Comorbidity of Migraine and Psychiatric Disorders—A National Population-Based Study

Authors

  • Nathalie Jette MD, MSc,

    1. From the Department of Clinical Neurosciences (Dr. Jette, Dr. Becker, and Dr. Wiebe) and Community Health Sciences (Dr. Patten and Mrs. Williams), University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author
  • Scott Patten MD, PhD,

    1. From the Department of Clinical Neurosciences (Dr. Jette, Dr. Becker, and Dr. Wiebe) and Community Health Sciences (Dr. Patten and Mrs. Williams), University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author
  • Jeanne Williams MSc,

    1. From the Department of Clinical Neurosciences (Dr. Jette, Dr. Becker, and Dr. Wiebe) and Community Health Sciences (Dr. Patten and Mrs. Williams), University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author
  • Werner Becker MD,

    1. From the Department of Clinical Neurosciences (Dr. Jette, Dr. Becker, and Dr. Wiebe) and Community Health Sciences (Dr. Patten and Mrs. Williams), University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author
  • Samuel Wiebe MD, MSc

    1. From the Department of Clinical Neurosciences (Dr. Jette, Dr. Becker, and Dr. Wiebe) and Community Health Sciences (Dr. Patten and Mrs. Williams), University of Calgary, Calgary, Alberta, Canada.
    Search for more papers by this author

  • Conflict of Interest: None

Dr. Nathalie Jette, Department of Clinical Neurosciences, University of Calgary, 1403-29th St. NW 12th floor, Room C1209, Calgary, Alberta, Canada.

Abstract

Background.— Migraine is common, with an estimated lifetime prevalence of 7-17%. Population-based studies have reported an association between various psychiatric conditions and migraine. This is a population-based study exploring the association between migraine and psychiatric disorders in a large cohort and assessing various health-related outcomes.

Objective.— (1) Determine the prevalence of various psychiatric conditions in association with migraine; (2) describe the patterns of association of these comorbidities with a variety of health-related outcomes.

Methods.— Data from the 2002 Canadian Community Health Survey were used. This is a national health survey which included administration of the World Mental Health Composite International Diagnostic Interview to a sample of 36,984 subjects. Health-related outcomes included 2-week disability, restriction of activities, quality-of-life, and mental health care utilization.

Results.— The prevalence of physician-diagnosed migraine (n = 36,984) was 15.2% for females and 6.1% for males. Migraine was most common in those between ages 25 and 44 years and in those of lower income. Migraine was associated with major depressive disorder, bipolar disorder, panic disorder, and social phobia, all occurring more than twice as often in those with migraines compared with those without. Migraine was not associated with drug, alcohol, or substance dependence. The higher prevalence of psychiatric disorders in migraineurs was not related to sociodemographic variables. Psychiatric disorders were less common in those over 65 years, in those who were in a relationship, and in those of higher income whether migraine was present or not. Health-related outcomes were worst in those with both migraines and a psychiatric disorder and intermediate in those with either condition alone.

Conclusion.— Migraine is associated with major depressive disorder, bipolar disorder, panic disorder, and social phobia. Migraine in association with various mental health disorders results in poorer health-related outcomes compared with migraine or a psychiatric condition alone. Understanding the psychiatric correlates of migraine is important in order to adequately manage this patient population and to guide public health policies regarding health services utilization and health-care costs.

Abbreviations:
CCHS 1.2

Canadian Community Health Survey

 

Mental Health and Wellbeing

MDD

major depressive disorder

HRQoL

health-related quality-of-life scale

SNRI

selective norepinephrine reuptake inhibitor

SSRI

selective serotonin reuptake inhibitor

WHO

World Health Organization

WMH-CIDI

World Mental Health Composite International Diagnostic Interview

INTRODUCTION

Migraine is common, with a lifetime prevalence in Canada estimated to be around 7-17%,1-5 in keeping with findings from non-Canadian research.6,7 Studies consistently show that migraines are up to 4 times more common in women than men,1-4,6-9 and peak in incidence between ages 25 and 44 years.1,10 The World Health Organization (WHO) has identified primary headaches including migraines as a major public health problem due to their high prevalence, widespread age and geographic distribution, and their significant functional and socioeconomic impact.11

In the 1970s, Feinstein coined the term comorbidity to refer to the “greater than coincidental association of two conditions in the same individual.”12,13 Over the years, it has become apparent from population-based studies that various psychiatric conditions are associated with migraine. These include major depression,8,14-27 bipolar disorder,28-31 panic disorder,16,26,32,33 phobia,26 and substance dependence.17 The risk of suicide has also been found to be increased in patients with migraine.17,34,35 Understanding the psychiatric correlates of migraine is very important for many reasons. For instance, depressive disorders are one of the leading causes of disability worldwide36 and the WHO estimates that major depressive disorder (MDD) will become the second leading cause of disease burden worldwide by the year 2020, second only to ischemic heart disease.37 Furthermore, individuals with migraine and psychiatric comorbidities are greater health resource users than migraineurs without psychiatric conditions. Finally, recognizing psychiatric comorbidities in those with migraine should result in improvement in patient management, by alerting physicians that first-line treatment should be targeted at both conditions, ie, migraines and the associated mental health disorder.

