Migraine Comorbidity

Kathleen Ries Merikangas

Nancy C. D. Low

Birthe Krogh Rasmussen

Comorbidity, introduced by Feinstein, refers to the presence of any additional coexisting ailment in a patient with a particular index disease (31). Failure to classify and analyze comorbid diseases can create misleading medical statistics and may cause spurious comparisons during the planning and evaluation of treatment for patients. Comorbidity can alter the clinical course of patients with the same diagnosis by affecting the time of detection, prognostic anticipations, therapeutic selection, and post-therapeutic outcome of an index diagnosis (44). In addition, it can also affect the length of hospital stay, response to somatic treatment and mortality (7,68,89).

Nonrandom co-occurrence of two conditions may be attributable to several methodologic artifacts, including samples selected from clinical settings that are nonrepresentative of persons with the index disease in the general population (“Berkson’s Paradox”) (4); assessment bias, in which the co-occurrence of two conditions is an artifact of overlap in the diagnostic criteria or in the assessments employed to ascertain the criteria; and the lack of an appropriate comparison (or control) group with which to account for factors that confound the association between the two conditions.

METHODOLOGY OF COMORBIDITY STUDIES

Associations between migraine and a variety of somatic and psychiatric conditions have been reported in the literature since it was first described as a discrete syndrome. Because most of the early descriptions of such associations were based on clinical case series, empirical evidence was lacking. There are several factors that complicate the investigation of comorbidity of migraine and other conditions. These include discrimination from “migraine equivalents” defined as alternate manifestations of migraine that occur in an “attack-like” fashion including abdominal pain, dizziness or vertigo, or visual symptoms; lack of specificity of symptom expression or constellations within individuals over time; and the involvement of several systems including the cardiovascular, gastrointestinal, and sensory organs, as well as both the peripheral and central nervous systems.

There is dramatic variability in the methodology of studies of comorbidity and migraine, and this limits the conclusiveness of the findings. Studies of comorbidity require valid definitions and reliable ascertainment of each of the disorders under consideration. The majority of migraine comorbidity studies were conducted prior to the introduction of the International Headache Society criteria and employed idiosyncratic definitions of migraine ranging from recurrent headaches to classical migraine with neurologic prodromes. Moreover, standardized definitions of both disorders are rarely included in clinical or community studies. In general, clinical series and case-control studies have employed the most thorough clinical evaluations of subjects. In contrast, community studies tend to apply less rigorous definitions of the disorders because collection of extensive diagnostic information on both conditions was precluded by the sheer magnitude of the studies.

Community studies generally have sufficient statistical power to detect associations between migraine and rare diseases. Indeed, the negative associations in the smaller sample sizes of the clinical and case-control studies are often the result of β-type errors rather than a true lack of association between migraine and other diseases. Another methodologic limitation of the majority of studies of migraine comorbidity is the failure to incorporate confounding risk factors that could explain the association between several diseases and migraine. In addition, the interrelationships between the comorbid disorders themselves are often unaccounted for in multivariate analyses, thereby yielding spurious associations. Finally, as noted, the samples of both the clinical series and case-control studies may be biased with respect to the increased probability that
persons with two or more conditions are represented in clinical samples (Berkson’s paradox). Thus, population-based studies are necessary to identify such biases in treatment samples. Case-control studies and cross-sectional epidemiologic studies mainly generate hypotheses about possible associations, and longitudinal population-based studies are necessary to test the hypotheses and reliably identify patterns of comorbidity.

EVIDENCE FOR MIGRAINE COMORBIDITY

Comorbidity of migraine and several other disorders has been reported in clinical series, case-control studies, and epidemiologic surveys. Using data from a large-scale epidemiologic study in the U.S. adult population, Merikangas and Fenton (58) found that migraine was associated with disorders of the cardiovascular, allergies/asthma, gastrointestinal, neurologic, and psychiatric systems. The following discussion focuses on associations between migraine and selected somatic conditions for which evidence has been presented in the literature.

Cardiovascular Diseases

Comorbidity between migraine and cardiovascular conditions including stroke, hypertension, and numerous manifestations of heart disease have been the most widely studied of all classes of somatic conditions. It is likely that the focus on these conditions derives from concern regarding migraine-induced stroke and the role of underlying vascular disease in triggering attacks of migraine.

Hypertension

In the classic epidemiologic study by Waters (85), no differences emerged between the levels of systolic and diastolic blood pressure among migraineurs compared to the remainder of the sample. Similar results were obtained from a large study of male Swedish conscripts (74) and in large-scale community surveys conducted by Chen et al. (22) and Weiss (87). In a cross-sectional survey of adults in the general population, Rasmussen and Olesen (71) found no association between headache disorders and arterial hypertension, but did find that women with migraine had a significantly higher diastolic blood pressure than those without migraine. In a prospective study of blood pressure and the risk of headache in the community, Hagen et al. (40) examined 22,685 adults in Norway and did not find a clear association between migraine and blood pressure, but did observe a trend in women with high systolic blood pressure to be at a reduced risk of migraine.

In contrast to the above results from epidemiologic studies, nearly all case-control studies examining the association between migraine and hypertension have reported a positive association. The first systematic study was conducted by Gardner et al. (36) in 1940, who found that the mean systolic blood pressure was higher in migraineurs over age 40 than among age-matched controls. More recently, a comparison between 200 subjects with recurrent headaches and the same number of age- and sex-matched controls, Featherstone (29) found a two-fold increase in the rates of hypertension among cases than controls. A study by Markush et al. (54) yielded similar findings among index cases of hypertension; there was an increased risk of migraine among hypertensives as compared to controls. Research has also shown that relatives of migraine patients are more likely to be hypertensive (24,49). Of note is that findings from studies using clinically ascertained cases may be vulnerable to Berkson’s fallacy.

