Migraine is characterized by episodes of headache with qualities such as unilateral location, throbbing pain and aggravation by routine physical activity. Additional symptoms include nausea, photophobia and phonophobia. Some patients have aura symptoms, usually visual, before the headache phase (Davidoff, 1995). Prodromal and accompanying symptoms of migraine attacks often are psychiatric in nature, such as depression, elation, irritability, anxiety, overactivity, difficulty thinking, anorexia or increased appetite. In some patients, an organic mental syndrome can be part of a migraine attack (Davidoff, 1995). In other patients, an acute psychotic condition is the dominating clinical feature. This presentation, with paranoid delusions, hallucinations and anxiety, has been described in families with hemiplegic migraine (Spranger et al., 1999). Migraine is, therefore, an important differential diagnosis in relation to episodic phenomena with a mixture of somatic and psychiatric symptoms. In addition, psychosocial stress is the most common precipitating factor for a migraine attack (Davidoff, 1995).

Migraine is an organic disorder with a clear genetic background, even if environmental factors also are important both etiologically and in the precipitation of individual attacks. It is now fairly well-established that migraine has a primary neuronal basis, although blood vessels also are involved in the series of events that constitute a migraine attack (Aurora and Welch, 2000).

The prevalence is usually between 10% and 15% in epidemiological studies, and migraine is more common in women than in men (Silberstein and Lipton, 1993). Neurochemical disturbances are thought primarily to involve the serotonergic (Silberstein, 1994) and the dopaminergic systems (Hargreaves and Shepheard, 1999). Drugs acting on serotonergic neurons or receptors may induce migraine headaches, and migraine patients are more sensitive than others to dopaminergic stimulation. In familial hemiplegic migraine, dysfunctional neuronal calcium channels have been found (Hargreaves and Shepheard, 1999).

Comorbidity of Migraine and Affective Disorders

A total of 102 patients, 79% of them inpatients between 18 and 65 years old, with major affective disorders were interviewed in two studies (Fasmer, 2001; Fasmer and Oedegaard, 2002). In the first study, we interviewed 62 consecutively admitted patients with major affective disorders and examined the frequency of migraine in patients with unipolar and bipolar disorders (BD) (Fasmer, 2001). In the second study, we recruited an additional 40 patients; and in the entire group of patients (n=102), we looked more closely at the clinical characteristics of the patients with migraine compared to those without migraine (Fasmer and Oedegaard, 2001). We used a clinical interview based on criteria from the DSM-IV, supplemented with Akiskal's criteria for affective temperaments (Akiskal and Akiskal, 1992; Akiskal and Mallya, 1987). Bipolar I disorder (BDI) was diagnosed according to DSM-IV, while bipolar II disorder (BDII) encompassed patients with either discrete hypomanic episodes or an affective temperament (cyclothymic or hyperthymic), in addition to major depressive episodes. We employed the criteria of the International Headache Society (1988) to diagnose migraine.

In both studies, we found migraine to be a common comorbid disorder in patients with unipolar depressive disorder or BD, affecting approximately half of the patients in each group. However, most of the patients we interviewed did not present migraine headaches as a prominent complaint, and often a history of migraine was not noted in the hospital records. The most interesting finding was a substantial difference between patients with BDI and BDII, with migraine being clearly more prevalent in the BDII than in the BDI group. In our second study, 82% of the patients with BDII had migraine, compared to 27% of the patients with BDI (Figure). There is much evidence, including our own, indicating that patients with BDI and BDII represent two different nosological conditions (Coryell, 1996). Our results are similar to those of Endicott (1989), who found, among patients with major affective disorders, the highest frequency of migraine (51%) in patients having characteristics similar to patients with BDII as defined in the present study.

The most noteworthy findings concerning the clinical characteristics were that patients with migraine had a higher frequency of affective temperaments (47% versus 22% in patients without migraine) and a higher number of anxiety disorders. They were more likely to have panic disorder (51% versus 24%) and agoraphobia (58% versus 27%) than the patients without migraine. Symptoms during depressive episodes were similar, except that the migraine patients reported irritability and suspiciousness with increased frequency.