Although various psychiatric conditions have been studied in migraine patients, this is the first comprehensive Canadian population-based study exploring the association between migraine and several comorbid psychiatric conditions in a large cohort. This study is also important for several other reasons: (1) it uses a fully structured, validated interview (face-to-face interviews in 86% of subjects) based on the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria for the assessment of mental disorders; (2) it is the first ever national survey on mental health; (3) it is representative of the entire Canadian population (all provinces); (4) it uses a very large cohort (n = 36,984); it has a very good response rate (77%); and (5) it not only addresses the psychiatric comorbidities but also addresses medication use, quality-of-life, disability, and various other functional and socioeconomic variables. The primary objectives of this study were to (1) determine the prevalence of various psychiatric conditions in association with migraine in a large population sample; and (2) describe the pattern of association of this comorbidity with a variety of health-related outcomes.

METHODS

Subjects.— The Canadian Community Health Survey, Mental Health and Wellbeing (CCHS 1.2) was conducted in 2002. It is a national/provincial level survey designed to provide reliable, comparable, and comprehensive data on mental health conditions. The CCHS has a 2-year collection cycle comprising 2 surveys: a regional survey in the first year (eg, Cycle 1.1, 2.1, etc) and a province-level survey in the second (eg, Cycle 1.2, 2.2, etc). Each second year of the survey cycle is designed to focus in depth on a particular topic. Mental Health and Wellbeing was the focus of Cycle 1.2. The CCHS is a cross-sectional survey that uses a multistage stratified cluster design combined with random sampling methods to select a representative sample of the Canadian population.38 Data collection for this survey began in 2002 and extended over 8 months. The CCHS 1.2 sample included 36,984 subjects ages 15 years and older, randomly selected from the population of household residents in Canada, with the exception of some native groups, armed forces personnel, and residents of some remote areas. See below under survey design for further information on the CCHS 1.2 survey design. However, for a comprehensive review of the CCHS, please refer to Gravel and Beland.38 Ethics approval was obtained for this study.

Survey Design.— The subjects were interviewed whenever possible in person at their homes by trained and experienced interviewers using computer-assisted interview methods.38 Telephone interviews were permitted only when travel was prohibitive or the respondent refused to conduct the interview in person. In total, 5173 (out of 36,984) interviews were conducted over the telephone. The interview included the World Mental Health Composite International Diagnostic Interview (WMH-CIDI).39 The WMH-CIDI is a lay-administered psychiatric interview that generates a profile of those with a disorder according to the definitions of the DSM-IV. The WHM-CIDI has been validated.39 More detailed information regarding the WHM-CIDI can be found in a review paper by Kessler and Üstün.39 The version of the WMH-CIDI used in the CCHS 1.2 produced diagnoses of MDD, bipolar disorder, panic disorder, agoraphobia, and social phobia. The WMH-CIDI interview also evaluated subjects for illicit drug dependence and, in the CCHS 1.2, this was supplemented by an assessment of alcohol dependence that used the CIDI Short Form.40 The CIDI Short Form generates 12-month prevalence estimates, but not lifetime prevalence estimates.

Diagnosis of Migraine.— The CCHS 1.2 subjects were read a list of chronic medical conditions and asked whether they had been diagnosed with one of these conditions by a health professional. The exact wording of the item was: “Now I'd like to ask you about certain chronic health conditions which you may have. We are interested in long-term conditions which are expected to last or have already lasted six months or more and that have been diagnosed by a health professional.” This was followed by: “Do you have migraine headaches?” A list of chronic conditions followed, and subjects were periodically reminded of the requirement for a diagnosis by a health professional with this statement: “Remember, we're interested in conditions diagnosed by a health professional. Do you have . . .”

Income definition.— To evaluate income, the CCHS 1.2 derived categories using total family income adjusted for family size. For the purposes of this study, low income was defined as falling below the middle income level, so that low and low middle income were combined. High income referred to the combination of middle, upper middle, and highest income categories in the CCHS. For example, the lowest household income would be less than $15,000 for 1 or 2 people; less than $20,000 for 3 or 4 people and less than $30,000 for 5+ people. The highest income category would be more than $60,000 for 1 or 2 or more than $80,000 for 3+ individuals.

Quality-of-Life Definition.— The CCHS 1.2 did not include a validated quality-of-life scale. As a measure of quality-of-life, a single item referring to “satisfaction with life in general” was used.

Two-Week Disability and Restriction of Activities.— The CCHS 1.2 defines 2-week disability as “the number of days in the last two weeks when the respondent stayed in bed or cut down in activities because of illness or injury.” Restriction of activities is a crude measure of the impact of long-term physical conditions, mental conditions, and health problems on 3 principal domains of life: home, work, or school, and other activities.

Definition of Mental Health Care Utilization.— Mental health care utilization is a derived variable indicating 12-month use of any resources for problems concerning emotions, mental health, or use of alcohol and drugs. It assesses whether the respondent was ever hospitalized overnight or ever consulted a professional, used an Internet support group or chat room, went to a self-help group, or used a telephone helpline in the 12 months prior to the interview.

Statistical Methods.— The CCHS 1.2 used a complex sampling strategy that involved both stratification and clustering. These design features require the application of sampling weights and specialized variance estimation procedures. Statistics Canada recommends a bootstrap procedure for variance estimation. This procedure was used, along with sampling weights in producing the estimates presented in this paper. Initially, prevalence was estimated in respondents with and without migraines. Next, these prevalence estimates were stratified on demographic variables. Finally, adjusted estimates of association were generated using logistic regression analysis.