In summary, in view of the largely negative studies done among large-scale community surveys, and despite the positive though inconsistent results of case-control studies, a definite association between migraine and hypertension has yet to be established. Moreover, all findings should be qualified by the fact that pain may result in higher blood pressure. Whether high blood pressure precedes the onset of migraine or rather is caused by the presence of migraine with recurrent headaches is unknown.

Heart Disease

Heart diseases including mitral valve prolapse, coronary artery disease, ischemic heart disease, angina, and arrhythmias have also been purportedly associated with migraine. There have been no large-scale, prospective epidemiologic studies specifically examining the association between migraine and these heart diseases; however, there are many case-control studies. In the examination of migraine and mitral valve prolapse, inconsistent measures and definitions of mitral valve prolapse and migraine, in addition to insufficient power to test this association adequately, are likely reasons for the discrepant findings among case-control studies (39,35).

Findings from case-control studies have been essentially negative when looking at the relationship between migraine or recurrent headache and heart disease (heart attacks, coronary heart disease, angina), after controlling for well-known cardiovascular risk factors (17,23,29,33, 51,72).

Several family studies have been undertaken to look at the relative risk among relatives of migraineurs compared to controls (24,34,49), and when taken together, these studies find an the average relative risk of approximately 1.5. Thus, overall there is little evidence for an association between heart disease and migraine, after considering the effects of the other cardiovascular risk factors including smoking and hypertension.

Neurologic Diseases

Regarding migraine and neurologic diseases, it may be important to note that migraine may be secondary to these disorders, that is, symptomatic migraine of established disease. This pertains, for example, to migraine after stroke, multiple sclerosis (MS), and tumors. Brain disorders may cause both symptomatic migraine and epilepsy, and in this way, give a false association. In examining comorbidity, the main interest is in primary migraines.

Stroke

The association between migraine and stroke has received renewed attention because of a recent paper reporting that persons with migraine have an increased prevalence of cerebellar infarcts compared to other relatives (45). Several population-based studies in the United States found an association between migraine and stroke (16,32,59,76). Buring et al. (16) reported an association between migraine and stroke in data from the Physician’s Health Survey (32), a large-scale prospective longitudinal study of the efficacy of aspirin in male physicians. When assessed after 5 years, men with migraine exhibited an elevated risk of stroke when compared to those without a history of migraine, yielding a risk ratio of 2.0 (95% confidence interval [CI] 1.1 to 3.6) for ischemic stroke, and 1.8 (95% CI 1.1 to 3.2) for all stroke. Merikangas et al. (59) examined the association in a large-scale epidemiologic study of 13,380 adults drawn from the U.S. adult population, and found those with migraine have an increased risk ratio of 1.5. As well, the risk of stroke, given a history of migraine, decreases with increasing age; for example, at age 40 the risk ratio is 2.8, compared to a risk ratio of 1.7 at 60 years.

Nearly all of the case series that report on the incidence of stroke among persons with migraine are limited to young adults. In examining this association, it is critical to discriminate between migraine-induced stroke and stroke that occurs among individuals with migraine. Reviews by Featherstone (30) and Welch and Levine (88) discuss the diagnostic criteria for migrainous cerebral infarction.

The majority of case-control studies of samples, which have included both sexes and a wider age range, suggest an association between migraine and stroke. In a study by the Collaborative Group for Study of Stroke in Young Females (67), 34.2% of women ages 15 to 44 with thrombotic stroke had a lifetime history of migraine. Henrich and Horwitz (42), using a hospital record review of migraine in a systematic series of first admissions for stroke among persons ages 15 to 65, found a significant association between classic, but not common, migraine and ischemic stroke. This association only emerged after controlling for other established risk factors for stroke. Carolei et al. (18) examined 308 adults of ages 15 to 44 years who were admitted for either transient ischemia attack or stroke, compared them with 591 age- and sex-matched controls, and found that for women younger than 35, a history of migraine was a significant risk factor for stroke (odds ratio [OR] = 3.1, 95% CI 1.5 to 9.0). Additionally, Chang et al. (20) found an increased risk of ischemic stroke, but not hemorrhagic stroke, in a case-controlled sample of migrainous women between 20 and 44 years. And finally, in their review of the literature between 1995 and 2001, Curtis et al. (25) confirmed that migraine in women using oral contraceptives is a risk factor for stroke.

However, two case-control studies have failed to report an association between stroke and migraine. Featherstone (29) did not report an increased risk of stroke among subjects with recurrent headache over controls; similarly, Tzourio et al. (81) found no overall association among 212 male and female patients admitted for stroke, matched for age, sex, and hypertension, except in a subset of women younger than 45 (OR = 4.3, 95% CI 1.2 to 16.3). Interestingly, in a follow-up case-control study with only women younger than 45, Tzourio et al. (82) found that classic migraine carried a larger risk for stroke compared to common migraine (OR = 6.2, 95% CI 2.1 to 8.0 versus OR = 3.0, 95% CI 1.5 to 5.8).

The family studies (24,49) cited previously that compared parents of migraine patients and controls also examined the rates of stroke. Leviton et al. (49) selected parental mating types in which only one parent had recurrent severe migraine headache and found no increase in the history of stroke in the parent with headache compared to the unaffected parent. In contrast, the family history study of Galiano et al. (34) also reported a nearly two-fold increased risk of stroke among the relatives of subjects with migraine compared to controls.

Only gold members can continue reading. Log In or Register to continue