Abdominal Migraine

Introduction

Abdominal migraine (AM) is an under-diagnosed cause of recurrent abdominal pain in the pediatric population.

The term AM was used by Brams, close to a century ago, to describe periodic attacks of epigastralgia with symptom-free intervals between the episodes and a history of migraine in either the patient or his relative.¹ Since then, more precise criteria have been proposed to aid in the diagnosis of AM.

In 2003, AM was identified as one of the “childhood periodic syndromes that are commonly precursors of migraine” in the International Classification of Headache Disorders (ICHD-2) published by the International Headache Society (IHS); this was further corroborated by ICHD-3 laid out in 2013. In 2006, Rome III Diagnostic Criteria for Functional Gastrointestinal Disorders separately proposed identifying markers for AM which succeeded the 1999 Rome II Gastroenterology Criteria.

In a series of patients from a pediatric gastrointestinal clinic, at least 4-15% of children with chronic, recurrent, idiopathic abdominal pain met diagnostic criteria for AM.² A UK study demonstrated a slightly lower prevalence of 2.4%.³

Interestingly, the application of Rome III criteria for AM classified a greater percentage of children as meeting the diagnostic criteria for AM when compared to the use of Rome II criteria (23.1% vs 5.7%).⁴

The exact pathophysiology of AM is not clearly understood. There is an overlap between AM and migraine as there is between the various functional gastrointestinal disorders such as cyclic vomiting and irritable bowel syndrome.

Presentation

ICHD-3 identifies AM as one of the “episodic syndromes that may be associated with migraine”.⁵

ICHD-3 describes AM as “an idiopathic disorder mainly seen in children as recurrent attacks of moderate to severe midline abdominal pain, associated with vasomotor symptoms, nausea, and vomiting, lasting 2-72 hours and with normality between episodes. Headache does not occur during these episodes”.

ICHD-3 also proposes the following diagnostic criteria for AM:

• A. At least five attacks of abdominal pain, fulfilling criteria B-D
• B. Pain has at least two of the following three characteristics:
  
  1. midline location, periumbilical or poorly localized
  2. dull or ‘just sore’ quality
  3. moderate or severe intensity

• C. During attacks, at least two of the following:
  
  1. anorexia
  2. nausea
  3. vomiting
  4. pallor
• D. Attacks last 2-72 hours when untreated or unsuccessfully treated
• E. Complete freedom from symptoms between attacks
• F. Not attributed to another disorder. In particular, history and physical examination do not show signs of gastrointestinal or renal disease, or such disease has been ruled out by appropriate investigations.

As per Rome III Diagnostic Criteria for Functional Gastrointestinal Disorders, AM must include all of the following:

1. Paroxysmal episodes of intense, acute periumbilical pain that lasts for 1 hour or more
2. Intervening periods of usual health lasting weeks to months
3. The pain interferes with normal activities
4. The pain is associated with 2 of the following:
   a. Anorexia

   b. Nausea

   c. Vomiting

   d. Headache

   e. Photophobia

   f. Pallor

5. No evidence of an inflammatory, anatomic, metabolic, or neoplastic process considered that explains the subject’s symptoms.

In children, it may be difficult to differentiate anorexia from nausea. Pallor may be accompanied by dark shadows under the eyes. Flushing may be the predominant vasomotor symptom in a subset of patients. Abdominal pain can be variable in character and location; ‘colicky’ and diffuse pain is reported in 22% and 16% of patients respectively.⁶ Behavior, mood changes, or nonspecific prodromal symptoms can precede the pain in about 14% of patients.⁶

The typical age of presentation ranges from 3-10 years (peak prevalence at 5-7 years of age) with males and females mostly affected equally; albeit a female preponderance has been reported by many authors.³

Strong family history is noted. 90% of patients have a history of migraines in a first-degree relative, especially their mothers.³ Also, a history of motion sickness is commonly found amongst either the patients with AM or their family members.³

Interestingly, a study by Heuschkel R proposed that AM may be more common than anatomic causes of abdominal pain in children with neurofibromatosis-1.⁷

Children with AM are believed to develop migraine headaches either concurrently or later in life, but abdominal migraine itself rarely persists into adulthood.³ Hence, essentially, abdominal migraine is a disorder of childhood, and therefore it is important for pediatricians to be aware of it.