RESULTS

Baseline Demographics.— A total of 36,984 subjects were successfully interviewed for the CCHS 1.2, representing a 77% response rate nationally. Table 1 lists the prevalence of migraine based on the subjects’ baseline characteristics. The sampling was highly representative of the population of household residents of Canada after weighing for age, sex, and geographic distribution.38 Overall, the prevalence of migraine in the total sample (n = 36,984) was 15.2% for females and 6.1% for males. Migraine was most common in those between the ages 25 and 44 (13.1%), followed by ages 45-64 (11.0%), ages 15-24 (9.9%), and least common in those greater than 65 years of age (5%). Individuals with the lowest or lower middle income (15.4% and 13.0% respectively) were more likely to suffer from migraine compared with those with the highest income (9.8%). The prevalence of migraine did not differ by education level, marital status or location of residence (urban vs rural).

Table 1.—. Prevalence of Migraine Based on Baseline Sample Characteristics (n = 3984/32,992)
CharacteristicNumberMigraine
(N = 3984)
% (95% CI)
  • Above percentages (%) represent weighted estimates of overall sample (N = 36,984), thus % for the row, not the column.

Gender
 Male10666.1 (5.6-6.6)
 Female 291815.2 (14.5-16.0)
Age (years)
 15-246479.9 (8.9-10.8)
 25-44174413.1 (12.3-13.9)
 45-64121811.0 (10.2-11.9)
 ≥653755.0 (4.3-5.7)
Education
 High school graduation or less177610.7 (10.1-11.3)
 Some postsecondary education218910.8 (10.1-11.6)
Marital status
 Single113210.1 (9.3-11.0)
 Married/common law204310.7 (10.1-11.3)
 Widowed/separated/divorced80511.9 (10.7-13.1)
Location of residence
 Urban315110.7 (10.2-11.2)
 Rural83310.9 (9.9-11.9)
Income level
 Lowest income26715.4 (12.8-18.1)
 Lower middle income43613.0 (11.3-14.7)
 Middle income82910.0 (9.1-10.8)
 Upper middle income123011.4 (10.5-12.2)
 Highest income8729.8 (8.9-10.6)

Prevalence of Psychiatric Comorbidities in Those with Migraine (Table 2).— The 12-month and lifetime prevalence of mental disorders was more than twice as high in those with migraine compared with those without migraine as shown in Table 2. For example, the lifetime prevalence of MDD was 18.8% (CI = 17.0-20.5) in migraineurs vs 9.8% (CI = 9.3-10.3) in nonmigraineurs, and the 12-month prevalence was 8.6% (CI = 7.3-9.8) in migraineurs vs 3.4% (CI = 3.1-3.7) in nonmigraineurs. A similar trend was noted for bipolar disorder, panic disorder (and agoraphobia with panic symptoms), and social phobia, all occurring more than twice as often in those with migraine compared with those without migraine. However, there was no difference in the 12-month adjusted prevalence of drug, alcohol or substance dependence in migraine subjects compared with nonmigraine subjects. The lifetime prevalence of drug, alcohol, or substance dependence in migraineurs was not available (see above).

Table 2.—.  Prevalence of Mental Health Disorders in Association With Migraine
Psychiatric comorbidityMigraine
(n = 3984)
%
(95% CI)
No migraine
(n = 32,992)
%
(95% CI)
Overall population
(n = 36,984)
%
(95% CI)
  • Interpret with caution: high sampling variability (CV 16.6-33.3%).

  • CV = coefficient of variation.

12 month prevalence
 Major depressive disorder8.6 (7.3-9.8)3.4 (3.1-3.7)3.9 (3.7-4.2)
 n = 1563
 Bipolar disorder2.8 (2.2-3.4)0.7 (0.6-0.9)1.0 (0.8-1.1)
 n = 393
 Panic disorder/agoraphobia with panic symptoms without panic disorder5.5 (4.5-6.4)1.6 (1.4-1.8)2.0 (1.8-2.3)
 n = 797
 Social phobia6.7 (5.7-7.8)2.5 (2.3-2.8)3.0 (2.7-3.2)
 n = 1189
 Drug dependence1.1 (0.6-1.5)0.7 (0.6-0.9)0.8 (0.6-0.9)
 n = 288
 Alcohol dependence2.3 (1.7-2.9)2.6 (2.4-2.8)2.6 (2.4-2.8)
 n = 1048
 Drug dependence3.1 (2.3-3.8)3.1 (2.8-3.3)3.0 (2.8-3.3)
 n = 1215
Lifetime prevalence
 Major depressive disorder18.8 (17.0-20.5)9.8 (9.3-10.3)10.8 (10.3-11.2)
 n = 4110
 Bipolar disorder5.4 (4.5-6.4)2.0 (1.8-2.2)2.4 (2.1-2.6)
 n = 938
 Panic disorder/agorophobia with panic symptoms without panic disorder9.6 (8.4-10.8)4.0 (3.7-4.4)4.6 (4.3-5.0)
 n = 1397
 Social phobia14.3 (12.8-15.8)7.3 (6.9-7.7)8.1 (7.7-8.5)
 n = 3061

