Comorbidity of Psychiatric and Behavioral Disorders in Pediatric Migraine


  • Ann Pakalnis MD,

  • Jessica Gibson MS,

  • Andrew Colvin PhD

  • From the Departments of Pediatrics and Neurology, The Ohio State University, College of Medicine, Columbus, OH (Dr. Pakalnis); Department of Psychology, Columbus Children's Hospital, Columbus, OH (Ms. Gibson and Dr. Colvin); and Department of Psychology, Ohio University, Athens, OH (Ms. Gibson).

  • This study was supported in part by a research grant from the National Headache Foundation.

Address all correspondence to Dr. Ann Pakalnis, Section of Neurology, Children's Hospital, 700 Children's Drive, Columbus, OH 43205.


Objective.—To determine whether behavioral and psychiatric disorders occur more frequently in school-age children with migraine headache. To also elucidate treatment response related to comorbid psychiatric or behavioral diagnosis.

Background.—Recurrent migraine headaches are common in school-age children. Concurrent behavioral or psychiatric diagnoses could significantly impact headache frequency, severity, and response to treatment.

Methods.—Healthy children from 6 to 17 years of age presenting to our headache clinic with migraine headache according to International Headache Society (IHS) criteria were identified. Parents/guardians were asked to complete the Child Symptom Inventory, 4th edition (CSI-4) after written informed consent. Children with positive rating scales underwent psychological interviews for confirmatory diagnosis. Results were compared to controls. Headache patients were assigned our usual treatment paradigm. Response regarding headache frequency was assessed at 3 months.

Results.—A total of 47 patients were diagnosed with migraine headaches. The mean age was 10.55 years. Thirty controls were identified. After completing the CSI-4 and confirmatory psychological interview, 14 of 47 headache patients fulfilled Diagnostic and Statistical Manual (DSM-4) criteria for a psychiatric or behavioral disorder. Oppositional defiant disorder (ODD) was significantly represented among children with migraine compared to the control group of children. Headache patients improved significantly post-treatment regarding their headache frequencies regardless of comorbid psychiatric or behavioral disorder. No significant differences were noted between boys and girls regarding diagnoses or treatment outcome.

Conclusion.—ODD was a significant comorbidity in our headache population. Although families complained of significant behavioral symptomatology in their children, most of these symptoms did not qualify their children for a psychiatric diagnosis and may be related to the stressors of headache on social/school disruption.

Migraine headaches are common in children and adolescents with prevalence rates ranging from 3.2% to 10.6%, depending on rigidity of criteria used for diagnosis of migraine.1 Recurrent headaches in children are a great source of stress on family dynamics with their impact on school performance, as well as loss of job productivity and work attendance in parent or guardian. Many new treatment options are being utilized including therapy with antiepileptic drugs such as valproate and topiramate and also abortive therapy with triptans that have commonly been used in the adult population.2–4 Alternative behavioral therapies such as biofeedback and relaxation therapy have also been noted to have utility in pediatric migraine and are being prescribed more frequently in children of appropriate age.5

In the adult migraine population, psychiatric comorbidity has been well defined.6 Co-occurrence of multiple psychiatric disorders with migraine has been repetitively documented including affective and anxiety disorders such as depression, panic disorder, bipolar disorder, and obsessive-compulsive disorder.7 When coexistent with migraine, these psychiatric conditions notably influence treatment selection in response to headache therapy. This is especially relevant since many of the migraine prophylactic agents such as amitriptyline and divalproex, display significant psychotropic properties.8,9 In addition, psychologic modalities may be an important consideration in the therapeutic armamentarium in these select patients with comorbid psychiatric disorders.

Although many families of children with migraine note concomitant behavioral issues, the presence of comorbid psychiatric or behavioral disorders in their children has not been well defined. Prior studies have not utilized validated psychiatric scales or strict International Headache Society (IHS) headache criteria in evaluation.10,11 The presence of concurrent psychiatric or disruptive behavioral problems could negatively impact factors that would aggravate headaches such as increased stress, diet changes, sleep deprivation, and compliance with medical therapies. Family and peer relationships, and school performance suffer with maladaptive behavioral issues. Children with behavioral disorders and their treatment remedies (such as neurostimulants) may decrease appetite and alter sleep schedules. The primary objective of our study was to evaluate children with recurrent headache to ascertain for presence of comorbid psychiatric or behavioral disorder diagnosis and secondarily to ascertain if they respond less favorably to therapeutic intervention to their headaches than their cohorts without associated psychiatric issues.


