Abdominal Migraine: Prophylactic Treatment and Follow-up

Abstract

Background: 

Abdominal migraine is a syndrome characterized by recurrent stereotypic episodes of paroxysmal abdominal pain and nausea and/or vomiting with wellness between episodes. It is often associated with a positive family history of migraine and no other apparent underlying disease. The purpose of this study was to report in patients diagnosed with abdominal migraine the outcome, the effect of prophylactic treatment, and the duration of treatment.

Methods: 

The records of 53 patients who underwent treatment after a diagnosis of abdominal migraine were retrospectively reviewed. Responses to treatment were graded as excellent (cessation of recurrent abdominal pain), fair (persistence of symptoms but milder and less frequent), or poor (no response). Follow-up data were available in 38 patients. Twenty-four patients were treated with propranolol and 12 with cyproheptadine. Four were not treated because of mild and infrequent symptoms.

Results: 

Among the children treated with propranolol, 18 (75%) had an excellent response, 2 (8%) had a fair response, and 4 (17%) had no response. In those treated with cyprohepatadine, 4 (33%) had an excellent response, 6 (50%) had a fair response, and 2 (17%) had no response. Patients were instructed to continue medication for 6 months or until cycles had stopped. However, 11 of 24 patients (46%) in the propranolol group took medication for less than 6 months and the remaining patients from 6 months 3 years. Six patients in the cyproheptadine group (50%) took medication less than 10 months and the remaining patients for 10 months to 3 years.

Conclusion: 

Patients with abdominal migraine may benefit from prophylactic treatment with propranolol or cyproheptadine.

Recurrent abdominal pain is a common problem in children and comprises a large portion of patients referred to the pediatric gastroenterologist. The prevalence of recurrent abdominal pain in children is approximately 10% ^(1,2). A subgroup of these patients has symptoms of abdominal migraine, which has a prevalence in the range of 1% to 4% of the pediatric population ^(3-5). The diagnosis of abdominal migraine is established by history, which includes paroxysmal recurrent abdominal pain with nausea and/or vomiting, a history of wellness between episodes, no identifiable alternative causes, and a family history of migraine among first-degree relatives ^(6-8).

Treatment of abdominal migraine is focused on preventing attacks of prophylactic medications. There are reports of successful treatment of abdominal migraine or the related (perhaps equivalent) condition called cyclic vomiting syndrome, with various medications including cyproheptadine, pizotifen, erythromycin, phenobarbital, propranolol, and imipramine ^(7,9-14). These studies included small numbers of patients with limited characterization of the response to treatment and the duration of treatment. We have seen and evaluated a relatively large group of patients with abdominal migraine in our pediatric gastroenterology clinic. The goal of our study was to determine whether these patients' responded to prophylactic treatment with propranolol or cyproheptadine and to evaluate the duration of treatment required for them to remain symptom free.

MATERIALS AND METHODS

We retrospectively reviewed the course of children younger than 18 years with a diagnosis of abdominal migraine in the pediatric gastroenterology clinic at the University of North Carolina at Chapel Hill between 1990 and 1996. Evaluation included review of medical records and telephone interviews of all patients' parents. In approximately 1200 patients referred to our clinic for chronic abdominal pain during this period, 53 patients had received a diagnosis of abdominal migraine. Criteria for diagnosis of abdominal migraine were a history of stereotypic attacks of paroxysmal recurrent abdominal pain with nausea and/or vomiting, wellness between episodes, and no identifiable alternative causes. A family history of migraine, although an important clue in diagnosis, is not an absolute criterion for making the diagnosis ⁽⁶⁾ and was not required. In 65% of patients, a first- or second-degree relative was identified who had symptoms consisten with migraine headaches or abdominal migraine. An additional feature in some patients was falling asleep during an attack and awakening feeling well. In many, the onset and resolution were sudden.

Relevant follow-up data, obtained by phone or letter, were available in 38 patients who had been treated medically. All patients had a history of paroxysmal abdominal pain and good health between episodes. Other clinical features are summarized in Table 1. Work-up for abdominal pain or vomiting was obtained in 23 patients. All findings were negative and included the following: upper gastrointestinal fluoroscopy and small bowel follow-through (21 patients), abdominal ultrasonography (3 patients), upper gastrointestinal endoscopy (1 patient), colonoscopy (1 patient), and computed tomography or magnetic resonance imaging of the head (5 patients). We were not able to gather reliable data retrospectively about headache, pallor, motion sickness, or photophobia, although in many cases these features were mentioned in the medical record and have been reported previously ^(3,5,6). Patients were treated with 10 to 20 mg propranolol twice or three times daily or 0.25 to 0.50 mg/kg cyproheptadine daily. Propranolol was selected first if there was no contraindication, such as asthma. If patients could not take propranolol or had no response, cyproheptadine was administered as an alternative. Response to treatment was graded as excellent (absence of further episodes of abdominal pain), fair (persistence of abdominal pain but improvement with less frequent and less intense episodes of abdominal pain), or poor (no response).