One-Year Prevalence of Mental Health Disorders in Migraineurs Based on Gender, Age, Education, Marital Status, Location of Residence, or Income Level.— (1) MDD (Table 3) – There were no differences in the sex-specific prevalence of migraine and MDD between males and females, although there was a trend for females to have a higher prevalence. Although not significant, there was also a trend for less comorbid migraine and MDD with age or with higher education level. Those who were widowed/separated/divorced were more likely than those who were married/common-law or single to have migraine and MDD. Those in the lowest and lower middle income categories were more likely to have migraine and MDD than those of higher income. Location of residence did not seem to be associated with combined migraine and MDD. Inclusion of all statistically significant covariates in a logistic regression model resulted in an adjusted odds ratio of 2.3 (95% CI 1.9-2.8). (2) Bipolar disorder (Table 3) – Gender, age, education, and location of residence were not associated with combined migraine and bipolar disorder, that is the same pattern of association was present regardless of whether subjects had migraine or not. However, there was a trend (although not significant) for combined bipolar disorder and migraine to be more prevalent in males and less prevalent with older age or higher education level. Similar to MDD, combined migraine and bipolar disorder was most common in those with the lowest and lower middle income compared with those of higher income, and was less frequent in those who were married/common law compared with those who were single. A logistic regression model simultaneously adjusting for each of the covariates yielded an adjusted odds ratio of 3.7 (95% CI 2.7-5.0). (3) – Anxiety disorders (Table 4) – Anxiety disorders (panic disorder and social phobia) and migraine were not associated with gender, age, education, or location of residence, that is the same pattern of association was present regardless of whether subjects had migraine or not. However, those who were married/common-law were less likely to have migraine and anxiety disorders compared with those who were widowed/separated/divorced. Those who were of the lowest and lower middle income were more likely than those of higher income to suffer from migraine and comorbid panic disorder, but not to suffer from social phobia. When all of the covariates were included in a logistic regression model, the adjusted odds ratio was 2.3 (95% CI 1.9-2.9) for social phobia and 2.8 (95% CI 2.2-3.6) for panic disorder/agoraphobia.

Table 3.—. Stratified One-Year Prevalence of Mood Disorders Based on Migraine Status by Age, Gender, Education, Marital Status, Location of Residence, and Income Level (% and 95% CI )
Migraine (n = 3984)
No migraine (n = 32,992)
Major depressive disorder
n = 1563
Bipolar disorder
n = 393
No mental health disorders
n = 31,503
Migraine
n = 375
No migraine
n = 1187
Migraine
n = 126
No migraine
n = 266
Migraine
n = 2942
No migraine
n = 28,559
  • Interpret with caution: high sampling variability (CV 16.6-33.3%).

  • Not releasable according to Statistics Canada guidelines: estimates unreliable (CV > 33.3%).

  • CV = coefficient of variation.

Gender
 Males7.42.6 3.20.875.487.5
 n = 16,773(5.3-9.6)(2.3-3.0)(2.1-4.2)(0.6-1.0)(72.1-78.8)(86.8-88.2)
 Females9.04.22.60.775.587.4
 n = 20,211(7.5-10.5)(3.8-4.6)(1.9-3.3)(0.5-0.8)(73.3-77.8)(86.7-88.1)
Age (years)
 15-249.74.6 4.81.469.181.6
 n = 5673(6.8-12.6)(3.8-5.4)(2.7-6.8)(1.1-1.8)(64.1-73.9)(80.1-83.0)
 25-449.13.63.30.974.586.9
 n = 12,813(7.0-11.2)(3.1-4.0)(2.2-4.3)(0.7-1.1)(71.4-77.5)(86.1-87.7)
 45-647.53.4 1.7 0.678.688.9
 n = 10,762(5.7-9.3)(2.8-4.0)(0.9-2.5)(0.4-0.8)(75.3-81.9)(88.0-89.8)
 65+ 7.21.682.892.3
 n = 7736(3.3-11.2)(1.2-2.0)  (77.7-87.9)(91.4-93.1)
Education
 High school graduation or less10.03.13.30.972.687.1
 n = 17,089(8.1-11.9)(2.7-3.5)(2.3-4.3)(0.6-1.1)(69.5-75.8)(86.3-87.8)
 Some postsecondary education7.33.62.30.677.987.8
 n = 19,664(5.6-9.0)(2.4-4.0)(1.6-3.0)(0.5-0.8)(75.4-80.4)(87.2-88.5)
Marital status
 Single9.34.95.01.267.382.0
 n = 9798(7.2-11.5)(4.1-5.6)(3.4-6.6)(0.9-1.5)(63.5-71.0)(80.9-83.1)
 Married/common-law6.62.4 1.70.580.490.3
 n = 19,184(5.1-8.2)(2.1-2.7)(1.1-2.4)(0.4-0.7)(77.9-82.8)(89.7-90.9)
 Widowed/separated/divorced15.65.23.5 0.968.284.9
 n = 7970(11.7-19.5)(4.5-6.0)(2.1-5.0)(0.6-1.2)(63.3-73.1)(83.6-86.2)
Location of residence
 Urban8.83.22.80.874.987.3
 n = 28,746(7.3-10.3)(3.2-3.8)(2.2-3.5)(0.6-0.9)(72.7-77.1)(86.8-87.9)
 Rural 7.62.8 2.50.678.288.2
 n = 8238(5.0-10.1)(2.1-3.4)(1.0-4.1)(0.3-0.8)(74.2-82.2)(86.9-89.4)
Income level
 Lowest and lower middle income12.03.94.40.870.186.0
 n = 13,049(9.6-14.4)(3.5-4.4)(3.2-5.5)(0.6-1.0)(66.6-73.6)(85.2-86.9)
 Middle or high income6.73.0 2.00.778.988.6
 n = 20,499(5.0-8.4)(2.7-3.4)(1.3-2.6)(0.5-0.9)(76.3-81.5)(88.0-89.2)
Table 4.—. One-Year Prevalence of Anxiety Disorders Based on Migraine Status Stratified by Age, Gender, Education, Marital Status, Location of Residence, and Income Level (% and 95% CI)
Migraine (n = 3984)
No migraine (n = 32,992)
NumberPanic disorder/ agoraphobia
n = 797
Social phobia
n = 1189
No disorder
n = 31,503
Migraine
n = 239
No migraine
n = 558
Migraine
n = 288
No migraine
n = 900
Migraine
n = 239
No migraine
n = 558
  • Interpret with caution: high sampling variability (CV 16.6-33.3%).

