Psychiatric Comorbidity of Migraine

Authors

  • Sandra W. Hamelsky PhD, MPH,

  • Richard B. Lipton MD


  • From the Department of Neurology, Albert Einstein College of Medicine, Bronx, NY, USA (Drs. Hamelsky and Lipton); Department of Epidemiology and Population Health, Albert Einstein College of Medicine, Bronx, NY, USA (Dr. Hamelsky); and Montefiore Headache Center, New York, NY, USA (Dr. Lipton).

Address all correspondence to Sandra W. Hamelsky, PhD, MPH, The Bradstreet Group, 1588 Route 130 North, North Brunswick, NJ 08902, USA.

Abstract

Migraine affects nearly 12% of the adult population in the United States and causes significant lost productivity and decrements in health-related quality of life. The burden of migraine and the challenge in managing it are increased by the comorbid psychiatric conditions that occur in association with it. Studies in both clinical and community-based settings have demonstrated an association between migraine and a number of specific psychiatric disorders. This review will focus on the relationships between migraine and depression, generalized anxiety disorder, panic disorder, and bipolar disorder. In large scale population-based studies, persons with migraine are from 2.2 to 4.0 times more likely to have depression. In longitudinal studies, the evidence supports a bidirectional relationship between migraine and depression, with each disorder increasing the risk of the other disorder. Migraine is also comorbid with generalized anxiety disorder (Odds Ratio [OR] 3.5 to 5.3), panic disorder (OR 3.7), and bipolar disorder (OR 2.9 to 7.3). A diagnosis of migraine should lead to a heightened level of diagnostic suspicion for these comorbid psychiatric disorders. Similarly, a diagnosis of one of these psychiatric disorders should increase vigilance for migraine. Treatment plans for migraine should be mindful of the comorbid conditions.

Migraine affects nearly 12% of the adult population in the United States,1,2 and causes significant lost productivity3,4 and decrements in quality of life.5 Part of the burden of migraine is produced by the psychiatric conditions that occur in association with it. When one disorder occurs with another with greater than chance frequency the disorders are said to be comorbid. Studies in both clinical and community-based settings have demonstrated an association between migraine and a number of specific psychiatric disorders. While the association between migraine and depression is most widely reported, there are also strong associations with other psychiatric disorders. Understanding the nature of the association between migraine and these psychiatric disorders has implications for diagnosis and treatment. The occurrence of comorbidity may also provide clues to the etiology of each disorder. This review will focus on the relationships between migraine and depression, generalized anxiety disorder, panic disorder, and bipolar disorder.

METHODOLOGICAL ISSUES

Studying comorbidity presents a series of methodological challenges. First, clinic-based studies are prone to overestimating comorbidity due to a phenomenon called Berkson bias. This bias arises when conditions may occur together in the clinic at increased frequency because of patterns of consultation and referral.6 For example, primary care doctors who tend to treat migraine with β-blockers may preferentially refer asthmatic migraineurs to neurologists, because β-blockers are contraindicated. As a consequence, a study in the neurologic offices might overestimate the association between migraine and asthma. Clinic-based studies should be viewed as hypothesis generating; studies of comorbidity are best conducted in representative samples of the general population. Second, studies of comorbidity must apply systematic methods for ascertaining both migraine and the other conditions under study. Comorbidity studies that rely on medical records review or medical claims data may overestimate comorbidity due to clinical detection and coding bias. For example, clinicians who code migraine may be more likely to code other pain disorders, potentially leading to overestimation of comorbidity. Third, studies of comorbidity can assess the cross-sectional or longitudinal association between migraine and another disorder. While cross-sectional studies demonstrate associations, they do not indicate directionality. Longitudinal studies make it possible to determine if one condition predisposes to the other (unidirectional relationship) or if each disorder predisposes to the other. Understanding directionality has implications for clinical practice and for exploration of disease mechanisms.