Investigations

AM, just like migraine, remains a clinical diagnosis. The importance of eliciting a good clinical history and performing a thorough physical examination cannot be undermined. General physical examination including a neurological examination during the symptom-free period is usually entirely unremarkable.

However, it is extremely important that one does not miss a potentially grave underlying gastrointestinal (pancreatitis, intermittent small bowel obstruction, chronic idiopathic intestinal pseudo-obstruction), renal (obstructive uropathy), neurologic (brain tumor with raised intracranial pressure, brainstem tumor, subdural hematoma, familial dysautonomia), endocrine (adrenal insufficiency) or metabolic disease (ornithine transcarbamylase deficiency, methylmalonic acidemia, acute intermittent porphyria) which can mimic symptoms of AM.⁸

Laboratory, radiological and procedural investigations, if any, should be performed to rule out a serious clinical condition at the earliest. Identifying certain “red-flags” which indicate the presence of an organic disease can be crucial in directing relevant investigations to avoid diagnostic delay. These “red-flags” are summarized in Table 1.^9,10 Table 2 provides a list of investigations that should be considered in a child presenting with clinical features suggestive of AM.¹¹ The list is exhaustive but by no means complete and each case merits its own set of investigations if any.

Interestingly, electroencephalography findings have been reported in patients with AM. Benign focal or rolandic spikes have been reported and this may mislead the clinician into believing that the clinical presentation is consistent with an epileptiform phenomenon.¹²

The diagnosis of AM is often delayed as most children get referred to a gastroenterologist before they are referred to a pediatric neurologist.

Non-pharmacological Therapy

A correct explanation of the disease process and also the fact that AM rarely persists into adulthood should be given to the families. It is important to emphasize that a dangerous gastrointestinal, renal, metabolic, endocrine, or neurologic disease process is absent in these children.

Theoretically, avoidance of triggers may help in a small group of patients; although, the evidence is minimal. Russell et al suggested avoidance of triggers viz., stress, travel, prolonged fasting states, exposure to flickering or strobe lights, physical exertion or exercise, and, alterations in sleep patterns.¹³

Dietary modifications with increased fiber, reduced lactose, and avoidance of certain food items have been proposed by various authors; again evidence is scarce to support these dietary changes. Feldman et al reported a significant benefit when fiber was increased in the diet of 52 children with AM studied by them. They concluded that 50% of the children in the fiber group had a 50% reduction in the number of episodes with abdominal pain, compared with only 27% in the control group.¹⁴ Barr and Liebman demonstrated that eliminating lactose in the diet of children with AM, a significant proportion of whom had an abnormal lactose tolerance test, provided a complete symptom resolution.^15,16 Russell et al also proposed that avoidance of certain food items (‘few-foods or ‘oligo-antigenic’ diet) known to trigger AM is helpful especially in children with frequent attacks (>2 per week).¹³ Most notable items include chocolates, citrus fruits, caffeine, cheese, and colorings.

Other non-pharmacologic options include psychotherapy in particular cognitive behavioral therapy. These therapies involve introducing the child to the concept of pain and the various coping strategies against it. Sanders et al studied forty-three children with recurrent abdominal pain between the ages of 7-14 years and randomized them to either the cognitive behavior therapy group or the standard pediatric care group without a cognitive therapy group. Clearly, the percentage of pain-free children as assessed by a diary and parents was higher in the cognitive behavior therapy group both at a 6 month and at a 12-month evaluation.¹⁷ Unfortunately, restricted availability of trained personnel providing cognitive behavior therapy can be a limiting-step for this treatment option.

Pharmacological Therapy

This is usually considered in children who fail non-pharmacological measures or in those with frequent and disabling symptoms where drug therapy can be combined with the non-pharmacological measures discussed above.