  • Not releasable according to Statistics Canada guidelines: estimates unreliable (CV > 33.3%).

  • CV = coefficient of variations.

Gender
 Males16,7734.01.17.02.375.487.5
  (2.7-5.4)(0.9-1.3)(4.9-9.2)(1.9-2.6)(72.1-78.8)(86.8-88.2)
 Females20,2116.02.26.62.875.587.4
  (4.8-7.3)(1.8-2.5)(5.4-7.9)(2.5-3.2)(73.3-77.8)(86.7-88.1)
Age (years)
 15-2456735.22.07.64.469.181.6
  (3.3-7.0)(1.5-2.5)(5.2-10.1)(3.6-5.1)(64.1-73.9)(80.1-83.0)
 25-4412,8136.12.07.72.874.586.9
  (4.6-7.5)(1.6-2.4)(6.0-9.3)(2.4-3.2)(71.4-77.5)(86.1-87.7)
 45-6410,7625.41.55.82.178.688.9
  (3.4-7.4)(1.2-1.9)(4.0-7.6)(1.7-2.5)(75.3-81.9)(88.0-89.8)
 65+77365.41.50.882.892.3
  (3.4-7.4)(1.2-1.9) (0.5-1.0)(77.7-87.9)(91.4-93.1)
Education
 High school graduation or less17,0896.01.97.02.772.687.1
  (4.3-7.7)(1.5-2.2)(5.6-8.5)(2.4-3.1)(69.5-75.8)(86.3-87.8)
 Some postsecondary education19,6645.11.56.62.477.987.8
  (3.8-6.3)(1.2-1.7)(5.1-8.1)(2.0-2.7)(75.4-80.4)(87.2-88.5)
Marital Status
 Single97986.22.17.94.367.382.0
  (4.6-7.8)(1.7-2.5)(5.9-9.8)(3.7-4.9)(63.5-71.0)(80.9-83.1)
 Married/common-law19,1844.31.35.61.980.490.3
  (3.1-5.5)(1.0-1.5)(4.2-7.0)(1.6-2.1)(77.9-82.8)(89.7-90.9)
 Widowed/separated/divorced79709.42.49.92.268.284.9
  (5.9-13.0)(1.6-3.2)(7.1-12.7)(1.8-2.6)(63.3-73.1)(83.6-86.2)
Location of residence
 Urban28,7465.81.77.02.674.987.3
  (4.7-6.9)(1.4-1.9)(5.8-8.3)(2.3-2.8)(72.7-77.1)(86.8-87.9)
 Rural82384.31.55.52.478.288.2
  (2.3-6.2)(1.0-2.0)(3.4-7.6)(1.8-3.0)(74.2-82.2)(86.9-89.4)
Income level
 Lowest and lower middle income13,0498.52.37.72.970.186.0
  (6.3-10.6)(1.8-2.8)(5.9-9.5)(2.5-3.3)(66.6-73.6)(85.2-86.9)
 High20,4994.21.46.32.478.988.6
  (3.1-5.3)(1.2-1.7)(4.9-7.7)(2.1-2.7)(76.3-81.5)(88.0-89.2)

One-Year Prevalence of Mental Health Disorders in Migraineurs Compared With Nonmigraineurs Based on Gender, Age, Education, Marital Status, Location of Residence or Income Level (Tables 3 and 4).— As noted in Tables 3 and 4, an association between migraine and mental health disorders is apparent across all demographic categories, with a higher prevalence of mental health disorders in the migraineurs compared with the nonmigraineurs regardless of baseline sociodemographic variables. However, similar epidemiologic patterns are noted in the nonmigraineurs and the migraineurs with respect to the possible effect of the various sociodemographic variables on the occurrence of mental health disorders. For example, mental health disorders appear to be less common in the older age group, in those who are married or in a common-law relationship, and in those of higher income whether one has migraine or not. Thus, having migraine appears to be contributing to a higher prevalence of the various mental health disorders but is not obviating the possible synergistic effects of various sociodemographic factors.

Medication Use in Relation to Migraine and Psychiatric Comorbidities.—Table 5 provides information regarding medication use in those of various migraine and mental health status. Antidepressant use was significantly higher in those with combined migraine and mental health disorder compared with those with neither or one of those conditions. For example, overall antidepressant use was 3.1% (CI 2.9-3.4) in those without migraine or MDD, compared with 8.2% (CI 7.0-9.32) in those with migraine alone, compared with 26.7% (22.8-30.4) in those with MDD alone, compared with 35.9% (CI 28.9-43.0) in those with combined MDD and migraine. Similar findings were found for those with bipolar disorder and anxiety disorder. Selective serotonin reuptake inhibitors (SSRIs) were the most commonly prescribed antidepressants. The only exception to the rule that antidepressant or antiepileptic use was higher in those with psychiatric conditions compared with those with migraine alone was for the tricyclic antidepressants. Their use was higher in those with migraine than those without migraine and a mental health disorder.

Table 5.—. Medication Use in Association with Migraine and Psychiatric Comorbidities (% With 95% CI)
 No migraineMigraineNo migraineMigraine
No mental health disorderNo mental health disorderMental health disorderMental health disorder
n = 31,514n = 3548n = 1187n = 375
  • Interpret with caution: high sampling variability (CV 16.6%-33.3%).