MIGRAINE AND DEPRESSION

The association between migraine and depression has been described in both clinic and community-based populations. Early studies established a cross-sectional association between the 2 disorders, while subsequent investigations examined the temporal sequence of the association. Persons with migraine were followed-up to determine if they developed depression at increased rates and persons with depression were followed-up to examine the incidence of migraine. To explain an association from migraine to depression, it has been hypothesized that unpredictable attacks of severe pain might lead to anxiety and depression, perhaps through mechanisms akin to those of learned helplessness (noncontingent punishment). Conversely, to explain a link from depression to pain, it has been postulated that headache could be a somatic manifestation of depression or impair the patients' ability to cope with pain.7,8 If the association is bidirectional (if each disorder predisposes to the other), the association may arise from an underlying common vulnerability.9

While a cross-sectional association between migraine and major depression has been reported in both clinical and population-based settings,10–17 herein we will focus on the studies in unselected populations because these studies support more robust causal inferences. Table summarizes the association between migraine and depression in a number of large scale community studies. In this table, estimates of the relative risk for the association between migraine and depression range from 2.2 to 4.0. In general, the cross-sectional studies measure strength of association using the Odds Ratio (OR), while most longitudinal studies measure association using the Hazard Ratio (HR).

Table Table.—. Select Studies of the Association Between Migraine and Depression
Author, YearIHS-BasedSource of ParticipantsCountryMigraine and Depression—OR, 95% CIBi-Directional Relationship—OR, 95% CI
  1. From Cox proportional hazards models; Age-adjusted; §Sex-adjusted; Age- and sex-adjusted; ††Sex-, age-, and education-adjusted.

Merikangas, 1990NoPopulationSwitzerland2.2, 1.1 to 4.8Not assessed
Breslau, 1994,‡YesHMO populationUSANot reportedNew-onset migraine 3.5, 2.2 to 5.6 New-onset depression 3.6, 2.6 to 5.2
Breslau, 1998§YesPopulationUSAMigraine with aura 4.0, 2.2 to 7.2 Migraine without aura 2.2, 1.2 to 4.0Not assessed
Breslau, 2000‡,§YesPopulationUSA3.5, 2.6 to 4.6New-onset migraine 2.8, 2.2 to 3.5 New-onset depression 2.4, 1.8 to 3.0
Swartz, 2000YesPopulationUSA3.1, 2.0 to 4.4New-onset migraine 0.68, .02 to 2.0
Breslau, 2003YesPopulationUSANot assessedNew-onset migraine 3.4, 1.4 to 8.7 New-onset depression 5.8, 2.7 to 12.3
Zwart, 2003††YesPopulationNorway2.7, 2.3 to 3.2Not assessed
McWilliams, 2004NoPopulationUSA2.8, 2.2 to 3.7Not assessed
Patel, 2004YesHMO participantsUSAStrict migraine 2.7, 2.2 to 3.3 Probable migraine 1.9, 1.5 to 2.4Not assessed

Cross-Sectional Studies.— Merikangas et al11 studied the association of psychiatric syndromes, including depression, and migraine headache in a prospective epidemiologic cohort of 27- and 28-year-olds in Zurich, Switzerland. In this study, 457 subjects (225 men and 232 women) were interviewed to assess presence of psychiatric syndromes and/or migraine headache. The criteria for classification of headache were based on the IHS criteria,18 but the diagnoses did not strictly adhere to the guidelines. Standardized psychiatric assessments were used to diagnose the psychiatric syndromes. Migraine was strongly associated with major depression (OR 2.2, 95% CI 1.1 to 4.8). This was the first study to demonstrate a strong association between migraine and major depression in an unselected sample.