TABLE 1: Clinical data on children with abdominal migraine

RESULTS

Eighteen of 24 patients (75%) treated with propranolol had an excellent response. Two patients (8%) had a fair response and continued taking medication. Four patients (17%) did not respond. The average number of abdominal pain attacks per patient in the propranolol group was 5.4 ± 0.6 (mean ± standard error) in the 6 months preceding treatment and 1.0 ± 0.3 attacks during the 6 months after treatment. There were no differences in response to treatment, when comparing the children with positive family history with those with negative family history. Four children had mild enough symptoms that they were not treated.

Of the four children who did not respond to propranolol, two responded to cyproheptadine, and in one patient functional abdominal pain was diagnosed later. In the cyproheptadine group (12 patients), four patients (33%) had an excellent response, six patients (50%) had a fair response, and two patients (17%) had no response. In one of the two patients with no response, gastroesophageal reflux was diagnosed later. As a whole, patients in the cyproheptadine-treated group had 5.8 ± 2.1 attacks during the 6 months before treatment and 1.3 ± 0.5 attacks during the 6 months after treatment. The response to treatment of these two groups is shown in Table 2. Although there was a trend toward a better response with propranolol, it did not reach statistical significance by chi square analysis.

TABLE 2: Response to treatment

Almost half of the patients in the propranolol group (n = 11) were given medication for approximately 3 to 6 months and did well after they discontinued the medication (Fig. 1). The remaining 13 patients received propranolol for 6 months to 3 years. Half of the patients in the cyproheptadine group (n = 6) took medication for less than 10 months.

FIG. 1: Percent of patients (n = 18) who had an excellent response (cessation of symptoms) to propanolol in relation to duration of treatment. About 50% of these patients required prophylactic treatment for less than 6 months duration.

Few side effects were reported by parents. In the propranolol group, side effects reported included shortness of breath (one patient), headache (one patient), and drowsiness (one patient). In the cyproheptadine group, side effects reported included drowsiness (one patient) and increased appetite with excessive weight gain (one patient).

DISCUSSION

Abdominal migraine is a syndrome that is not widely recognized by general pediatricians. Repeated episodes of vomiting are well known to be associated with migraine and have been called "migraine equivalent" or "cyclic vomiting syndrome" ^(15-17). Attacks are believed to be precipitated by abnormal electrical discharge from the hypothalamus extending to the cortex and autonomic nervous system ⁽¹⁸⁾ or by changes in blood flow velocity in the cerebral artery ⁽¹⁹⁾. Criteria to establish the diagnosis have been developed over several years ^(6-8). Visual evoked response may show unusual spiking activity in most patients but is not required to make the diagnosis ^(18,19). Arriving at the diagnosis requires ruling out other organic or functional causes ^(8,13,16).

Abdominal migraine prophylaxis is recommended primarily if a patient has symptoms that disrupt his or her normal activity. Symon and Russell ⁽⁹⁾ studied treatment of 14 children with pizotifen, an antimigraine with antiserotonin (5-HT₂) and antihistamine (H-1) effects and reported that it was superior to placebo for prophylaxis. Pfau et al. ⁽¹¹⁾ reported complete cessation of symptoms in six of eight patients with abdominal migraine treated with propranolol, and in three of four patients treated with cyproheptadine. In the recent report by Gokhale et al., ⁽²⁰⁾ of the 14 patients diagnosed with cyclic vomiting syndrome treated with prophylactic phenobarbital, 11 had complete resolution of symptoms, and the remaining 3 had marked improvement ⁽²⁰⁾.

Propranolol has been used in treating migraine headache in children, with good results ^(21,22). The majority of our patients with abdominal migraine were treated with propranolol, and we found that this drug was effective, as did others ^(6,11). The mechanism of action of propranolol for migraine is not well understood; however, it is proposed to be related to β-blockade ^(23,24), possibly inhibiting vasospasm of cerebral arteries in the early stages of the attack.

Cyproheptadine has also been used in the treatment of migraine headache ^(23-27). The mechanism of action of cyproheptadine is probably an antiserotonin effect ^(24-26) or a calcium-channel block ^(23-26,28). Lundberg ⁽⁷⁾ reported nine patients with abdominal migraine treated with cyproheptadine. Three were free of symptoms, three improved, and three had no change in symptoms. Similarly, approximately one third of our patients in the cyproheptadine group had a good response, but others had a fair response. There was no statistically significant difference between treatment with propranolol and cyproheptadine.

The results of our retrospective study of one of the largest series of children with abdominal migraine suggest that most patients with abdominal migraine benefit from prophylactic treatment. Our data highlight the observation that abdominal migraine is not rare in children and is treatable. The diagnosis of abdominal migraine is based on clinical features, and this necessitates a thorough history and physical examination and a good follow-up. Our study was a noncontrolled study and does not rule out a possible placebo effect, although the relatively large number of episodes before treatment suggest a well-established cyclic pattern that we think would not have responded to placebo. The side effects of both drugs in this study were few and not severe. A placebo-controlled trial should be undertaken to confirm our findings.