SEARCH

SEARCH BY CITATION

Keywords:

  • headache;
  • psychiatric comorbidity;
  • depression;
  • anxiety;
  • psychopathology;
  • methodology

Abstract

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

The comorbidity of headache and psychiatric disorders is a well-recognized clinical phenomenon warranting further systematic research. Affective disorders occur with at least three-fold greater frequency among migraineurs than among the general population, and the prevalence increases in clinical populations, especially with chronic daily headache. When present, psychiatric comorbidity complicates headache management and portends a poorer prognosis for headache treatment. However, the relationship between headache and psychopathology has historically been misunderstood, and measures of psychopathology have not always met the standard of formal Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV) criteria. In some cases, headache has been inappropriately attributed to psychological or psychiatric features, based on anecdotal observations. The challenge for future studies is to employ research methods and designs that accurately identify and classify the subset of headache patients with psychiatric disorders, evaluate their impact on headache symptoms and treatment, and identify optimal behavioral and pharmacologic treatment strategies. This article offers methodological considerations and recommendations for future research including: (i) ascribing dual-International Classification of Headache Disorders, 2nd ed. (ICHD-2) headache and DSM-IV psychiatric diagnoses according to reliable and valid diagnostic criteria, (ii) differentiating subclinical levels of depression and anxiety from major psychiatric disorders, (iii) encouraging validation studies of the recently published ICHD-2 diagnoses for “headache attributed to psychiatric disorder,” (iv) expanding epidemiological research to address the range of DSM-IV Axis I and II psychiatric diagnoses among various headache populations, (v) identifying relevant psychiatric and behavioral mediator/moderator variables, and (vi) developing empirically based screening and treatment algorithms.

Abbreviations: 
(ICHD-2)

International Classification of Headache Disorders, 2nd ed.

(DSM-IV)

Diagnostic and Statistical Manual of Mental Disorders, 4th ed.

(BDI)

Beck Depression Inventory

(MMPI)

Minnesota Multiphasic Personality Inventory

(SCID-I)

Structured Clinical Interview for DSM-IV Axis I Disorders

(SCID-II)

Structured Clinical Interview for DSM-IV Personality Disorders

Headache has been associated with psychiatric illness in the medical literature for well over a century. Unfortunately, as noted by Silberstein and colleagues,1 the relationship between headache and psychopathology has been clinically discussed far more often than it has been systematically studied. This relationship remains probably one of the most poorly understood, while at the same time clinically important, areas for future headache research. The present article provides a brief historical context for research examining headache and psychiatric comorbidity. Despite an arguably dubious genesis emerging from psychoanalytic case reports with little evidence supporting their generalizability to the wide range of those with headache, empirical research has begun to emerge in recent years. As discussed below, psychiatric comorbidity impacts headache symptoms and management. Epidemiologic and clinical investigations are confirming associations between headache and select psychiatric disorders, and revealing implications for clinical practice. Methodological considerations within this literature are discussed. Recommendations are presented for future research.

HISTORICAL CONTEXT

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

Although some of the earliest scholarly works on the topic of headache acknowledged the robust influence of psychological and behavioral factors in these disorders, it was not until the end of the 19th century that Freud categorically associated concepts of psychopathology with commonplace migraine.2,3 While others working within the emerging field of psychosomatic medicine carried forward Freud's notions regarding psychopathology and headache,4,5 it was not until the late 1930s that Harold Wolff first applied careful, systematic observation to the study of these associations in those with migraine who were not seeking treatment for psychiatric illness.6 For more than 25 years, Wolff vigorously pursued efforts to integrate knowledge from the social and medical sciences that would lead to better understanding of headaches and other “psychophysiological” disorders—efforts that foreshadowed contemporary behavioral medicine.7–10 Along the way, Wolff has been credited with developing the influential notion of “the migraine personality” that he characterized as a medley of “personality features and reactions dominant in individuals with migraine” including “feelings of insecurity with tension manifested as inflexibility, conscientiousness, meticulousness, perfectionism, and resentment” (p. 348).8

Since Wolff's era, hundreds of additional published empirical studies have systematically examined interrelationships between psychological factors and recurrent headache.11 Wolff's notions regarding the generalizability of the migraine personality have not withstood the test of time; in fact, current research suggests that there is no one dominant personality profile among those with migraine. Similarly, psychodynamic conceptualizations of migraine in psychopathological terms once popularized within psychosomatic medicine are now largely relegated to a historical footnote in the medical literature. Nevertheless, from these roots a biopsychosocial conceptualization of headache has emerged with few headache clinicians or researchers prepared to dispute the significance of psychological, behavioral, and social factors in precipitating and exacerbating recurrent headache disorders.