Breslau et al17 conducted a population-based study of people 25 to 55 years of age with migraine or other severe headaches to examine the relationship between migraine and major depression. This study used Cox proportional hazards models to estimate the risk for the first occurrence of migraine associated with prior major depression and the risk for depression associated with prior migraine.19,20 Similar Cox proportional hazards models were used to estimate the risk for the first occurrence of severe headache associated with major depression and vice versa. In this study, migraine was strongly associated with major depression (Sex-adjusted OR 3.5, 95% CI 2.6 to 4.6). The association was higher in people with migraine with aura (Sex-adjusted OR 4.9, 95% CI 3.3 to 7.2) than in people with migraine without aura (Sex-adjusted OR 3.0, 95% CI 2.2 to 4.1). Interestingly, the association was also present for severe headache (Sex-adjusted OR 3.2, 95% CI 2.1 to 4.7).

In a prospective, population-based study, Swartz et al21 examined the relationship between specific psychiatric disorders and migraine. The study sample included 1,343 participants identified from the Baltimore, MD cohort of the Epidemiologic Catchment Area Study. In cross-sectional analyses, major depression was strongly associated with migraine (OR 3.1, 95% CI 2.0 to 4.8).

Zwart et al22 conducted a large, cross-sectional population-based study to examine the association between migraine, non-migrainous headache, and headache frequency. Data were derived from the Nord-Trondelag Health Study, conducted in Nord-Trondelag County, Norway. The study sample included more than 50,000 participants. Data were collected through questionnaires. Headache diagnoses were based on IHS criteria, and mental assessment was based on HADS, a 2-dimensional self-rating instrument for depression (HADS-D) and anxiety (HADS-A). In this study, the OR for depression was significantly higher in people with migraine and non-migrainous headache, compared to headache-free individuals (non-migrainous headache OR 2.2, 95% CI 2.0 to 2.5; migraine headache OR 1.9, 95% CI 1.6 to 2.3). In addition, there was a strong linear trend (P < .001) of higher prevalence OR of depression with increasing headache frequency.

A study reported by McWilliams et al23 used data from the Midlife Development in the United States Survey to investigate associations between 3 pain conditions (arthritis, migraine, and back pain) and 3 common psychiatric disorders (depression, panic attacks, and generalized anxiety disorder) in a large sample (n = 3,032) of adults aged 25 to 74 in the U.S. general population. In this population, 28.5% of subjects with migraine were considered clinically depressed, while only 12.3% of subjects without migraine fit the same criteria (OR 2.8, 95% CI 2.2 to 3.7).

Patel et al24 studied the prevalence of major depression in individuals with migraine, probable migraine (a subtype of migraine missing just one migraine feature), and controls. Participants were identified from members of a mixed model health maintenance organization (HMO) using computer-assisted telephone interviews. The overall prevalence of major depression was 28.1% for persons with migraine, 19.5% for those with probable migraine, and 10.3% for the control group. The prevalence of major depression was elevated in all migraine groups when compared with controls on both crude and adjusted (by age, sex, education) prevalence ratios.

Longitudinal Studies.— Several studies have examined the bidirectional associations of migraine and depression. Three of these studies found a positive association14,17,25 while one found no association.21 In 1994, Breslau et al14 examined the association between migraine and depression in a prospective sample of 1,007 adults between the ages of 21 and 30 years. Subjects were members of a large HMO in Southeast Michigan. Using the Cox proportional hazards model, the sex- and education-adjusted HR for new-onset migraine in subjects with major depression was 3.1 (95% CI 2.0 to 5.0). Conversely, the sex- and education-adjusted HR for new-onset major depression in subjects with migraine was 3.2 (95% CI 2.3 to 4.6).

In 2000, Breslau et al17 reported a HR of 2.4 (95% CI 1.8 to 3.0) for first onset of major depression associated with previous migraine, controlling for sex. The HR for the reverse association: onset of migraine associated with previous major depression was 2.8 (95% CI 2.2 to 3.5). This study also investigated the bidirectional relationship between severe headache and major depression. The HR was significant for severe headaches and first-onset major depression (3.6, 95% CI 2.4 to 5.3), but not for major depression and first occurrence of migraine (1.6, 95% CI 0.9 to 2.8).

