Pharmacological Management of Mood and Anxiety Disorders in Headache Patients


  • James L. Griffith MD,

  • Maryam Razavi MD

  • From the Department of Psychiatry and Behavioral Sciences, The George Washington University Medical Center, Washington, DC.

Address all correspondence to Dr. James L. Griffith, Department of Psychiatry 8th Floor, 2150 Pennsylvania Avenue, N.W., Washington, DC 20037.


There is emerging evidence that treatment of comorbid mood and anxiety disorders can improve headache treatment outcome when implemented within a comprehensive program. Effective treatment for comorbid mood and anxiety disorders requires screening headache patients and accurately diagnosing specific psychiatric disorders when present. Specific dual-action antidepressant, anticonvulsant, and atypical antipsychotic medications can serve as dual agents that simultaneously treat both headaches and a mood or anxiety disorder. Serotonin reuptake inhibitors and most other antidepressant, anxiolytic, and mood-stabilizing medications are generally ineffective for headache prophylaxis. However, they can be safely added to a headache regimen for treatment of a comorbid psychiatric disorder. Treatment of comorbid psychiatric disorders in headache patients requires patient education about the psychiatric disorder, its treatment, possible side-effects, and expected benefits. Clinicians need to be sensitive to possible stigma that some patients fear from a psychiatric diagnosis or its treatment.


serotonin-reuptake inhibitor


Diagnostic and Statistical Manual of Mental Disorders-4th ed. Text Revision


generalized anxiety disorder


obsessive-compulsive disorder


posttraumatic stress disorder

Headaches, particularly migraine and chronic daily headaches, have high rates of comorbidity with psychiatric mood and anxiety disorders. There is emerging evidence that these mood or anxiety disorders play important contributory roles toward headache chronicity and refractoriness to treatment.1 There is also increasing evidence that pharmacological interventions targeting both headaches and comorbid depressive or anxiety disorders can lead to better headache treatment outcomes.2 Such interventions are implemented within a comprehensive program of treatment that integrates treatment of the psychiatric disorder with behavioral, educational, and pharmacological headache treatments.3 Evidence-based guidelines for treatment of comorbid mood and anxiety disorders are premature at present. Large-scale, longitudinal studies have not yet been conducted.1,2 However, provisional guidelines can be proposed that take into account the current state of knowledge:

  • 1Headache patients need to be screened for anxiety and depression due to high rates of comorbidity.1,2
  • 2When a patient is positively identified through a screening instrument, the psychiatric diagnosis should be confirmed according to valid and reliable diagnostic criteria.1
  • 3Treatment for a comorbid mood or anxiety disorder can select a medication that has established dual efficacy for both the psychiatric disorder and the patient's type of headache, in essence “killing two birds with one stone.”
  • 4Mood or anxiety disorder medications should be monitored for possible adverse effects upon headache frequency or severity. Conversely, headache medications should be monitored for possible adverse effects upon mood or anxiety symptoms.
  • 5Problems in treatment adherence can be exacerbated when headaches and mood or anxiety disorders are comorbid, emphasizing a need for patient education for both disorders.3
  • 6Potentially stigmatizing effects of a psychiatric diagnosis should be anticipated and managed with an appropriate clinical strategy.


Major depressive disorder is diagnosed when a patient has had for most of the preceding 2 weeks at least 5 of the following 9 symptoms: depressed mood (eg, anguish, sadness, emptiness, tearfulness), anhedonia (diminished capacity for pleasure), significant weight loss or appetite change, insomnia or hypersomnia, psychomotor agitation or retardation, fatigue or loss of energy, feelings of worthlessness or guilt, difficulties concentrating, or recurrent thoughts of death or suicide.4,5

For nonheadache depressed patients, a broad range of antidepressant medications produce symptom remission in 60% to 70% of cases.6 A patient's past history of preferential response to a particular class of medications, or a family history of preferential response, gives guidance to selecting an antidepressant.6,7 However, evidence-based criteria are otherwise still lacking for selecting one medication versus another based upon treatment efficacy, with the exception of atypical depression, which is specifically responsive to monoamine oxidase inhibitors.6–8