PSYCHIATRIC DISORDERS AND HEADACHE

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

An extensive review of the empirical literature examining more than 300 studies of psychological factors and headache found not only little empirical support for earlier personality stereotypes, but also little evidence of significant psychopathology in those with “average” headache.11 The review examined 190 empirical studies comparing psychological test data of those with headache with normative controls and revealed that although headache patients often differ from “normal” controls, they generally do so in “less pathological” ways (such as amount of daily stress experienced or the predominant stress coping strategies employed). Clinically meaningful psychopathology is observed among only a minority of all those with headache, including those in the community at large, and those who volunteer for research studies.11,12

Nevertheless, numerous epidemiological and community studies have revealed that psychiatric disorders (eg, depression and anxiety) occur with greater frequency among recurrent headache patients than among the general population, and the prevalence of psychopathology increases and is over represented in clinical populations.13 There is evidence indicating patients with elevated psychological symptoms are more likely to seek medical assistance.14 When present, psychiatric comorbidity often complicates headache management and portends a poorer prognosis for headache treatment.15–17

Psychiatric comorbidity is in fact relatively commonplace among headache patients who present for treatment (especially those presenting for specialist care) and an important consideration in headache treatment planning.17 Patients with migraine and tension-type headache exhibit psychiatric illnesses at a disproportionately higher rate than individuals with no history of recurrent headache. These comorbid relationships have been identified in epidemiological research12,13,18–29 as well as clinical studies of treatment-seeking patients.13,15,30,31 Affective disorders occur with at least three-fold greater frequency among migraineurs than among the general population, and the prevalence increases in clinical populations, especially with chronic daily headache.

There are significant gender differences in comorbidity. Women are significantly more likely than men to receive lifetime diagnoses of both migraine (24% vs 9%) and major depression (24% vs 13%) by age 30, with relative female risk increasing for migraine in late adolescence and for major depression after about age 20. Longitudinal data indicate that relative to men, women are four-fold more likely to develop migraine and two-fold more likely to develop major depression.12,25

The comorbidity of headache and psychiatric disorders is now a well-recognized clinical phenomenon warranting further systematic research. However, the relationship between headache and psychopathology has historically been misunderstood. In some cases headache has been inappropriately attributed to psychological or psychiatric features, based on anecdotal observations. Measures of psychopathology have not always been met the criteria promulgated in the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV),32 which have evolved over time. In addition, the diagnosis of headache has not always followed standard criteria, complicated by the emergence of new terms and criteria. Examples include the first publication of International Headache Society's classification criteria in 1988, proposed new criteria for “chronic daily headache” published by Silberstein et al in 1994, and the new diagnosis of “chronic migraine” in the 2004 revision of the International Classification of Headache Disorders, 2nd ed. (ICHD-2).33 The challenge for future studies is to employ research methods and designs that accurately identify and classify the subset of headache patients with psychiatric disorders, evaluate their impact on headache symptoms and treatment, and identify optimal behavioral and pharmacologic treatment strategies. This article offers methodological considerations and recommendations for future research.

DSM-IV codes psychiatric disorders along a multiaxial coding system.32 Axis I includes diagnoses of clinical disorders and other conditions that may be a focus of clinical attention, while Axis II includes diagnoses for personality disorders and mental retardation. The following sections represent selected Axis I and II psychiatric disorders that have at least some empirical association to headache, and demonstrate that the prevalence and impact of these disorders among headache patients render them worthy of careful consideration when designing empirical studies of recurrent headache disorders.

Axis I: Clinical Disorders.— The lifetime prevalence rates for selected Axis I psychiatric disorders among migraineurs are presented in Table 1, and they are shown to be elevated relative to control subjects. The strongest epidemiologic associations are observed for the depressive and anxiety disorders with at least a three-fold increase in migraineurs over nonheadache controls.12,23 Some comorbid psychiatric disorders, such as major depression described below, are quite commonly encountered, occurring in over 10% of the general population and in 30% or more of the migraine population.21 With increasing frequency and severity of headache in clinical samples of patients with migraine-related disorders34,35 and with medication overuse,36 there is a concomitant increase in the likelihood of comorbid psychiatric illness.

Table 1.—.  Lifetime Prevalence of Migraine and Selected Psychiatric Disorders12,23
DiagnosisMigraine Group (%)Control Group (%)Odds Ratio*
  1. *Odds ratios adjusted for sex.