Breslau et al25 reported similar results in a 2-year longitudinal population-based cohort from the Detroit metropolitan area. The results of this study showed that, over a 2-year period, having baseline depression increased the risk of incident migraine (OR 3.4, 95% CI 1.4 to 8.7), but not other severe headaches (OR 0.6, 95% CI 0.1 to 4.6). In addition, the risk of incident depression was significantly higher in those with baseline migraine (OR 5.8, 95% CI 2.7 to 12.3), but not in those with severe headaches (OR 2.7, 95% CI 0.9 to 8.1). These results, in combination with those reported by Breslau et al in 2000, suggest that the bidirectional relationship is specific to migraine, and not all severe headaches.

In contrast, Swartz et al21 reported no excess incidence of adult-onset migraine among people with pre-existing depression (Age- and sex-adjusted OR 0.68, 95% CI 0.02 to 2.0). This study did not assess the onset of depression among people with pre-existing migraine. In addition, this study did not obtain data about subjects' history of migraine at baseline, and it excluded all people with a history of unspecified headache (21% of the sample). Subjects with a history of unspecified headache were considered not at risk for the first onset of migraine during the follow-up period. This exclusion is likely to lead to underestimation of the relationship between depression and first-onset migraine. Other explanations for the inconsistency of results between these studies include: differences in the samples ages, length of recall period, assessment procedures, and the definition of people at risk for first-onset migraine.

Conclusions.— Migraine and depression are comorbid: in large scale population-based studies, persons with migraine are from 2.2 to 4.0 times more likely to have depression. Some studies14,17,25 support a bidirectional relationship between migraine and depression, with each disorder increasing the risk for the subsequent first onset of the other. In the studies that report a bidirectional relationship, the risk of first-onset migraine in people with pre-existing depression ranged from 2.8 to 3.5 in the studies we sampled. Conversely, the risk of first onset of depression in people with pre-existing migraine ranged from 2.4 to 5.8. While Swartz did not support the bidirectional relationship, we believe that methodologic limitations may account for the negative study. These findings have important implications for clinical practice. Patients with migraine or depression should be evaluated for the other disorder. In addition, for patients with co-existing conditions, treatment choices that might improve both conditions should be considered.

MIGRAINE AND ANXIETY DISORDERS

Anxiety disorders are also associated with migraine. This relationship has been observed in both clinic and community-based populations.11–13,23,26–36 Herein, we summarize 2 large scale, prospective community studies11,23 that have demonstrated a cross-sectional relationship between migraine and various anxiety disorders.

The study reported by McWilliams et al23 showed a link between migraine and depression (see above), as also between migraine and anxiety. In this study, 9.1% of subjects with migraine, compared with only 2.5% of people without migraine had comorbid generalized anxiety disorder (OR 3.9, 95% CI 2.5 to 6.0). The association remained significant, even after adjusting for demographic variables including other common pain conditions (arthritis and back pain).

Another study reported that the association between migraine and anxiety disorders was even stronger than that for the affective disorders.11 In this prospective study among young adults, general anxiety disorder (OR 5.3, 95% CI 1.8 to 15.8) and social phobia (OR 3.4, 95% CI 1.1 to 10.9) were the types of anxiety disorders that exhibited the greatest association with migraine. Although the phobias were generally associated with migraine, agoraphobia did not show this relationship. This may be explained by the rarity of the disease.