For headache patients, a preferable antidepressant is one that has shown efficacy in treating both headaches and depression. “Dual-action antidepressants” that combine potent serotonin reuptake inhibition with norepinephrine reuptake inhibition are candidates for this role.9,10 Amitriptyline, some other tricyclic antidepressants, and venlafaxine are dual-action antidepressants whose efficacy in treating both headaches and depression are strongly supported by clinical evidence. For migraine prophylaxis, Loder and Biondi reviewed the evidence for efficacy of available medications according to US Headache Consortium criteria for evidence-based treatment.11 Among antidepressants, amitriptyline had the most evidence supporting efficacy (Level A), while other tricyclic antidepressants, including nortriptyline, protriptyline, and doxepin, had less extensive evidence (Level C). However, treatment of major depressive disorder typically requires daily doses of 150 mg or higher for amitriptyline and equivalent doses for other tricyclics.12 These doses are substantially greater than those often needed for headache or chronic pain control. The main limitations of amitriptyline and other tricyclic antidepressants are their side-effects—sedation, dry mouth, constipation, blurry vision, urinary retention, weight gain, EKG changes from slowed intracardiac conduction.12 For suicidal patients, doses only 3 to 5 times the therapeutic dose of a tricyclic antidepressant can be fatal.7,12

Venlafaxine is a nontricyclic antidepressant that recently has met Level A criteria with the support of both open label and randomized controlled treatment trials.13,14 Venlafaxine is a more potent reuptake inhibitor for serotonin than for norepinephrine. For this reason, it functions as an SSRI at low dosages, becoming a dual-action antidepressant at progressively higher dosages above 100 mg daily.10 Although limited by study designs and small sample sizes, published studies suggest that at least 75 to 150 mg daily is needed for migraine treatment.11,12 Venlafaxine usually produces few side-effects, including little effect upon body weight. Some patients experience gastrointestinal symptoms or sexual dysfunction, and mild hypertension occurs in 5% of patients, primarily with regimens above 300 mg daily.7,8

Mirtazapine is a dual-action antidepressant with a complex mechanism in which enhancement of serotonergic and noradrenergic neurotransmission is accompanied by blockade of post-synaptic serotonergic 5-HT-2A and histaminic receptors. Case reports have reported mirtazapine-induced improvement in migraine patients.15 Duloxetine is a new dual-action antidepressant with an FDA indication for relieving diabetic neuropathy pain, but thus far has not been studied systematically for possible efficacy in treating migraine.16

Among other antidepressants, nefazodone, an antidepressant that blocks serotonergic 5-HT-2A receptors, showed efficacy for chronic daily headache prophylaxis in an open-label trial (Level B).17 Other than sedation, nefazodone side-effects are typically uncommon, although rare occurrences of liver toxicity have been reported.7 Unlike many antidepressants, nefazodone is usually free from weight gain or sexual side effects. The monoamine oxidase inhibitors, phenelzine and tranylcypromine, have efficacy similar to that of tricyclic antidepressants (Level C), but are now reserved only for patients who have initially failed to respond to other available antidepressant medications, due to insomnia, weight gain, sexual side effects, dietary restrictions, and potentially lethal side effects of hypertensive crisis or central serotonergic syndrome.8,11

While serotonin reuptake inhibitors (SSRIs), including fluoxetine, citalopram, sertraline, paroxetine, and escitalopram, are the most commonly prescribed antidepressants for depressed psychiatric outpatients, they have not been found to be consistently effective for headache treatment. For example, a randomized trial of chronic tension-type headaches found amitriptyline to significantly reduce headache frequency and duration, but not the SSRI citalopram or placebo.18 Fluoxetine similarly failed to show robust effectiveness for migraine in a double-blinded study, although some benefits for depressed patients with chronic daily headaches.19 A randomized study of sertraline failed to show benefits over placebo in a migraine study that excluded co-morbid depression. The authors' accompanying review of other SSRIs concluded that this class of medications is ineffective for migraine prophylaxis.20