  • Not releasable according to Statistics Canada guidenlines: estimates unreliable (CV > 33.3%).

  • §

    Subset of TCA group.

  • CV = coefficient of variations; MDD = major depressive disorder; SNRI = selective norepinephrine reuptake inhibitor; SSRI = selective serotonin reuptake inhibitor; TCA = tricyclic antidepressant.

MDD
 n = 1563
 Overall antidepressant use3.1 (2.9-3.4)8.2 (7.0-9.32)26.7 (22.8-30.4)35.9 (28.9-43.0)
 n = 2143
 TCA
 n = 422
0.6 (0.5-0.7)2.1 (1.5-2.7)4.0 (1.8-6.1)
 Amitryptiline§
 n = 288
0.4 (0.3-0.4)1.8 (1.2-2.3)
 SSRI
 n = 1292
1.9 (1.6-2.0)4.4 (3.6-5.3)16.1 (13.2-19.0)22.0 (16.2-27.8)
 SNRI
 n = 336
0.4 (0.3-0.5)1.4 (0.8-1.9)7.1 (4.2-10.1)10.3 (5.8-14.7)
 Trazadone
 n = 122
0.2 (0.1-0.2)0.7 (0.3-1.0) 1.2 (0.6-1.7)3.0 (1.1-5.0)
 Anticonvulsants
 n = 176
0.3 (0.2-0.3)0.9 (0.4-1.4) 1.4 (0.6-2.3)
Bipolar disorder
 n = 938
 Overall antidepressant use
 n = 2143
3.9 (3.6-4.2)10.3 (9.0-11.5)17.6 (11.6-23.6)30.8 (21.6-40.0)
 TCA
 n = 422
0.7 (0.6-0.8)2.2 (1.6-2.7)
 Amitryptiline
 n = 288
0.4 (0.3-0.5)1.8 (1.2-2.3)
 SSRI
 n = 1292
2.4 (2.1-2.6)5.6 (4.6-6.6)11.6 (7.0-16.1)22.5 (15.1-29.8)
 SNRI
 n = 336
0.6 (0.5-0.7)2.2 (1.6-2.8)
 Trazadone
 n = 122
0.2 (0.1-0.2)0.7 (0.4-1.0)
 Anticonvulsants
 n = 176
0.3 (0.2-0.3)0.9 (0.4-1.4)
Anxiety disorder
 n = 1803
 Overall antidepressant use
 n = 2143
3.3 (3.0-3.6)8.4 (7.2-9.6)18.8 (15.8-21.7)29.0 (23.1-34.9)
 TCA
 n = 422
0.6 (0.5-0.7)1.8 (1.2-2.3)2.4 (1.1-3.8)5.5 (2.8-8.2)
 Amitriptyline
 n = 288
0.4 (0.3-0.5)1.5 (1.0-2.0)0.5 (0.2-0.8)4.3 (1.8-6.9)
 SSRI
 n = 1292
1.9 (1.7-2.2)4.7 (3.8-5.6)12.8 (10.4-15.2)16.7 (11.8-21.7)
 SNRI
 n = 336
0.5 (0.4-0.7)1.7 (1.1-2.3)2.7 (1.4-4.0)6.3 (3.2-9.4)
 Trazadone
 n = 122
0.2 (0.1-0.2)0.6 (0.3-0.8)0.7 (0.3-1.2)
 Anticonvulsants
 n = 176
0.2 (0.2-0.3)0.9 (0.4-1.4)1.6 (0.8-2.4)

Migraine and Mental Health Disorders and Their Association With Functional Status and Socioeconomic Variables.— The Figure illustrates the association between having migraine and a mental health disorder (MDD, bipolar disorder or anxiety disorder) on 2-week disability, restriction of activities, quality-of-life or mental health care use. Having migraine and any of the above psychiatric comorbidities increased the likelihood of a poor outcome. Thus, those with comorbid migraine and a mental health disorder were more likely to have 2-week disability, restriction of activities, poorer quality-of-life or mental health care use compared with those with neither condition or those with only one condition.

Figure Figure.—.

Figure Figure.—.

Migraine and mental health disorders and their association with functional status and socioeconomic variables. (A) major depressive disorder (MDD); (B) bipolar disorder; (C) anxiety disorder. Error bars represent 95% confidence interval.

Figure Figure.—.

Figure Figure.—.

Migraine and mental health disorders and their association with functional status and socioeconomic variables. (A) major depressive disorder (MDD); (B) bipolar disorder; (C) anxiety disorder. Error bars represent 95% confidence interval.

Association Between Migraines and Mental Health Disorders Nationally.— The odds ratios associated with various mental health disorders were generally comparable between the various Canadian provinces. The only significant finding was that the odds ratio of migraine predicting social phobia in British Columbia (OR 4.8 with a CI 1.8-3.0) was higher than in Alberta (OR 1.3 with a CI 0.6-1.4). In the absence of plausible explanatory hypotheses, this result may represent a type I error. Otherwise, there were no significant regional differences in the odds ratios for various mental health disorders in relation to migraines.

DISCUSSION

The key findings of this study are: (1) Migraine is more than twice more common in females than in males, and is most common in those aged 25-44 years and in those of lower income. (2) Migraine is associated with MDD, bipolar disorder, panic disorder, and social phobia, all occurring twice as often in those with migraine compared with those without it. However, migraine is not associated with drug, alcohol, or substance dependence. (3) The higher prevalence of mental health disorders in migraineurs is not accounted for by sociodemographic variables. However, psychiatric disorders are less common in those over 65 years, in those who are in a relationship, and in those of higher income whether migraine is present or not. (4) Antidepressant and antiepileptic drug use is generally highest in those with migraine and a mental health disorder, intermediate in those with a mental health condition but without migraine, and lowest in those with migraine alone or neither condition. (5) Health-related outcomes are poorer in those with both migraines and a psychiatric disorder than in those with either condition alone.