Over a 9-month period, school-age children (ages 6 to 17 years) with a diagnosis of episodic migraine (with or without aura) according to IHS criteria were prospectively identified from a multidisciplinary pediatric headache clinic at Columbus Children's Hospital.12 The diagnosis was made by one of the authors (AP), a board certified neurologist. Children with chronic daily headache (CDH) or medication overuse/analgesic rebound headaches were purposefully excluded because of the difficulties inherent in reproducible criteria for diagnosis in children compared to adults. Prior studies have also delineated higher degree of maladaptive coping skills and psychologic dysfunction in these children with CDH.13 The study protocol and consent/assent forms were approved by the Institutional Review Board, Columbus Children's Hospital. Consent was obtained from the parent/guardian of their child participant.

Complete histories and neurologic examinations were performed as part of their headache clinic evaluations. Imaging studies, such as magnetic resonance imaging, were also performed when appropriate. Children enrolled had no history of serious medical problems, epilepsy, or progressive central nervous system disorders. Four potential families recruited from our clinic declined, citing participation time constraints or disinterest.

The parents of children presenting to our clinic who fulfilled these study criteria were asked to complete either the Child Symptom Inventory, 4th edition (CSI-4) or the Adolescent Symptom Inventory, 4th edition (ASI-4), which are well-accepted validated diagnostic studies with utility in evaluating children with behavioral or psychological parameters consistent with one or more DSM-4 diagnoses. Children with high predictive scores for behavioral or psychological disorder, as defined by the CSI scoring scale, were interviewed using a semistructured interview (Barkley & Murphy 1998) to confirm a diagnosis under the supervision of a licensed psychologist.14–17 Seventeen of 47 (37%) headache patients were interviewed after meeting CSI/ASI criteria. Other demographic data obtained included age, gender, duration of headaches, and response to different treatment regimens, which were prescribed, for their headache disorder.

Families presenting to the headache clinic were instructed to keep a headache diary of their child's headaches prospectively before their appointments and on a continuing basis to assess their response to medication. Based on frequency of headaches, patients were stratified to treatment groups; utilizing abortive therapy alone or in conjunction with prophylactic medication. Those with at least three or more migraine headaches per month over the past 3 months were referred for preventive therapy including amitriptyline, divalproex, topiramate, levetiracetam, and cyproheptadine. Triptan agents, naproxen sodium or ibuprofen were used for abortive therapy of migraine headaches. At the time of their initial clinic visit all children aged 8 years and above were also referred for at least one session of biofeedback, relaxation therapy, and stress counseling performed by a clinical psychologist; with an average being two sessions. All patients/families received general education regarding lifestyle issues and headache exacerbation (such as dehydration, diet, and sleep deprivation).

A control group of 33 healthy school-age children without a history of headaches were randomly recruited from local pediatrician offices in the community during routine physicals. Results of statistical analyses showed no significant differences between the control and migraine groups regarding demographic variables (ie, age, sex, race, socioeconomic situation). Three declined to participate due to lack of interest. Parents/guardians of the remaining 30 children were asked to complete the CSI-4 or ASI-4 after written informed consent. Seven of 30 control patients were interviewed after scoring the CSI/ASI for a behavioral/psychological diagnosis.

Statistical analysis using χ2 was performed comparing our control group and headache patient group to determine the frequency of psychological and behavioral diagnoses. Severity of reported behavioral or psychiatric symptoms on the CSI/ASI were also analyzed in both groups using independent samples t-tests. A Bonferroni correction was not used, therefore there is an increased risk of type I error. However, there is a decreased risk of type II error.

Parents of the headache patients were contacted by phone after a 3-month interval to review headache diaries to ascertain treatment response. A regression analysis was conducted that used headache frequency as the dependent variable and the presence of behavioral/psychiatric diagnoses as the independent variables in order to evaluate differential treatment responses in migraine children with comorbid behavioral/psychiatric illness versus children without a behavioral/psychiatric diagnosis


Over a 9-month time interval, 47 patients were diagnosed with migraine headaches according to IHS criteria. Eleven of these had migraine with aura. Twenty-one patients were girls. The mean age was 10.55 years. Thirty controls were identified from local pediatrician practices and 17 of these patients were girls. The mean age was 10.93 years. Six of 47 (13%) headache patients had a prior psychiatric diagnosis; 4 of 30 (13%) control children had similar history. Eight of 47 patients had more severe headaches and were treated with a preventive medication in combination with abortive therapy. Thirty of 47 migraine patients were treated with triptans, while three children were treated with over-the-counter analgesic medications.