The relationship between migraine and panic disorder has also been reported,13,37–39 but the temporal relationship between the 2 disorders has not been thoroughly explored. One recent study40 examined the panic disorder-migraine association in a population-based survey in the Detroit, MI area. In this study, the OR of panic disorder associated with migraine, adjusted for history of major depression and sex, was 3.7 (95% CI 2.2 to 6.2). Similarly, the OR for panic disorder associated with severe headache, adjusted for major depression and sex, was 3.0 (95% CI 1.5 to 5.8). Using a Cox proportional hazards model, and controlling for history of major depression and sex, the investigators examined the temporal relationships between migraine and panic disorder, and also severe headache and panic disorder. The adjusted HR for the first onset of panic disorder among migraineurs was 3.6 (95% CI 2.2 to 5.8), and among severe headache patients was 5.8 (95% CI 2.7 to 12.3). The depression and sex-adjusted HR for first onset of migraine in people with prior panic disorder was 2.1 (95% CI 1.4 to 3.1). Similarly, the HR for the first occurrence of severe headache associated with prior panic disorder was 1.9 (95% CI 0.7, 4.8). The results of this study indicate that the comorbidity of panic disorder is not specific to migraine, but includes other headache types as well. The investigation of the temporal relationship between the disorders suggests that the influence appears to be in the headache-to-panic disorder direction, rather than the reverse.

Co-Occurrence of Depression and Anxiety in People With Migraine.— Two studies have reported on the co-occurrence of depression and anxiety in people with migraine.11,12 Merikangas et al11 examined this relationship in a prospective study by categorizing subjects into mutually exclusive categories of anxiety and depression. Based on this analysis, they found that there was a 2-fold increased risk of migraine among subjects with both major depression and an anxiety disorder. In contrast, subjects with depression only were 2 times less likely to have migraine. In a similar, but somewhat different analysis, Breslau et al12 found that the combined effects of anxiety disorder and migraine increased the odds for major depression far more than might be expected. While the OR for migraine and major depression was 2.7 (95% CI 1.1 to 6.3), the statistic for major depression in people with migraine and any anxiety disorder versus people with neither disorder was 22.8 (95% CI 12.7 to 41.2).

Merikangas et al11 also reported the results of a logistic regression analysis conducted to examine the diagnostic overlap among those psychiatric disorders most strongly associated with migraine. The model controlled for sex and risk group while assessing the effects of major depression, bipolar spectrum, general anxiety disorder, and social phobia. The best fitting model included only general anxiety disorder.

Conclusions.— Migraine and anxiety disorders are comorbid, and in some studies, the relationship is even stronger than that between migraine and depression. In the studies reported herein, the OR for migraine and generalized anxiety disorder ranged from 3.5 to 5.3. In addition, most people with depression also have anxiety disorders, but many people with anxiety disorders do not have depression. For this reason, it is important to screen for both depression and anxiety in persons with migraine. Migraine and other severe headaches are also comorbid with panic disorder, and the data suggest that the temporal relationship is in the headache-to-panic disorder direction, rather than the reverse.

The co-existing relationship between migraine, depression, and anxiety disorders can have important clinical implications, depending on the chronology of these conditions. Treatment of one condition could help prevent progression to one or both of the other two.

MIGRAINE AND BIPOLAR DISORDER

The relationship between migraine and bipolar disorder is less well studied than the inter-relationships between migraine and depression or migraine and anxiety disorders. The relationship between migraine and bipolar disorders has been studied in both clinic41–46 and community-based populations,11,26 but this review will emphasize the population-based studies.

Merikangas et al.11 reported that 8.8% of the 61 migraineurs, compared with only 3.3% of the 396 individuals without migraine had bipolar spectrum disorder (defined by a manic or hypomanic episode) (OR 2.9, 95% CI 1.1 to 8.8). Breslau26 reported similar results in their study of a Detroit sample of young adults. In this study, the OR for Bipolar I was 7.3 (95% CI 2.2 to 24.6) and Bipolar II was 5.2 (95% CI 1.4 to 19.9).

OVERALL CONCLUSIONS

Migraine is consistently associated with several psychiatric disorders, including depression, anxiety, and bipolar disorders. It is important to maintain diagnostic vigilance for comorbid conditions. When migraine and a comorbid psychiatric disorder are present, it is important to take both disorders into account in formulating a treatment plan.

Conflict of Interest:  None

Ancillary