The Prevalence of Migraine is Consistent With Prior Population-Based Studies.— The prevalence of migraine obtained in this population-based study of 15.2% in females and 6.1% in males is consistent with prior studies,1-3,5-7,10 but lower than in one Canadian study reporting a prevalence of 7.8% in males and 24.9% in females.4 One possible explanation for the lower prevalence rates could be the different methods used to diagnose migraine. In the CCHS, subjects answered “yes” to having migraine only if they were diagnosed by a health professional. International Headache Society criteria were not incorporated in the survey.41 Thus, although questions regarding the presence of medical conditions in the CCHS specified that the condition be diagnosed by a health professional, there remained the potential for individuals to over- or under-report migraine. Under- or over-ascertainment of migraine could have occurred for several reasons including for example: (1) bias due to self-report; (2) misdiagnosis or (3) undiagnosed as subjects never seeked medical attention for migraine. In fact, studies have shown that 27-70% of migraineurs have never been diagnosed by a physician.4,10,42-44 A recent Canadian survey found that 48% of women with migraine had never consulted a physician for their headaches.44 However, when a diagnosis of migraine is made by a primary care physician, it is usually accurate (98% of the time in the Landmark study).45 Thus, those who self-reported having physician-diagnosed migraine in our study most likely do indeed have migraine. It is more likely rather that under-ascertainment of migraine occurred as many migraineurs do not seek medical attention for their headaches.44 For that reason, some may argue that those with severe migraine are more likely to seek medical attention and be diagnosed, and that as a result, our findings are more applicable to severe migraine sufferers. However, it has been shown that substantial disability also occurs in a high proportion of migraineurs who never consulted a physician (UK 60%, USA 68%), in those who never received a correct medical diagnosis (UK 64%, USA 77%) and in those who were only treated with over-the-counter medications (UK 72%, USA 70%).46 Thus, although the prevalence of migraine is likely underestimated in our study, we are still confident that the majority of those who are undiagnosed migraineurs are not mild migraine sufferers, but rather are similar to our diagnosed sample of migraineurs. Furthermore, although International Headache Society criteria were not used in our study, our findings are in keeping with prior published reports on the epidemiology of migraine,1-3,5-7,10 suggesting that although we may be underdiagnosing migraine, our migraineurs are still fairly representative of the typical migraine sufferer.

Our study is also consistent with the literature in that migraine is up to 4 times more common in women than men,1-4,6-10 and peaks in incidence between ages 25 and 44 years.1,10 The prevalence of migraine was also highest in those of lower/lowest income, consistent with a prior large US study with 20,468 subjects,6 but inconsistent with a smaller prior Canadian study with 1573 subjects which found no association between income and migraine.1 It is possible that the prior Canadian study was underpowered to detect this effect (type II error).

Psychiatric Comorbidities Are Associated With Migraine.— The previous association described between migraine and MDD,8,14-27 bipolar disorder,28-31 panic disorder,16,26,32,33 and phobia26 was confirmed in this study. Conversely, the association between migraine and substance dependence reported in a prior study was not reproduced in our study.17

One of the main strengths of this study is its high participation rate and sampling procedure, providing data which can be interpreted as representative of all Canadians and resulting in less bias than clinical samples. However, our analysis was based on cross-sectional data, thus causal inference cannot be supported. The associations observed between migraine and the various psychiatric conditions may be due to an etiological effect in either direction, probably in both directions. A bidirectional association was found to exist in prior studies between migraine and depression,18,19 and migraine and panic disorder,32 but this could not be evaluated in our study due to the cross-sectional nature of the survey. However, the results obtained through cross-sectional analyses are extremely useful in understanding the burden associated with migraine and its comorbidities and in guiding public health policies and funding for health-care services. It is also important to understand whether an association exists between migraine and the various mental health disorder because treating depression in subjects with comorbid medical conditions may lead to improved symptom levels and functioning.47,48

In this study, mental health disorders were found to be higher in those with migraine compared with those without migraine regardless of sociodemographic variables. However, psychiatric disorders were less common in those over 65 years, who were in a relationship or were in the higher income bracket regardless of migraine status. These epidemiological findings are consistent with the literature reporting a decrease in major depression in older age groups.24,49 There was also a trend toward major depression being more common in females than males in the migraine group; and this was statistically significant in the nonmigraineurs. Thus, even though migraine and depression may be linked at a medical or even biological level, these sociodemographic factors are important determinants of mental health, and the trends observed have important implications for service delivery, to ensure adequate provision of support (eg, social work, etc) in relevant clinical settings.