Abortive agents used for AM are similar to those used in migraine headaches and include acetaminophen or ibuprofen which is usually most effective when given early during the attack. Symptomatic treatment with anti-emetics is indicated in those with nausea or vomiting.

The prophylactic agents found useful in AM are those primarily used in migraine headaches in children - beta-blockers (propranolol), serotonin antagonists (cyproheptadine, pizotifen), and calcium channel blockers (flunarizine).

A study by Worawattanakul et al concluded that propranolol or cyproheptadine are effective prophylactic agents in the majority of the patients with AM.¹⁸ The authors retrospectively reviewed records of 53 patients who underwent treatment for AM for six months or until migraine cycles had stopped. Among the children given propranolol (10-20 mg, 2-3 times daily), 75% had an excellent response identified as complete cessation of symptoms, 8% had a fair response which meant persistence of symptoms but milder and less frequent and 17% had no response. In those treated with cyproheptadine (0.25-0.5 mg/kg/day), 33% had an excellent response while 50% had a fair response while the non-responders remained the same at 17%. 46% of patients in the propranolol group took the medication for less than six months and 50% in the cyproheptadine group took medication for less than 10 months while the rest of the patients in both the groups continued the respective therapy for three years.

Propranolol is typically started at 10 mg divided into two or three daily doses. Propranolol is best avoided in children with asthma as it can cause bronchospasm. Cyproheptadine is used in doses of 2-8 mg divided into two doses; increased appetite, weight gain, and sedation are the most common side-effects.

Symon and Russell in a double-blind crossover trial of fourteen children with AM using pizotifen or placebo demonstrated pizotifen to be clearly superior to placebo. Pizotifen which has anti serotonin, antihistamine, and anticholinergic activity; and, is widely used in Canada and Europe; was found to be an effective prophylactic drug in 70% of the children with AM.¹⁹ Some authors have reported continued benefits even after the drug is stopped.²⁰ Pizotifen is started at 0.5 mg/kg/day and titrated based on the clinical response up to 4-6 mg/day. Common side-effects are similar to cyproheptadine.

Flunarizine, a calcium-channel blocker was found to be a safe, once-daily option for AM prophylaxis in a study by Kothare.²¹ Ten children with AM were treated with a mean dose of 7.5 mg/day and were followed up after a mean average of 13 (6-24) months. 61% reported a reduction in frequency and 51% a reduction in the duration of the attacks. Flunarizine, not available in the US, is typically given at a dose of 5 mg at bedtime. Common side effects include constipation, weight gain, and hypotension.

Small case series report response to valproic acid and triptans. Others report scarce use of topiramate, amitriptyline, verapamil, nadolol, lynesternol (progestin not available in the US), flumedroxone (progesterone derivative not available in the US). The response is believed to be due to similarities in the pathophysiology of migraine and AM.^22,23

Treatment

Very little is concluded regarding the treatment of children with AM. Not all children require treatment; especially if the frequency and severity of the episodes do not affect the quality of life or schoolwork. In these cases, in addition, to providing reassurance; the risk-benefit ratio of different treatment options should be explained to the family and they should be totally involved in the decision-making process.

Broadly, the management of AM is divided into non-pharmacological and pharmacological modalities.

Complications

Complications from mistaking an underlying serious condition for AM and vice-versa can be devastating. An incorrect diagnosis involves subjecting the child to unnecessary and often expensive investigations. A delay in diagnosis can cause significant impairment in the quality of life of both children and the parents, with school and work absenteeism respectively. Hence, early and correct diagnosis is the key.

As mentioned before, children with AM are known to develop migraine headaches in later life. In a study of 54 pediatric patients with AM, 52% developed migraine headaches at a 7-10 year follow-up period, compared to 20% of age-matched controls.⁶ These patients can develop the usual complications of a migraine headache which include status migrainosus, persistent aura without infarction or a migraine-aura triggered a seizure in those with migraine with aura, and, migrainous infarction.⁵