram in some patients suggests a neuronal origin and possible channelopathies (63). Genetic factors, such as calcium channels, dopamine and insulin receptors, Na+/K+ ATPase pump subunits, and components of mitochondrial energy metabolism, have all been linked to influencing the risk of developing migraine or its variants (27). Further, Boles and colleagues have demonstrated maternal inheritance in children with cyclic vomiting syndrome with or without neuromuscular disorders, suggesting a common predisposing genetic factor of mitochondrial DNA (13; 14).

The same research group has reported that two common mitochondrial DNA polymorphisms, 16519CàT and 3010GàA polymorphisms, are highly associated with migraine headache and cyclic vomiting syndrome in children with haplogroup H (87). However, these two polymorphisms are not associated with adult-onset cyclic vomiting syndrome, suggesting a degree of genetic distinction between pediatric and adult-onset cyclic vomiting syndrome (16; 82). Lee and colleagues demonstrated that the mitochondrial RYR2 gene, encoding a stress-induced calcium channel present in many neurons, was the only gene demonstrating a statistically significant difference in the proportion of conserved sequence variants among patients with cyclic vomiting syndrome and controls by using a next-generation sequencing platform (18/75 cyclic vomiting syndrome, 24%, vs. 3/60 controls, 5%; p=0.0018, OR=6.0, 95% CI=1.7-22) (48). These genetic components might also play a role in the etiology of abdominal migraine; however, at this point, more research evidence is needed to investigate or replicate these findings.

The pathogenesis and pathophysiology of abdominal migraine are unknown. Indirect mechanisms have revealed that abdominal migraine and migraine might share some common mechanisms. In a population-based study of schoolchildren in Sweden, children with abdominal migraine had demographic and social characteristics similar to those of children with migraine. It has also been noted that patients with abdominal migraine had similar patterns of associated recurrent painful conditions, trigger and relieving factors, and associated symptoms during attacks (01). Children with abdominal migraine are more likely to develop migraine without aura (25). In addition, several small-scale clinical trials also showed that patients with abdominal migraine were responsive to the non-analgesic migraine prophylactic agents propranolol, cyproheptadine, and pizotifen (86). The similarities between abdominal migraine and migraine are enough to suggest that they have a common pathogenesis. One report demonstrated the effectiveness of sumatriptan in aborting abdominal migraine and childhood functional abdominal pain syndrome and further strengthened the link between migraine and abdominal migraine (46). On the other hand, it has been suggested that cyclic vomiting syndrome may be an anticipatory stress response in migraine, in which emesis-induced or nausea-induced arginine-vasopressin release mediates a key neuroendocrinological role (33). It might be possible that abdominal migraine bears similar mechanisms. It was shown that mitochondrial DNA control region sequence variations are associated with both migraine without aura and cyclic vomiting syndrome (83). Additional findings also suggest that the mitochondrial C16519T and G3010A polymorphisms might predispose the development of migraine without aura and cyclic vomiting syndrome through an effect on energy metabolism (87). These studies suggest that energy depletion plays a role in the pathogenesis of migraine and its variants and also points to possible therapeutic strategies. As with migraine, the pathogenesis of abdominal migraine might be multifactorial. Further studies are needed to explore the underlying mechanisms.

Epidemiology

Unfortunately, the methods used in surveys and in making diagnoses should lead to critical questioning of the published results and conclusions. Because of the controversy about the existence of the disease entity of abdominal migraine and the unsettled diagnostic criteria, studies on the prevalence in the clinic- or community-based cohorts are rare. In the community, a UK study showed that the prevalence of abdominal migraine with or without migraine was 2.4% (59). Using similar criteria to those published in the ICHD-II (35), the prevalence of abdominal migraine in 2165 children aged 5 to 15 years was 4.1% in a Swedish population of school-aged children (01). A Turkish study reported that the prevalence of cyclic vomiting syndrome in urban school children was 1.9% with a 2:1 female-to-male ratio (26). An Irish study found an annual incidence of cyclic vomiting syndrome of 3.15 per 100,000 children (95% CI 2.19-4.11) (30). A Japanese study based on claims data reported a low prevalence of 0.32 per 1000 population in the general population, with 2.10 in 10,000 in pediatric and 0.05 per 1000 in adult population (23). Of note, one Italian study using the Rome III criteria showed a marked increase of the prevalence of abdominal migraine from 5.1% in 2018 to 13.9% in 2021 in children and adolescents during the COVID-19 pandemic (28).