Behavioral or psychiatric diagnoses including attention deficit hyperactive disorder (ADHD—inattentive, ADHD—hyperactive/combined), conduct disorder (CD), oppositional defiant disorder (ODD), depression, generalized anxiety disorder, and dysthymia were evaluated in the children with migraines. Based on findings from the CSI-4 and semistructured interview, there were significant differences between the migraine group and control group regarding the frequency of ODD (Table 1). Specifically, results of these analyses found that children with migraines have a significantly higher number of ODD diagnoses than control children, χ2(1) = 6.51, P < .05. There were no other significant differences between the two groups regarding psychiatric diagnoses.

Table 1.—.  Behavioral/Psychiatric Diagnosis in Children With Migraine
(n = 30)
Migraine Group
(n = 47)
  1. *Significant (χ2 analysis).

  2. Eight of the headache patients had more than one diagnosis, 3 of the control patients had more than one diagnosis.

ADHD combined13
Conduct disorder02
ODD09 (P= 0.01*)

This study also examined differences between children with migraines and healthy controls regarding the presences of psychological symptoms, as reported by parents on the CSI/ASI-4 (Table 2). Although T-scores of patients with headache were all within normal range, they were significantly higher than the control group T-scores on all of the following scales: ODD (t(62) =−5.08, P < .05), generalized anxiety disorder (t(62) =−4.87, P < .05), and CD (t(62) =−3.16, P < .05).

Table 2.—.  Migraine Patients Have More Severe Psychiatric Symptoms
 T-Scores—Severity of CSI Symptoms
(50th Percentile—Normative Mean)
Range of Scores 50 to 78
n = 30
Migraine Children
n = 47
  1. *Significant (paired t-tests with Bonferroni correction).

ADHD/inattentive52.76 ± 5.4455.65 ± 7.95 
ADHD/hyperactive52.21 ± 5.3855.24 ± 9.29 
ADHD combined53.24 ± 6.0155.48 ± 7.96 
Conduct disorder50.45 ± 1.3753.62 ± 6.83P= 0.04*
ODD52.73 ± 4.1257.81 ± 9.95P= 0.03*
Anxiety54.36 ± 6.6459.24 ± 9.44P= 0.04*
Depression 55.36 ± 55.3655.24 ± 8.32 
Dysthymia 59.09 ± 10.0860.38 ± 9.87 

At 3-month follow-up, response to the prescribed headache treatment was evaluated in the headache children. Results of a repeated measures t-test found that the headache patients reported a statistically significant decrease in headache frequency post-treatment (t(37) = 2.94, P < .01). However, there was no statistically significant difference regarding response to therapy in migraine children who qualified for different behavioral or psychiatric diagnoses (eg, ODD vs ADHD). There was also no relationship between response to therapy in headache children with a comorbid diagnosis and response to therapy in headache children without a comorbid diagnosis.


Psychiatric comorbidity has been well described in adults with migraine headache. Association with anxiety, obsessive-compulsive disorder, and in particular mood disorders such as depression have been noted in adult migraineurs. Depression rates have been documented between 3.8% and 54.3% compared with general population lifetime rates, which range from 4.4% to 12.6%.6,18–21 A correlation between migraine and adult patients with bipolar disorder has also been noted.22 Anxiety and panic disorders are more frequently diagnosed in adults with headache in population-based studies. In a study of adolescents and young adults, age adjusted risks for migraine were elevated for both men and women with panic disorder.7

In striking contrast, psychiatric disorders coexisting with migraine in children and adolescents has not been well studied. Guidetti, et al observed a comorbidity with anxiety disorder and depression in adolescent and young adult migraine patients (mean age—17.9 years). These factors negatively influenced headache progression.10 Egger et al evaluated children with psychiatric disorders for the presence of headaches.11 While the criteria used for evaluation of the semiology of headaches were not clear and IHS guidelines were not utilized, they found in their study that girls with depression and anxiety had greater prevalence of headaches but this relationship was not present in boys. CD in boys had a significant association with headache. These studies did not include a control group of healthy children.

Guidetti and Galli evaluated 54 children with mixed headache types including migraine (36), episodic tension (8), chronic headache (6), and unclassified (4), for the presence of psychiatric comorbidity.23 Sleep disorders, anxiety, and mood disorder were frequently represented. IHS and DSM-4 criteria were used for diagnosis. Another study evaluated children with migraine headache utilizing diagnostic interviews for anxiety and depression.24 Male first-born children were found to have higher predisposition for migraine and associated anxiety and depression. Control groups of healthy school-age children were not used in these studies and retrospective analysis was applied.