  2. The hypomanic episode does not result in marked impairment.

Major depression34104.5
Dysthymia 9 24.4
Bipolar II 4 15.1
Manic episode 5 15.4
Panic disorder11 26.6
Generalized anxiety disorder (GAD)10 25.7
Obsessive compulsive disorder (OCD) 9 25.1
Phobia40212.6
Illicit drug use20102.2
Nicotine dependence33182.2

Mood and Anxiety Disorders.— Epidemiologic and clinical research consistently associate depressive, bipolar, and anxiety disorders with migraine.12,18–26,37 Although other headache diagnoses are much less frequently studied, anxiety and depression also appear more common among individuals with chronic tension-type headache15,30,31,38 and chronic daily headache35 than among the individuals without headache. In the case of chronic daily headache, a comorbid psychiatric disorder assessed by standard American Psychiatric Association criteria was present in 90% in a sample of 88 clinical patients.35 In a multicenter study of patients with tension-type headache (equally divided between episodic and chronic forms), psychiatric comorbidity also reached very high levels: 84% of 109 patients with chronic tension-type headache, and 70% of 108 patients with the episodic form.31

The exact nature of the relationship between migraine and mood disorders remains unclear. Although chronification of pain may breed demoralization and associated psychiatric distress,39 it is unlikely that depression or anxiety results only as a consequence of the burden of living with recurrent headache, or that psychiatric distress is the cause of migraine. Recent characterizations of psychopathology and headache have implicated shared neuropathic mechanisms between migraine and affective disorders40,41 and bidirectional influences.25,42 The concept of limbically augmented pain disorders was initially set forward by Rome and Rome in a seminal article elaborating the extent to which limbic activation may lead to a complicated intertwining and progression of chronic pain (regardless of its location) and affective distress over time.43 Cady and others have more recently expanded this concept to headache disorders in particular, suggesting cosensitization of the sensory and affective components of head pain as a possible phenomenon underlying observed comorbid relationships.44 Both concepts refer to neuroplastic processes in corticolimbic structures, where an expanding corticolimbic field becomes activated by both nociceptors and psychological stimuli over a period of time, resulting in an integrated relationship between migraine (or pain) and psychiatric disturbance in susceptible individuals.

Comorbid Anxiety-Migraine-Depression as a Syndrome.— Epidemiological family genetic studies find an association between anxiety, migraine, and depression across generations, where the presence of all three disorders in one individual emerges from one generation to the next.13,27,45 In contrast, family epidemiologic studies have not found evidence that migraine in one parent and depression in the other was associated with a combination of both in their offspring.13,27,45 The natural history observed in these studies is typically characterized by the onset of the anxiety disorder prior to the emergence of migraine, with subsequent depression that may persist throughout adult life. However, depression may emerge prior to the onset of migraine in approximately one-third of cases.42

Somatoform Disorders.— In episodic migraine patients without other psychiatric comorbidity, somatic prevalence is comparable to a sample of primary care patients. Somatization has rarely been addressed in epidemiologic or clinical studies.46 However, there is at least some evidence that headache can be a manifestation of a somatoform disorder, where the physical complaints and/or associated impairment is in excess of what would be expected based on the history, physical examination, or laboratory findings. Somatoform disorders include somatization disorder, conversion disorder, hypochondriasis, or pain disorder associated with psychological factors (where psychological factors are judged to play the major role in the onset, severity, exacerbation, or maintenance of pain).32 In two recent studies of comorbidity in tertiary care, between 6% and 22% of those with headache met diagnostic criteria for a somatoform disorder.15,35 The two studies observed somatoform disorders equally among women and men. The most common somatoform disorder associated with headache was “undifferentiated somatoform disorder.”15 In somatoform disorder, headache would represent only one of many medically unexplained somatic complaints such as fatigue, loss of appetite, gastrointestinal symptoms, and urinary complaints.

Substance Dependence and Abuse.— As shown in Table 1, in an epidemiologic study of young adults enrolled in a large Health Management Organization (HMO), both nicotine dependence and illicit drug use were significantly more common in those with migraine than those with no migraine history.23 Mitsikostas and colleagues found the highest risk for major depression in headache outpatients in those with “drug abuse headache,”36 a somewhat misleading term that bridges the more recent diagnosis of “medication overuse headache” in ICHD-233 and medication misuse or abuse. In general, there is a relative paucity of research in this area compared to studies of comorbidity with mood and anxiety disorders. Saper and others have recently observed that there appear to be two types of patients with medication overuse headache: those with simple analgesic overuse, and a more complex type of medication overuse driven by behavioral factors that include the search for sedation, altered consciousness, and the willful disregard of prescription guidelines.47 They argue for a distinction between simple and complex medication overuse headache, based on these biobehavioral factors, and note that complex patients have a more challenging prognosis.