Supplementing a categorical DSM-IV-TR depression diagnosis with a quantitative assessment of depression severity best discriminates patients likely to respond to antidepressants. Demoralization and grief in medically ill patients commonly produce sleep, energy, and concentration disturbances that overlap with symptoms of major depressive disorder, but show little response to antidepressant medications. A Hamilton Depression Scale score of 17 or higher, or equivalent score on another well-validated mood scale, predicts a high likelihood of response to medication. Use of these quantitative scales also aids accuracy of assessment of treatment response after a medication is prescribed. Partial response is defined as greater than 25% but less than 50% decrease in depression assessment scale scores; treatment response is a 50% or greater reduction in scores with a final Hamilton Depression Scale score less than 15; remission is the absence of depression symptoms or minimal residual symptoms, such as a Hamilton Depression Scale score of 7 or less.6,21 Remission is the goal of depression treatment.21,22


Bipolar disorder is diagnosed when periods of elevated, expansive, or irritable mood, at least a week in duration, occur in addition to other periods of depression. A diagnosis of bipolar disorder also requires significant presence of 3 or more associated symptoms of grandiosity, decreased need for sleep, talkativeness, racing thoughts, distractibility, heightened goal-directed activity, and excessive involvement in pleasurable activities that have a high potential for harmful consequences.4,5 Mania is diagnosed when impairment of work, social, or relational functioning is marked or psychotic symptoms occur, while hypomania refers to a lesser severity. The simultaneous occurrence of depression and mania is referred to as a mixed state. Bipolar I disorder is diagnosed when there are episodes of mania; bipolar II disorder is diagnosed when there is only hypomania alternating with depressive episodes.4,23 For nonheadache patients, treatment of mania or mixed episodes is initiated with lithium carbonate in combination with an antipsychotic medication, or valproate in combination with an antipsychotic. For less ill patients, monotherapy with lithium, valproate, or an atypical antipsychotic such as olanzapine may be sufficient. In mixed states, valproate is preferred over lithium.23,24

For headache patients with bipolar disorder, a mood-stabilizing medication can be selected that also has efficacy for headaches. Valproate has been shown to reduce migraine frequency by up to 27% from baseline in multiple drug trials.11 Other atypical antipsychotic medications have established efficacy in treating bipolar disorder, as well as accruing evidence for efficacy in treating migraine.25 Silberstein et al found olanzapine treatment of refractory migraine or chronic daily headache patients to produce a significant reduction in pain intensity and an increase in headache-free days.26 One small open trial has reported marked effectiveness of olanzapine in abortive treatment of cluster headaches.27 Lithium is ineffective for migraine, but has held a longstanding role in treatment of cluster headaches.28

Although the efficacy of topiramate for migraine treatment is strongly supported by research studies, it has been found ineffective for treatment of bipolar disorder.29 Likewise, gabapentin has shown some efficacy for migraine treatment in an open trial, but it has been ineffective for bipolar disorder.30,31 Carbamazepine and lamotrigine are anticonvulsants whose efficacy for bipolar disorder treatment is well supported by randomized controlled studies, with lamotrigine particularly effective for bipolar disorder with depression.24 However, neither medication has been shown to be effective for headache treatment. It should be noted that antidepressant treatment of depression in a patient with bipolar disorder poses a risk for precipitating mania. Patients with bipolar disorder should be treated with a mood-stabilizing medication at all times, whether or not an antidepressant is added during periods of depression.24


Generalized anxiety disorder (GAD) is characterized by persistent worry about events of daily life together with heightened physiological arousal. The duration, frequency, and intensity of anxiety are out of proportion to the level of threat.5,32