Medication Use in Migraineurs is Different Than Medication Use in Those With a Mental Health Disorder Alone.— Antidepressant and antiepileptic drug use was highest in those with combined migraine and a mental health disorder, followed by those with a mental health disorder without migraine, those with migraine but no mental health conditions, and finally those with neither condition. These trends are not surprising as many antidepressants or antiepileptic drugs are used for migraine prophylaxis and for the treatment of mood or anxiety disorders. The only exception is tricyclic antidepressants which were used more in those with migraines than those without migraine and with a mental health disorder. Tricyclic agents are very commonly used for migraine prophylaxis, but their use in the management of mood disorders in particular has declined over the years with the introduction of the SSRIs and selective norepinephrine reuptake inhibitors (SNRIs). This is likely the explanation for the trend observed. Another interesting observation is the frequent use of antidepressants in those with migraine but without a mental health disorder, consistent with a prior study using the CCHS.31 In the study by Beck et al, migraine, fibromyalgia, anxiety disorder, or past depression was present in more than 60% of those taking antidepressants without a past-year episode of depression. They concluded that antidepressant use is now employed extensively for indications other than depression. Another population-based survey study of 9428 subjects also found that the use of anxiolytics and antidepressant was high in those with migraine when compared with other chronic conditions.50

Migraine and Psychiatric Comorbidities Are Associated With Poorer Health-Related Outcomes.— Subjects with comorbid migraine and a mental health disorder were more likely to have 2-week disability, restriction of activities, poorer quality-of-life or mental health care use compared with those with just one of the 2 conditions, and even more so than those with neither condition. Breslau et al17 found similar results in a prospective study of 1007 young adults. Those with migraine were more likely to have a history of various mental health disorders and were also more likely to report job absenteeism, assess their general health as fair or poor, and use mental health services. McIntyre et al,30 using the CCHS found that bipolar males with migraine were more likely than those without migraine to utilize mental health care services and that bipolar females with comorbid migraine were more likely to require help with personal or instrumental activities of daily living when compared with bipolar females without migraine. Other studies which have examined health-related outcomes such as disability, restriction of activity, quality-of-life or mental health care utilization are restricted to migraine sufferers without taking into consideration associated psychiatric comorbidities. For example, in an earlier study on the impact of migraine on lifestyle,5 consulting behavior and medication use, Edmeads et al found that regular activities were limited in 78% of migraine attacks, consistent with our results showing restriction of activities in those with migraine. They also found that 11% of the reported migraine events caused the headache subjects to leave or not to report to work and that many describe their migraine as adversely affecting relationships with family, friends, and colleagues. Some subjects even indicated that having migraine affected their job decisions. Lipton et al found that 85% of those with migraine reported substantial reductions in their ability to do household work and chores, 45% missed family social and leisure activities, and 32% avoided making plans for fear of cancellation due to headaches.51 A Canadian clinic-based study found that the degree of disability suffered by migraineurs referred to their headache specialty clinic was very significant using the Headache Impact Test and the Migraine Disability Questionnaire.2 One limitation to our study is that a validated quality-of-life scale was not used. However, although there are no studies to our knowledge examining the validity of single-item quality-of-life scales for headache research, the use of single-item rating scales to assess various aspects of health and function is well established.52-56 Recently, evidence from large cohorts of patients confirmed that single-item scales are valid and reliable to assess symptom severity, psychosocial function, and quality-of-life.56 However, satisfaction with life in general was lower in those with migraine and a mental health disorder, which is consistent with prior studies. For example, in a nationwide population-based survey performed in France on 10,532 adult subjects, quality-of-life was measured using the health-related quality-of-life scale (HRQoL), a well-validated scale.57 Migraine subjects showed significantly lower mean scores than nonmigraine subjects at all 8 HRQoL concepts except for physical functioning. A similar impact of migraine on quality-of-life using the HRQoL was also found by Lipton et al.15,58

Another limitation to our study is that our health-related outcome variables are not migraine or mental health-specific. There may be several mechanisms linking poor mental health, migraine, and poor quality-of-life. For example, poor quality-of-life in migraineurs with mental health problems may reflect altered clinical status or life circumstances, altered perceptions of life circumstances, or both. Our cross-sectional data have allowed us to effectively quantify these associations at a point in time, but prospective studies will be needed to disentangle underlying mechanisms.

This study was not designed to address the mechanisms responsible for comorbid migraine and mental health disorders. Three basic theories have been described in the literature:59 (1) that mental disorders cause migraine; (2) that migraine causes mental health disorders; and (3) that a common shared biological factor exists to explain the co-occurrence of these conditions. A bidirectional relationship has been described between migraine and depression,18,19 and migraine and panic disorder,32 providing evidence for the above described theories. Various biologically based studies have looked at the association between migraine and mental disorders. A particular dopamine D2 receptor genotype was found to be associated with migraine, major depression, generalized anxiety disorder, panic attacks, and phobia.60 Low tyramine conjugation, a marker of endogenous depression, was associated with a lifetime history of major depression in migraineurs compared with nonmigraineurs.61 This finding ruled out the possibility that the depression was a psychological reaction to migraine attacks. Serotonin receptors and transporters as well as adrenaline have also been implicated in migraine and various mental health disorders.62,63 There are also numerous studies supporting the use of various antidepressants in the prevention or treatment of migraine including tricyclic antidepressants, anxiolytic, SNRI, and other forms of antidepressants.24 Thus, although various theories have been suggested to explain the association between migraine and various psychiatric conditions, more studies are needed to elucidate the mechanism of this association.

CONCLUSION

In summary, the data reported in this paper suggest an association between migraine and various mental health disorders. Having both types of conditions is associated with poorer health-related outcomes including decreased quality-of-life, restricted activities, increased disability, and increased mental health services utilization. From the public health perspective, these findings emphasize the important contribution of medical morbidity to the burden of mental health disorders in modern society. These findings are important in planning health services and provision of adequate medical therapy in individuals with migraine and comorbid mental health conditions.

Acknowledgments

Acknowledgment: This study was supported by an operating grant from the Research Coordinating Committee of the Institute of Health Economics (http://www.ihe.ab.ca) received by one of the paper's authors.

Ancillary