Migraine equivalents in infancy, childhood, and adolescence are common in daily pediatric practice. A retrospective clinical database study estimated that 108 of 5848 patients (1.8% of total, 9.8% of migraineurs) had migraine equivalents. These included 11 with benign paroxysmal torticollis (10.2% of patients with migraine equivalents), 41 with benign paroxysmal vertigo (38%), 20 with abdominal migraine or cyclic vomiting (18.5%), 31 with acephalgic migraine (28.7%), and five with acute confusional migraine (4.6%). More typical coexisting migraines were observed in 10% (benign paroxysmal torticollis) to 70% (abdominal migraines or cyclic vomiting) of the cases (04). Of note, one retrospective study in Taiwan reported that seven patients (35%) with cyclic vomiting syndrome (n=20) developed migraine (55). In a Bulgarian pediatric neurology clinic, Pacheva and Ivanov reported that migraine variants accounted for 24.3% of all migraine cases (65). Abdominal migraine only occurred in 1.8% of the cases. A German study using a claims dataset reported the following: (1) several episodic syndromes, including abdominal migraine and cyclic vomiting, were risk factors for later migraine; (2) the risk of migraine at primary school age in most exposed children is low, except for abdominal migraine (> 50%); and (3) the proportion of later migraine in children aged 6 to 10 years explained by pre-school episodic syndrome is lower than 3% (03).

Prevention

There are no studies on the prevention of abdominal migraine attacks, although it has been suggested that abdominal migraine might be prevented by the avoidance of triggers, such as travel, prolonged fasting, alteration of sleep patterns, exposure to flickering or glaring lights, exercise, and stress related to school and family activities, and possibly medication overuse (62). Simple dietary management or a few-foods diet is suggested for subjects with frequent attacks (72). However, one study demonstrated the dietary habits and functional gastrointestinal disorders based on Rome IV criteria, including abdominal migraine, in an Italian pediatric cohort aged 4 to 18 years. Furthermore, adherence to the Mediterranean diet was associated with a lower prevalence in younger children (aged 4 to 9 years) but not older children (aged 10 to 18) (22). As for cyclic vomiting syndrome, a systemic review found sumatriptan worked for acute attacks, whereas tricyclic antidepressant were effective as a prophylactic agent (49). Of note, the majority of the studies were non-placebo-controlled studies.

Differential diagnosis

In the absence of pathognomonic features, abdominal migraine is often diagnosed through a process of exclusion. Catto-Smith and Ranuh proposed important differential diagnoses, including the following (21):

Apley, in his careful follow-up study, found that 20 of 100 patients with recurrent abdominal pain in childhood had structural abnormalities (06; 05):

Two reviews added gastroesophageal reflux, Crohn disease, bowel dysmotility, sorbitol and lactose malabsorption (60), and irritable bowel syndrome (44).

Altered awareness or level of consciousness raises the possibility of epilepsy. Millichap and associates showed that cyclic vomiting could be of epileptic origin; Prensky considered half his patients with recurrent abdominal symptoms to have epilepsy; and Hockaday had two patients with structural temporal lobe lesions (57; 67; 41).

Diagnostic workup

Both the Rome III pediatric criteria for abdominal migraine (69) and the ICHD-II criteria for abdominal migraine (35) stress that the disease should be diagnosed by exclusion. The most important evaluation of a child who presents with recurrent abdominal pain is a careful history and physical examination. If there is doubt about abdominal symptoms, a gastroenterologist should be consulted for a second opinion before the diagnosis of abdominal migraine is made. It was reported that a trial of empiric therapy with medications such as pizotifen, propranolol, or cyproheptadine with upper gastrointestinal and small-bowel radiological studies is cost-effective (64). Total gastric empty time evaluated by real-time ultrasonography or scintigraphic methods could also be considered an ancillary diagnostic evaluation. Migrainous children with functional gastrointestinal disorders had significantly more prolonged total gastric empty time than subjects without (11). On the other hand, it was found that adult patients with cyclic vomiting syndrome had rapid gastric emptying and tachygastric electrogastrogram (61). Why these two closely related syndromes had such different manometric profiles should be investigated to see if the difference could be used as a useful clinical marker. In addition, ictal electroencephalogram might provide ancillary diagnostic value in certain patients (63). The North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) provided an official recommendation for diagnosing and treating cyclic vomiting syndrome in children and adolescents in 2008 (51). The consensus recommended targeted testing instead of a "shotgun" approach. This targeted testing focused on four red flags: abdominal signs (eg, bilious vomiting, tenderness), triggering events (fasting, high-protein meal), abnormal neurologic examination (altered mental status, papilledema), and progressive worsening or a changing pattern of vomiting episodes. A subsequent retrospective study focused on white matter hyperintensities in children with cyclic vomiting syndrome because this finding is commonly seen in patients with migraine (68). The authors found that three of the 13 patients in the cyclic vomiting syndrome with migraine group had migraine-like white matter hyperintensities, compared to zero of the five patients in the cyclic vomiting syndrome without migraine group.

Management

Because of the great debate about the existence of the disease entity and no widely accepted diagnostic criteria for abdominal migraine before, evidence-based studies of this illness are rare. The publication of the diagnostic criteria for abdominal migraine proposed by the ICHD-II and Rome III pediatric committee provided a chance to improve its diagnosis and treatment (35). In fact, the difference between “recurrent abdominal pain” and abdominal migraine needs strict validation.