Axis II: Personality Disorders.— Personality disorder have been less frequently examined in the empirical headache literature than affective disorders, but are considered to markedly complicate headache management.48 Personality disorders represent an inflexible and pervasive pattern of behavior leading to clinically significant distress or impairment in functioning.32 The behavior patterns are generally stable and enduring, but may become more intense during times of stress, acute pain, anxiety, or depression. Personality disorders are present in 10% to 13% of the general population.49

Borderline personality disorder is probably the most difficult of the personality disorders.48 The essential feature is a pervasive pattern of unstable swings in interpersonal relationships, self-image, and mood, often complicated by impulsivity, self-destructive behaviors, and inappropriate, intense anger. Borderline patients may further complicate treatment with manipulative behavior and substance misuse. Borderline personality is found in approximately 2% of the general population, about 10% of psychiatric outpatients, and close to 20% of psychiatric inpatients.32,48 Childhood emotional trauma and abuse (physical, sexual) are reported much more frequently among borderline personality disorder (40% to 86% with alleged sexual abuse) than among the general population (22% to 34%).50 There are no epidemiologic studies of headache and comorbid borderline personality. However, significant headaches are a complaint of about 60% of patients with the personality disorder presenting for acute treatment to a hospital emergency department.51

Though even less empirical or systematic information is available for other personality disorders, some are believed to at least in some cases complicate headache treatment because of the difficult interpersonal style of the patient (eg, histrionic patients may greatly exaggerate physical complaints, passive/dependent patients may excessively rely on healthcare providers, narcissistic patients may be demanding).16,48,52,53

CLINICAL IMPLICATIONS

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

In recent years, research has begun to address the clinical implications of psychiatric disorders on headache. There is some preliminary evidence described below that psychiatric illness may impact the natural history for migraine. Preliminary evidence is also emerging suggesting that addressing the psychiatric illness may favorably impact headache management.

Psychiatric Comorbidity Associated With Chronification and Intractability of Migraine.— The presence of comorbid psychiatric illness may contribute to the intractability and chronification of migraine. An 8-year follow-up study of 100 young adults with headache examined the relationship between psychiatric disorders at initial evaluation and headache status at follow-up.38 In this study, sleep disturbance was defined as a psychiatric disorder, in addition to depression and anxiety conditions, among others. For those with two or more psychiatric disorders at initial evaluation, 57% exhibited no improvement or deterioration in headache at follow-up, 29% were improved, and only 14% were headache free. In contrast, patients with no psychiatric disorder or only one comorbid condition (eg, sleep disturbance) exhibited greater headache improvement 8 years after the initial evaluation: only 7 to 15% were the same or worse, 46 to 53% had improved, and 39 to 40% were headache free.

Moreover, there is emerging evidence that a number of behavioral/psychological risk factors are associated with progression of headache from episodic to chronic and daily54,55 and that psychological distress may play an even greater role in the transformation and chronification of headache than does analgesic overuse/abuse.56 Clearly, large-scale and long-term longitudinal studies are called for to better understand not only the natural history of migraine over time but also the influence of comorbid psychopathology upon the prognosis and progression of migrainous headache.

Differential Headache Treatment Response in Patients With Psychiatric Comorbidity.— Psychiatric comorbidity is often a negative prognostic indicator for headache treatment,11,17,52,57 highlighting the general relevance of screening and tracking psychiatric disorders that may account for variance in headache outcomes in clinical trials. Behavioral and psychosocial variables (including specific forms and levels of psychopathology) can play an important role as “mediator” and “moderator” variables in influencing treatment outcomes. A moderator is a variable that predicts the level of impact a treatment will have on persons with a certain characteristic. Identification of influential moderators is a key component for establishing the clinical utility of a treatment by revealing whether a specific individual is likely to benefit from a specific treatment. As it relates to clinical trials, a mediator potentially explains “treatment mechanisms”—why or how treatments have their effect (mediator and moderator variables are addressed in detail in a companion article within this issue58).

Interestingly, few studies have moved beyond the prognostic implications of psychiatric comorbidity and conducted post hoc analyses of headache outcomes in relation not only to psychiatric disorders but also treatment modality. As described below, differential headache outcomes have been observed with modalities that would be presumed to impact psychological or psychiatric processes (eg, antidepressant medication, stress management) as well as headache. The studies below have observed differential headache outcomes in part related to the use of a treatment modality that would be considered appropriate to the psychiatric disorder, and they demonstrate the potential value of additional research specifically designed to examine the role of psychopathology in mediating and moderating headache outcomes.