Venlafaxine has received an FDA indication for long-term treatment of GAD.32 As such, venlafaxine can be recommended for initial treatment of GAD co-morbid with headaches. Gabapentin has efficacy for migraine prophylaxis and can be utilized as an adjunctive agent for GAD.30,31 Buspirone is a nonsedating partial serotonin agonist that is FDA approved for GAD treatment and was recently found effective for migraine prophylaxis in a double-blind, randomized study.33 The anti-anxiety effects of buspirone are often modest and 1 to 4 weeks may be required for benefits to be seen. Since it is not cross-tolerant with alcohol, buspirone can be utilized safely with alcohol-dependent patients for whom benzodiazepines would not be appropriate.34

For nonheadache patients, serotonin-reuptake inhibitors, including fluoxetine, paroxetine, citalopram, escitalopram, and fluvoxamine, are most commonly used for initial treatment of GAD. Benzodiazepines, including diazepam, chlordiazepoxide, lorazepam, alprazolam, and clonazepam, are also highly effective. Because their antianxiety effects typically occur within 1 to 2 hours after ingestion, benzodiazepines can be used for the initial treatment of GAD, then tapered after venlafaxine or another antidepressant induces remission of symptoms after 2 to 6 weeks of treatment.32


Panic disorder is characterized by recurrent panic attacks. Panic attacks are sudden-onset episodes of intense fear of dying or losing control, together with such physical symptoms as hunger for air, palpitations, sweating, trembling, dizziness, and hot and cold flashes. Agoraphobia, as fear of being trapped in a place from which escape might be difficult, frequently accompanies panic disorder.4,5 Both panic disorder and posttraumatic stress disorder (PTSD) patients have high rates of medical health care utilization due to anxiety symptoms misinterpreted as symptoms of physical disease.32

There are multiple medication regimens that can effectively treat panic disorder. For headache patients with panic disorder, venlafaxine, tricyclic antidepressants, and monoamine oxidase inhibitors can be effective for both disorders.11,32 For non-headache patients, serotonin-reuptake inhibitors are generally regarded as holding the most favorable balance of efficacy versus adverse effects. Alternatively, benzodiazepines are used when there is a need for rapid control of symptoms.32

Bupropion and trazodone are antidepressants that are effective for depression but ineffective for panic disorder.32 Open-label studies have suggested that gabapentin may have a role as an add-on medication, but not as primary treatment for panic disorder.31 While buspirone is effective for generalized anxiety, it is ineffective for panic disorder.32


Obsessive-compulsive disorder (OCD) is a chronic psychiatric disorder characterized by a range of obsessive thoughts and compulsive behaviors. Obsessions are recurrent or persistent thoughts, images, or impulses that a patient considers inappropriate but cannot control with will power (eg, aggressive or sexual thoughts, pathological doubt). Compulsions are repetitive acts felt to have an irresistible urgency (eg, cleaning, checking, counting).4,5

Pharmacological treatment of OCD requires high doses of antidepressants that potently inhibit serotonin reuptake. Effective doses are typically 3 to 4 times the level required for treatment of depression. A longer treatment latency before symptom remission, often 9 to 12 weeks, is common. Symptom remission is most often partial, and relapse typically occurs if medications are discontinued.32

Clomipramine is the only tricyclic antidepressant that is also a serotonin-reuptake inhibitor. As a dual-action antidepressant, clomipramine may hold advantages in migraine treatment over other SSRIs when headaches are comorbid with OCD. The disadvantage of clomipramine is the frequency of side-effects including dry mouth, weight gain, postural hypotension, sedation, and lowering of seizure threshold. In nonheadache psychiatric populations, SSRI antidepressants, such as fluoxetine 60 to 80 mg daily or sertraline 200 to 250 mg daily, are most commonly employed for OCD treatment.12,32


Patients with PTSD show chronic reexperiencing symptoms (nightmares, flashbacks), avoidance symptoms (emotional numbing, behavioral avoidance of reminders of trauma), and hyperarousal symptoms (insomnia, irritability, excessive startle reflex) after a traumatic life event.4,5 Rates of PTSD in headache populations have been little studied. However, clinicians treating PTSD commonly find headaches to be persistent among patients who have largely recovered from other posttraumatic symptoms. Since PTSD has high rates of comorbidity with major depressive disorder and other anxiety disorders,32 it is not yet clear whether such headaches are primarily related to the PTSD or another associated disorder.