Before making the diagnosis of abdominal migraine, a pediatric gastroenterologist should be consulted, especially when no migraine headache can be traced by a careful medical history. At present, the ICHD-II criteria or the Rome III pediatric criteria can be applied for diagnosis even though the criteria need further evaluation (35; 69). When the diagnosis is uncertain, a study for upper gastrointestinal and small intestine problems is warranted.

Management of abdominal migraine is indicated if it causes significant morbidity depending on the severity and frequency of the attacks. Russell and colleagues suggested the following steps for the management of abdominal migraine (72):

Abortive treatment for abdominal migraine has not been well studied. Sumatriptan was reported to be effective in treating abdominal migraine in adult and pediatric patients (58; 62) and a 19-year-old man with cyclic vomiting (08). Kakisaka and colleagues reported that intranasal sumatriptan was found to be effective in treating two pediatric patients (ages 9 and 12) with abdominal migraine (46). A Japanese study showed that subcutaneous sumatriptan effectively treated 19 of 35 attacks in 11 patients with cyclic vomiting syndrome; nasal sumatriptan was effective in two of six attacks in five patients (39). They also noted that a positive family history of migraine was predictive of response. However, clinicians should carefully weigh the risks and benefits of using triptans because their efficacy and safety in children have not been well documented. A trial of antimigraine medications is indicated if abdominal migraine causes a significant disability. These include pizotifen, propranolol, and cyproheptadine. In a small-scale, double-blind, placebo-controlled crossover trial, pizotifen, a serotonin receptor antagonist that is not available in the United States, provided effective treatment for abdominal migraine (77). Propranolol and cyproheptadine were also reported effective for both short- and long-term use (6 months to 3 years) in a retrospective clinic study (86) in which patients reported significant improvement after treatment. Propranolol was also found to be effective as prophylactic treatment in 92% of a group of children with cyclic vomiting syndrome, showing better results than amitriptyline, effective in 56% (34). The latter was found to be effective in suppressing attacks in adult patients with cyclic vomiting syndrome (61). Flunarizine, which is still unavailable in the United States, has been employed as a prophylactic drug in a study. A 61% reduction in frequency and a 51% reduction in duration of abdominal migraine attacks were demonstrated (47). However, this study was retrospective, and the study cohort was small (n=10). A larger-scale study demonstrated that flunarizine was effective in decreasing the frequency and duration of migrainous episodes as well as the gastrointestinal symptoms in children with migraine and functional gastrointestinal disorder, with a quantifiable decrease in total gastric emptying time as measured by ultrasonography (11). Nevertheless, a randomized, controlled trial is still needed to affirm and solidify this therapy. In addition, intravenous valproic acid was found to be effective in aborting abdominal migraine (79). Oral valproate in a dose of 10 to 40 mg/kg per day was also found to be an effective prophylactic drug for cyclic vomiting syndrome in a small uncontrolled trial (40). Antiepileptic drugs such as topiramate in a single case report (63) and zonisamide or levetiracetam in an uncontrolled case series (24) were found to be potentially beneficial in patients with cyclic vomiting syndrome; however, these findings await further confirmations. An Internet survey showed that both coenzyme Q10 and amitriptyline were effective in patients with cyclic vomiting syndrome (72% vs. 68%); however, the side effects differed significantly (0% vs. 50%) (15). In a retrospective chart review of 42 patients with cyclic vomiting syndrome, Boles reported that a protocol consisting of mitochondrial-targeted cofactors (co-enzyme Q10 and L-carnitine) plus amitriptyline may be effective in the prevention of cyclic vomiting episodes (12). In 2008, the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition had made their therapeutic recommendations for cyclic vomiting syndrome after carefully reappraising previous works. However, all the pharmacological recommendations were off-label as no large-scale controlled or open-label trial had been conducted (51). In a review on adult patients with abdominal migraine, topiramate was suggested as a preventive treatment (85). In a review, Li suggests cyproheptadine for children younger than 5 years of age and amitriptyline for those 5 years of age or older, with propranolol as the second-line agent (54). Management should start with lifestyle alterations and extend to medications, supplements, and stress reduction therapies. Drug therapy can be divided into four phases of the illness: (1) interictal (preventative medications and mitochondrial supplements); (2) prodromal (abortive agents); (3) vomiting (fluids/energy substrates, antiemetics, analgesics, and sedatives); and (4) recovery (supportive care and nutrition). Due to the increasing recognition of cyclic vomiting syndrome in adults, the American Neurogastroenterology and Motility Society and the Cyclic Vomiting Syndrome Association made recommendations of treatment for adults of cyclic vomiting syndrome, including prophylactics with tricyclic antidepressants as first line followed by topiramate or aprepitant and abortive treatments with serotonin antagonists, such as ondansetron; triptans, such as sumatriptan; or aprepitant (neurokinin-1 receptor antagonist) (81). The guidelines were generally similar for the pediatric patients (31).