For example, Holroyd and colleagues reported results from a large, randomized, placebo-controlled trial comparing behavioral and pharmacologic treatments for chronic tension-type headache.59 Following 8 months of treatment, the mean headache improvement was 64% for the combination of stress management with tricyclic medication, 38% for stress-management alone, 35% for medication alone, and 29% for placebo. Of particular interest is a post hoc analysis, which found that differential outcomes could be accounted for by pretreatment depression or anxiety.60 When high pretreatment levels of headache severity and disability coexisted with an anxiety disorder, stress management alone was superior to placebo medication. When comorbid anxiety or depression was present along with high headache severity and disability, tricyclics were superior to placebo. However, for patients with lower levels of headache activity in the absence of anxiety or depression, tricyclics did not differ from placebo.

Studies of pharmacotherapy alone also point to the importance of distinguishing between patients with and without comorbid psychiatric illness. Post hoc analysis of a small open label study of the antidepressant medication nefazodone for chronic daily headache found evidence for differential treatment response in depressed versus nondepressed patients.61 Although headache severity indices did not differ between the two groups at baseline, depressed patients showed a mean 85% improvement during the third month of treatment, versus a mean improvement of 41% for nondepressed patients. In this study, depression was defined by a Beck Depression Inventory II score of 15 or higher (median = 22, mean = 25, range = 15 to 43). Depressed patients also improved more than nondepressed patients when researchers identified clinical depression using the Hamilton Rating Scale for Depression. For patients with anxiety identified by the Hamilton Rating Scale for Anxiety, outcomes were marginally better.

As yet, there is no cogent evidence that treating depression or anxiety per se improves headache outcomes, an important area for further research discussed in the article on future directions in this issue of Headache.62

METHODOLOGICAL CONSIDERATIONS

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

As has been pointed out by Merikangas and Stevens,13 there is substantial diversity in the methods employed in studies of comorbidity of headache and psychopathology conducted to date. A particularly key concern is that many such studies have employed suboptimal assessment and diagnostic methodologies. This includes not only idiosyncratic definitions of headache syndromes (many studies were conducted prior to the release of the initial International Headache Society classification in 198863), but also nonstandardized definitions and measurements of psychiatric disorders and psychopathology.

Measurement Issues and Interpretation Bias in Assessing Psychiatric Symptoms.— Extensive clinical research has revealed that although levels of psychological symptoms observed in headache patients frequently have exceeded those of normal control samples, clinically meaningful levels of psychopathology have been observed less often.11 The majority of studies reporting increased maladaptive behavior or genuine psychopathology among headache patients employed the Minnesota Multiphasic Personality Inventory (MMPI) or other validated psychometric measures of psychological symptoms or personality (ie, Beck Depression Inventory, Spielberger State/Trait Anxiety Inventory, Eysenck Personality Inventory). When clinically meaningful levels of psychopathology have been reported, it is usually among only a subset of the headache patient sample. This point is exemplified by the three studies that have employed cluster analytic techniques to identify homogeneous subgroups of headache patients based upon their MMPI profiles.64–66 Each of the three studies identified only one subgroup of patients with psychopathological MMPI profiles (ie, patients exhibiting clinically significant elevations on several psychological symptom scales), and only a small proportion of the patients (between 5% and 13%) have exhibited these “psychopathological” profiles. The majority of patients were deemed to have an essentially normal MMPI profile (62% to 72% of patients). Note, however, that neither the MMPI nor the revised MMPI-2 alone is adequate for conferring a psychiatric diagnosis following DSM-IV criteria.

Headache diagnosis and severity of headache symptoms are also likely to play roles in the relationship between psychopathology and headache as suggested by a recent large-scale study by Bigal and colleagues.39 The latter study revealed elevated MMPI-2 scores among only a small proportion of patients with episodic migraine or new daily persistent headache. The authors did find elevated scores in a considerably greater proportion of patients with chronic migraine or analgesic rebound headache leading them to speculate that the observed psychopathological factors appear to be a consequence of the chronification process.