In two epidemiological studies, one from Zurich, Switzerland, (Merikangas et al., 1990) and one from Detroit (Breslau and Davis, 1992), a clear relationship between migraine and major affective disorders has been found (Breslau et al., 1994). In the Zurich study, people with migraine had a threefold-increased one-year prevalence of bipolar spectrum disorders (9% versus 3%), a nonsignificant increase in manic episodes and a twofold-increased prevalence of major depression (15% versus 7%).

Although these results cannot be directly compared to ours, they show that the association of migraine and affective disorders is not only found in such a selected group as we have studied. In these epidemiological studies, people with migraine also had an increased frequency of anxiety disorders. In the study by Breslau and Davis (1992), the frequency was doubled, compared to people without migraine, and the association was especially strong for panic disorder, with a sixfold increase. In contrast to these findings in patients with affective disorders, a study of patients with schizophrenia found no increased frequency of migraine (Kuritzky et al., 1999).

In our second study, the age of onset of the first anxiety disorder (most often a specific phobia) for patients with migraine was 15 years of age. This was earlier than the onset of migraine (21 years), which again was earlier than the onset of the first depressive episode (26 years). The first hypomanic episode occurred at age 28 (Figure). These chronological relationships are in agreement with previous studies. The high prevalence of anxiety disorders in patients with major affective disorders and comorbid migraine supports the hypothesis that there is a syndromal relationship between migraine, anxiety and depression (Merikangas et al., 1990). We would add that bipolar features should be included as part of this syndrome, and possibly the presence of migraine may be used to delineate a distinct subgroup of the major affective disorders.

Treatment Considerations for Both Disorders

To our knowledge, there are no studies that have specifically examined responses to drug treatment in patients with major affective disorders and comorbid migraine. Guidelines for pharmacological treatment must, therefore, be based on data from the neurological literature combined with data from the treatment of major depressive disorder, BDII and panic disorder.

Concerning antidepressants, amitriptyline (Elavil, Endep) is the drug that has been best studied in the prophylactic treatment of migraine and has been shown to reduce the frequency of attacks by 40%. This effect seems to be unrelated to its effect on depression (Ramadan et al., 1997). Selective serotonin reuptake inhibitors are less effective than either amitriptyline or propanolol (Inderal) (Silberstein, 1998).

In open studies, lithium has been shown to be useful in some patients with migraine (Medina and Diamond, 1981), however, others have reported worsening of migraine with lithium (Peatfield and Rose, 1981). Carbamazepine (Tegretol) does not seem to have any effect in patients with migraine (Post and Silberstein, 1994).

Several studies, both open and controlled, have shown that valproate (Depakene) has prophylactic effect in migraine, reducing the number of attacks, duration of headache and intensity of pain (Silberstein, 1996). Valproate thus has effect on the three main symptom groups in patients with migraine and comorbid affective disorders: headaches, mood instability and panic attacks (Freeman et al., 2002).

In the acute treatment of migraine, triptans, which exert their effect by a combination of vasoconstriction and decreased release of inflammatory mediators (Blier and Bergeron, 1995), are usually employed. The oldest and best-studied is sumatriptan (Imitrex). Although sumatriptan apparently has limited ability to penetrate the blood-brain barrier (Millson et al., 2000), it has been implicated in adverse events resembling the serotonin syndrome, when combined with centrally acting serotonergic drugs. However, the number of reported cases is small, and most patients seem to tolerate this combination without problems (Gardner and Lynd, 1998).

It is theoretically possible that the risk of depressive illness may be increased by the use of triptans, especially the newer ones which have enhanced lipophilicity, but this could not be confirmed in a recent study of consulting rates in general practice (Millson et al., 2000).

Acknowledgement

This research has been supported financially by the legacy of Gerda Meyer Nyquist Gulbrandson and Gerdt Meyer Nyquist.

Dr. Fasmer is associate professor of medicine in the department of psychiatry at University of Bergen and senior consultant at Haukeland Hospital, Bergen, Norway.

Dr. Oedegaard is research fellow in the department of psychiatry at University of Bergen and senior consultant at Haukeland Hospital, Bergen.