SSRI antidepressants are generally regarded as first-line treatments for PTSD. However, venlafaxine, nefazodone, and mirtazapine may hold advantages over SSRIs as initial treatment for PTSD comorbid with headaches due to their efficacy for migraine prophylaxis.32 Atypical antipsychotics, including risperidone, olanzapine, and quetiapine, have found increasing support in research studies as agents that improve sleep quality and augment the potency of antidepressants in treating PTSD.35 Since olanzapine has been shown in open trials to be effective in migraine treatment, it may fill the role of a preferred second-line agent added to venlafaxine or an SSRI.26


Successful psychopharmacological treatment of comorbid mood or anxiety disorders in headache patients can require special attention to issues impacting treatment adherence. The likelihood of success is enhanced when clinician and patient share expectations regarding: (1) criteria for gauging success or failure of the treatment, including expected time course of improvement; (2) medication side effects to be anticipated, including a plan for their management; and (3) the next step to take if the initial treatment fails.

Failure to experience a rapid treatment response and untoward side effects are common reasons why patients discontinue medications. When treating both headaches and comorbid psychiatric symptoms, a special concern is the possible exacerbation of headaches by the psychiatric medications or exacerbation of mood or anxiety symptoms by the headache regimen. Propranolol can aggravate depression, and topiramate can produce anxiety, depression, or agitation in some vulnerable patients.6,29 Divalproex and olanzapine can produce marked weight gain or sedation that further demoralize an already depressed patient.24 Tricyclic antidepressants, monoamine oxidase inhibitors, and lithium carbonate also can produce weight gain.6,12,24 While manufacturers' data for venlafaxine, fluoxetine, paroxetine, and sertraline reported no greater frequency of headache side-effects over placebo, clinicians have commonly noted headache exacerbations from each of these antidepressants, with remission when the drug was discontinued.36,37 Approximately 6% of patients taking buspirone have headaches as side-effects.38

Stigma from a psychiatric diagnosis is a different reason why some patients refuse pharmacological treatment for depression or anxiety. Some patients with refractory headaches experience recommendations for depression or anxiety treatment as diverting attention from “the real problem” or feel abandoned by their headache specialist who makes a psychiatric referral. An important component of pharmacological treatment is to inquire about the meaning a patient attributes to a mood or anxiety disorder diagnosis. Patient education should address concerns about diagnosis, treatment, treatment side-effects, and prognosis. However, issues of stigma are more complex when a patient fears belittling responses from family or friends, hence not as amenable to simple educational measures. Some clinicians attempt to avoid stigma by labeling the psychiatric medication as a treatment given solely for the headaches or for another specific symptom, such as insomnia. However, this strategy risks a patient later feeling deceived, and it deprives the clinician and patient from an authentic collaboration in treatment. A more collaborative approach is for the clinician to inquire whether the patient has concerns about taking a medication for depression or anxiety (eg, “Are you comfortable taking an antidepressant?”) and to learn whether there are fears of deprecatory responses from others who find out about the medication. The clinician can then adopt a proactive stance by helping the patient strategize how he or she might respond to queries—what information will be shared with whom at home or at work and how information will be presented.


The current state of research leaves important clinical questions with less than satisfactory answers. Treatment recommendations for comorbid mood and anxiety symptoms often must be gleaned from a handful of open trials and case reports with headache patients, sometimes supplemented by a single randomized controlled trial. By comparison, 16 controlled studies have supported the efficacy of carbamazepine in psychiatry for treating mania, giving clinicians some sense of surety for its role, even though the drug never received FDA approval for that purpose.39 The need for further research is great. For clinicians treating migraine, some salient research questions also suggest what may be prudent guidelines for clinical practice given the state of current evidence:

(1) Does effective treatment of comorbid mood and anxiety disorders also improve the clinical course of headaches? Unless disconfirming data emerge, it is safest for our patients to assume there is benefit in treating comorbid mood and anxiety disorders to be best of our capabilities. The current standard of practice is to seek symptom remission, not symptom reduction, for mood and anxiety disorders.21,22

(2) How do antidepressant medications produce improvement for migraine headaches? Do they alter fundamental pathophysiological mechanisms? Or does headache improvement reflect only the general augmentation of brain antinociceptive systems? Until research provides a finer discrimination, clinical wisdom suggests relying on medications with established track records as broad-spectrum agents that can: (1) effectively treat both mood and anxiety symptoms; (2) provide effective migraine prophylaxis; and (3) provide general anticoceptive effects for chronic pain.40 This suggests reliance upon dual-action antidepressants for comorbid depression and anxiety disorders and mood stabilizers such as divalproex and atypical antipsychotics as first-line agents for comorbid bipolar disorder.

(3) For medications with dual efficacy for both migraine and comorbid psychiatric disorders, are effective doses the same or different for the 2 disorders? Do migraine patients have special side-effect vulnerabilities to psychiatric medications? Until well-designed studies target these questions, a conservative clinical strategy is to utilize dosing guidelines extrapolated from the psychiatric clinical literature, as summarized in Table, while monitoring patients closely for unexpected side-effects or medication sensitivities.

Table Table.—.  Dosing Guidelines for Some Mood and Anxiety Medications6
MedicationStarting Dose (mg per day)Increments (mg)Full Dose (mg per day)
 Venlafaxine37.537.5 increments every 3 to 5 days150 to 225
(Slow-release Effexor-XR)
 Nefazodone50 twice daily100 weekly300 to 500
 Amitriptyline25 (at night)25 every 3 days150 to 300
Imipramine, or Doxepin
 Clomipramine25 (at night)25 every 3 days150 to 250
 Mirtazapine15 to 30 (at night)15 every 3 days15 to 45
 Citalopram1010 every week20 to 40
 Escitalopram1010 every week10 to 20
 Paroxetine1010 every week20 to 50
 Sertraline5050 every week50 to 200
 Phenelzine1515 every 2 days90 (or until postural hypotension occurs)
Mood stabilizers
 Lithium carbonate300 twice daily300 mg after 1 to 2 days900 to 1500 (or until 0.7 to 1.2 mEq/l serum level for acute treatment)
Mood stabilizers (atypical antipsychotics)
 Olanzapine2.5 to 5.0 at night2.5 mg daily5 to 20 mg at night
 Risperidone0.5 to 1.0 daily0.5 mg daily2 to 3 mg twice daily
 Quetiapine25 twice daily25 mg daily100 to 600 mg, with 2/3 dose at night
 Ziprasidone20 mg twice daily40 mg every 2 days120 to 160 daily
Mood stabilizers (anticonvulsant)
 Divalproex15 per kg daily 15 to 60 per kg daily
 Gabapentin100 to 300100 to 300 weekly900 to 3600 (in 2 to 3 divided doses)
Other anxiolytics
 Buspirone5 twice daily5 every 3 days30 to 60 (in 2 to 3 divided doses)

(4) To what extent is antidepressant-induced improvement in headaches a function of mood effects upon cognition (ie, less depressed or anxious people perceive and experience pain differently than when symptomatic), rather than direct effects upon the headaches? For example, a study with nefazodone found that, as the depression improved, depressed patients showed a greater differential improvement in headache pain than non-depressed patients.17 Similarly, chronic daily headache patients treated with fluoxetine showed improvements in headache pain that paralleled their improvement in depression, even though fluoxetine failed as a prophylactic medication for migraine.19 Our fascination with neurobiological mechanisms should not distract us from the larger truth that treating migraine patients often is more about helping people bear suffering and less about removing headaches. When a patient is no longer depressed or anxious, suffering that had been unbearable can become bearable. This argues for assertive treatment of comorbid psychiatric disorders, using both medications and evidence-based psychotherapies.3

Conflict of Interest:  None