Even these figures may overestimate the number of those with recurrent headache exhibiting significant psychopathology. In the relevant research and in clinical practice, psychopathology typically is identified using measures that in part reflect somatic or vegetative symptoms of chronic headache. Unfortunately, most studies examining psychopathology among those with headache to date have ignored the potential impact of transdiagnostic symptoms (ie, symptoms common to both psychopathology and headache). For example, research has shown headache patients' elevated Beck Depression Inventory scores to arise through their endorsement of more somatic items on the (eg, “I am very worried about physical problems such as aches and pains,”“my appetite is much worse now”), but not more nonsomatic items (eg, “I am sad all the time,”“I am disappointed in myself”).67 Thus, in many instances, the elevations in depression scale scores observed among headache patients probably reflect either direct headache-related symptoms, or shared somatic symptoms with depression, and not depression per se. Note that major depression requires either (i) depressed mood or (ii) loss of interest or pleasure as primary symptoms required for the diagnosis.32 More generally, because many psychological symptom measures similarly do not distinguish the somatic symptoms of psychiatric disorders from the somatic symptoms of headache, they likely overestimate psychopathology when administered to those with recurrent headache.11,67

On a related note, investigators should be careful to distinguish changes in psychometric scales from clinically meaningful changes in psychopathology. All too often, investigators in the past have used shorthand terminology in referring to an observed change in a self-report measure of psychopathology which, when taken out of context, generates the impression that the observed change reflects a legitimate shift in psychopathology. To cite a common example, a number of researchers have referred to an observed reduction in a Beck Depression Inventory (BDI) score following treatment simply as a “decrease in depression.” While a BDI score may indeed reflect a genuine shift in psychopathology, it often does not do so. In fact, in most studies, the typical headache patients' BDI scores at pretreatment fall within the “subclinical” range. While subclinical levels of depression or anxiety may be of significance in treatment, a posttreatment reduction in BDI score does not necessarily reflect a true reduction in “depression” per se. Investigators are urged to employ more precisely descriptive terminology when reporting the findings of psychopathology measures (ie, “There was a significant reduction in BDI score” rather than “There was a significant decrease in depression”) and reserve the use of terminology referring to psychiatric disorders and psychopathology (a la DSM-IV32) for their intended purpose.

Maladaptive Coping Strategies.— As noted above, the average headache patient does not have a psychiatric disorder. Nevertheless, it is at least as important to address psychological patterns such as headache-relevant stressors, coping strategies, and beliefs that impact headache symptoms but may not reflect a psychiatric disorder. Measures designed to assess levels of stress, cognitive and behavioral strategies for coping with stress, emotional reactivity to pain, dysfunctional beliefs about headaches, factors influencing perceived personal control of headaches, and headache-related disability may often prove to be more valuable in the study of mediators and moderators of treatment outcome than measures designed to assess more severe psychopathology (eg, MMPI, Eysenck Personality Inventory). For example, studies conducted to date have consistently shown: (i) the level of daily, minor stress is positively correlated with headache activity level,68–72 and (ii) headache patients are likely to engage in maladaptive coping behaviors in managing stress or headaches.73–78 Collectively, these findings suggest that we might more profitably shift the focus of psychological assessment from psychopathology toward other, “less pathological” behavioral and cognitive variables in future research seeking to better delineate the role of psychological factors in recurrent headache.

Diagnosis of Headache and Psychiatric Disorders.— As indicated by Merikangas and Stevens,13 inconsistencies in diagnostic methods for psychiatric and headache disorders is a limitation of earlier literature. Without question, additional research addressing the comorbidity of headache is needed using the current ICHD-2 headache classification as well as well validated and reliable methods for assessing and diagnosing psychopathology, for example, Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) and Structured Clinical Interview for DSM-IV Personality Disorders (SCID-II),79-81 or the explicit use of DSM-IV.32

Headache Attributed to Psychiatric Disorder.— The latest revision of the ICHD-2 classification (Headache Classification Subcommittee, 2004) has made a provision for the small number of patients for whom headache ostensibly is causally related to psychiatric illness.33 The classification of headache attributed to psychiatric disorder (ICHD-2 12. code type; pp. 121-123) is reserved for those patients for whom a headache occurs in the context of a psychiatric condition where headache is believed or known to be a symptom of the psychiatric illness, and not a primary disorder in its own right.

The classification provides criteria for only two varieties of this form of headache: headache attributed to somatization disorder (ICHD-2 12.1), and headache attributed to psychotic disorder (ICHD-2 12.1; see Table 2). When a new headache occurs in close temporal relation to a psychiatric disorder, it is coded as a secondary headache attributed to that disorder (ie, using the 12. code type). When a pre-existing headache is made worse in close temporal relation to a psychiatric disorder, the patient can be given a diagnosis of either the pre-existing headache, or else given both the pre-existing headache diagnosis and a “code 12.” diagnosis. The code 12 diagnosis becomes definite only when the headache resolves or greatly improves after remission of the psychiatric disorder. If the psychiatric disorder does not remit spontaneously or when there has been insufficient time for this to happen, a diagnosis of headache probably attributed to a psychiatric disorder is applied.