Relationship of the disruptive behavioral disorders (ADHD, ODD, and CD) and comorbidity with headache in children and adolescents has not been studied previously. Behavioral issues such as these are common in school-age children and many of the parents/guardians of our headache patients complain of these issues in concert with their children's headaches. When present in children with an underlying headache disorder, they could aggravate migraine frequency and severity with their disruption of the academic or family milieu. As shown by the significant increase in T-scores, many behavioral/psychologic issues were frequently reported by families but only ODD was significantly represented compared to our control cohort when strict diagnostic criteria were applied. These psychologic complaints appear to be generalized and multifactorial, with only oppositional-defiant issues attaining clinical significance to make a psychiatric diagnosis.

Our study raises important differences with adult comorbidity studies. Mood or anxiety disorders were not found in a significantly higher frequency in our children with headache. Our study excluded children with CDH and this group of patients with very frequent headaches or medication overuse may be more prone to psychiatric or behavioral disorders. These may be more commonly represented in adult migraineurs due to the chronicity of their underlying headache disorder. This observation emphasizes the relative importance of aggressive early treatment of headache disorders in children. Although ODD was significantly more frequently represented in our headache population than in controls, the incidence of 19% in our headache cohort compares similarly with prevalence in US school-age children at 16%. ADHD in our headache population was more frequent; 17% in our headache group versus 3% to 5% of national norms. The more severe behavioral diagnosis of CD was underrepresented in our headache children (4.2%) compared to 6% to 16% of boys and 2% to 9% of girls in the general population as a whole.25 This may be related to our small patient sample and a selection bias for our families that may have greater resources and support systems compared to population as a whole where CD may be over represented in children in residential facilities or other nonfamily units.

Regardless of the presence of comorbid behavioral or psychiatric diagnosis, our headache patients in their pilot study, tended to respond similarly to treatment for their headaches. A previous study by our headache center had shown a poorer response in headache outcome in these children with concurrent psychiatric diagnosis when treated prophylactically with divalproex sodium for their headaches.26 However, this group of children had heterogeneous psychiatric diagnosis and more frequent headaches in general than our current study patients in whom only a minority received prophylactic medication for their migraine headaches. In addition, our present therapeutic paradigm in headache clinic significantly utilizes behavioral/psychologic therapy as a treatment option for families, which was not as generally widely accepted previously. With our current treatment plan utilizing psychologic modalities, this may improve outcome in children with psychiatric comorbidities. Future studies could further elucidate more specific information regarding psychological treatment modalities and headache response in a prospective care-centered design.

The relationship between ODD and headache, particularly migraine, raises several possibilities as far as semiologies for this co-occurrence in our patient cohort. Biochemical differences may play a role with central serotonergic function. Snoek et al studied children with ODD with responses to sumatriptan injection (5HT1B/ID agonist) to ascertain serotonergic responsiveness.27 Growth hormone response was significantly higher in children with ODD compared to control healthy children suggesting the postsynaptic 5HT1B/1D receptor in these children is more sensitive. This may play a role in the higher level of aggressiveness in children with ODD and the central serotonergic pathway has been well documented to be involved in migraine pathophysiology. Decreased central serotonergic function may play a role in migraine induction.28 Children with comorbid migraine and ODD may have innate differences in serotonin susceptibility.

Frequent headaches and stress secondarily engendered may also play a role in inducing many symptoms referable to ODD such as irritability, emotional lability, missing school and social functions, and argumentativeness.29 These children with migraine may react in maladaptive patterns to social/academic stressors related to their headache symptoms' effects on peer relationships and school absences. Children and adolescents with frequent headaches exhibit more disruptive behavioral symptoms in an oppositional pattern and family members/parents could consequently report more negative symptomatology in their children. With response to headache treatment, some of these disruptive behaviors may abate.

Further studies evaluating this relationship between behavioral disorders and ODD in particular in children with migraine is warranted. Coexisting psychologic disorders and headaches can create a complex management problem. Evaluating a larger patient sample with broader spectrum psychological rating scales would be helpful. Bidirectional studies evaluating children with pre-existing behavioral diagnosis such as ODD and then historically or possibly prospectively evaluating for concurrent headache disorder utilizing defined IHS criteria is indicated.

Our group of headache children tended to respond no differently to appropriate therapy for their headaches regardless of concurrent behavioral or psychiatric diagnosis. This in part may be attributed to our paradigm of aggressive utilization of psychologic intervention including stress counseling, biofeedback, and relaxation therapy in headache patients of appropriate age and level of cooperativity. In the future, a treatment study evaluating the benefits of psychologic intervention in children with comorbid psychiatric issues may be appropriate. Offering therapy of this type to pediatric headache patients may be an important part of the therapeutic treatment plan to achieve optimal success in decreasing headache burden in children and their families.


Acknowledgments: We appreciate the assistance of Ms. Jessica Olsheski who typed the manuscript. Ms. Gibson was a graduate student training under Kenneth Holroyd, PhD.