Table 2.—.  ICHD-2: Headache Attributed to Psychiatric Disorder: Two Coded Subtypes (12.1, 12.2) and Seven Appendix Listings (A12.3-A12.9)*
  1. *Excerpted from the International Classification of Headache Disorders, 2nd ed.33

12.1 Headache attributed to somatization disorder
History of many physical complaints beginning before age 30 that occur over a period of several years and result in treatment being sought and/or in significant impairment in functioning.
At least four pain symptoms, two nonpain gastrointestinal symptoms, one sexual or reproductive symptom, and one pseudoneurological symptom.
After appropriate investigation, each of these symptoms cannot be fully explained by a known general medical condition or the direct effects of a substance or medication.
12.2 Headache attributed to psychotic disorder
Delusional belief about presence or etiology of headache occurring in the context of a psychotic disorder (eg, delusional disorder, schizophrenia, major depressive episode with psychotic features, manic episode with psychotic features).
Headache occurs only when delusional.
Headache resolves when delusions remit.
Headache not attributed to another cause.
A12.3 Headache attributed to major depressive disorder
A12.4 Headache attributed to panic disorder
A12.5 Headache attributed to generalized anxiety disorder
A12.6 Headache attributed to undifferentiated somatoform disorder
A12.7 Headache attributed to social phobia
A12.8 Headache attributed to separation anxiety disorder
A12.9 Headache attributed to posttraumatic stress disorder

As acknowledged in the ICHD-2, there are circumstances where headache may occur exclusively during common psychiatric disorders, leading some clinicians and researchers to question whether the headache symptoms may best be considered as attributed to these disorders. Examples include major depressive illness, panic disorder, generalized anxiety disorder, undifferentiated somatoform disorder, and more (see Table 2). However, given the paucity of empirical evidence addressing these circumstances, the Committee (Headache Classification Subcommittee, 2004) refrained from forming diagnostic codes for these cases, instead listing them in an Appendix (pp. 146-149) to encourage further research into this area.33

Clearly, much research is needed to validate and establish the utility of this new classification. In our opinion, there may be very few cases where headache could actually be attributed to a psychiatric disorder. For example, the emergence of severe migraine headaches during episodes of a paranoid delusional disorder, followed by the remission of both headaches and delusions with neuroleptic medication, would best be considered an example of shared neurophysiological mechanisms underlying the two disorders—not as evidence that the delusional disorder was the cause of the headache.

IMPLICATIONS FOR FUTURE RESEARCH

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

The external validity or generalizability of research results requires that the research sample be representative of the general population of those with headache with a specific headache diagnosis. That is, the research sample should represent and track in a systematic manner the common concomitant conditions. In pharmaceutical studies assessing medication effectiveness, the United States Food and Drug Administration's criteria for medication approval ensure that patient samples are large and represent the population of patients for whom the treatment would eventually be applied. However, in most cases patients with significant psychopathology are specifically excluded from the study.

Behavioral treatments are developed and tested in a much different manner, usually within academic or patient care settings. In their initial stages, behavioral studies may require intensive interaction with those with headache beyond what is required to administer a drug. There is often no available external funding, at least in the initial phases of research, and no large corporations with a financial interest in funding these studies. Consequently, behavioral treatment studies often rely on relatively small samples. The inclusion of patients with severe psychiatric disturbance may seriously bias results. This potential for bias becomes progressively less of a concern as the sample size increases and patients are randomized to treatment.

For behavioral treatment, it is also reasonable, appropriate, and in fact necessary to develop and test the effectiveness of therapies for headache patients with comorbid psychiatric disorders. In clinical practice, referrals for behavioral treatment often include patients with severe levels of chronic daily headache, who failed to respond adequately to medication therapies, and who experience significant psychiatric disturbance. The high prevalence of anxiety and depression in the general population, odds ratios that indicate an association between migraine with psychiatric disorders well beyond what would be expected by chance, evidence in support of the anxiety-migraine-depression syndrome as a distinct disorder, and the association of comorbid psychiatric illness with headache intractability, are all significant factors that support the need for further research in this area. For a more in-depth discussion of related headache research design issues that make allowances for and systematically examine comorbid psychopathology, readers are referred to contributions by Lipchik and colleagues82 and Nicholson and colleagues58 in this volume.

RECOMMENDATIONS FOR RESEARCH

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

In closing, we offer a number of additional research recommendations for addressing psychiatric comorbidity. Psychiatric comorbidity is relevant to the randomization and stratification of patients in outcome studies82 and the choice of secondary outcome measures83 as discussed in companion articles in this issue. The importance of developing and adequately testing treatments for patients with significant comorbidity also is underscored in the article on future directions62 as a research priority for behavioral headache research.

  • (i) 
    Identify Dual Headache and Psychiatric Diagnoses With Appropriate, Valid, and Reliable Criteria. When conducting studies of psychiatric disorders comorbid with headache, formal “dual diagnoses” should be made according to both ICHD-2 headache and DSM-IV psychiatric diagnoses. Psychiatric diagnoses, as with headache diagnoses, should follow formal diagnostic assessment methods and clearly specified diagnostic criteria. When psychometric tests are used to assist in diagnosis, clinically relevant cut-off scores should be identified.
  • (ii) 
    Differentiate Subclinical Levels of Depression and Anxiety from Diagnosed Psychiatric Disorders. Subclinical levels of depression and anxiety remain worthy of study, but should not be identified as clinical depression or anxiety disorders. Using median or quartile cut-off scores that suggest a mild mood disturbance below the level of clinical significance should not be reported as indicative of “depression” or “anxiety.” When subclinical levels of disturbance are investigated, special care should be made to address the specific symptoms comprising the subclinical syndrome, with attention to transdiagnostic symptoms.67 In some cases it may be more helpful to focus on the symptom (eg, sleep disturbance) in its own right, rather than as a symptom of “depression.”
  • (iii) 
    Validate New ICHD-2 Diagnoses. Research is needed to validate the new ICHD-2 diagnostic codes wherein headache is specifically attributed to a specific psychiatric disorder (Headache Attributed to Psychiatric Disorder, Table 2).
  • (iv) 
    Expand Epidemiological Research. It would prove valuable to better identify the prevalence of patients with various comorbid psychiatric disorders within various headache populations (eg, primary care practice setting, specialty setting, general population). The psychiatric disorders should include various Axis I (clinical syndromes) and Axis II (personality disorders) diagnoses, and should also evaluate the impact of these disorders upon headache presentation and prognosis. Similarly, it would prove valuable to better identify the prevalence of specific headache disorders among various psychiatric patient samples.
  • (v) 
    Identify Relevant Psychiatric and Behavioral Mediator/Moderator Variables. Investigators are encouraged to assess, track, and evaluate psychiatric/psychological factors when carrying out clinical trials (perhaps in stratified research designs82) in order to differentiate empirically those disorders which impact headache symptoms and treatment. Research methods should specify the presence of psychiatric comorbidity in study inclusion/exclusion criteria to accurately identify and classify the subset of headache patients with psychiatric disorders. When the number of patients is sufficient, investigators should perform and report post hoc analyses of differential dropout and response to treatment for patients with and without comorbid psychiatric disorders. Studies should be designed that target patients with comorbid conditions in order to aid in the identification of psychiatric features that mediate/moderate treatment response and development of optimal behavioral and pharmacologic treatment strategies.58
  • (vi) 
    Develop Empirically Based Screening/Treatment Algorithms. Future investigations should develop and test algorithms for screening and managing headache patients with psychiatric comorbidity. Ideally, screening measures should be validated on headache or similar medical patient populations. It would be particularly advantageous to develop and validate brief screening (eg, self-report) tools that can be applied efficiently in physicians' practice settings to initially flag potential psychopathology.46,84 With respect to treatment, both pharmacologic and behavioral therapies should be evaluated specifically on populations of headache patients with psychiatric disorders, who are routinely excluded from clinical trials research. The impact of treating the comorbid conditions, such as depression or anxiety, on headache outcomes should be evaluated as well.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES

A substantial number of those with headache exhibit psychiatric disorders, and their presence generally portends a poorer outcome for headache management. Psychiatric comorbidity is a significant moderator of treatment outcome, and can be a negative prognostic indicator for some pharmacologic and behavioral treatments. Unfortunately, there are no evidence-based algorithms to guide research and practice in this important clinical arena. The prevalence and potential complicating influence of psychiatric disorders on headache treatment have prioritized psychiatric comorbidity on the agenda for behavioral headache research.62 Historically, nonscientific methods have misrepresented the relationship between headache and psychopathology, in some cases inappropriately stigmatizing those with headache and perhaps suppressing the biological science of headache. Examination of psychiatric comorbidity poses special considerations with respect to research design, analysis, and reporting. Objective and systematic experimental methods are recommended for future empirical investigations.

REFERENCES

  1. Top of page
  2. Abstract
  3. HISTORICAL CONTEXT
  4. PSYCHIATRIC DISORDERS AND HEADACHE
  5. CLINICAL IMPLICATIONS
  6. METHODOLOGICAL CONSIDERATIONS
  7. IMPLICATIONS FOR FUTURE RESEARCH
  8. RECOMMENDATIONS FOR RESEARCH
  9. CONCLUSIONS
  10. REFERENCES