Endometriosis Is Associated With Prevalence of Comorbid Conditions in Migraine
From the Department of Neurology, University of Toledo Medical Center, Toledo, OH, USA (Drs. Tietjen, Herial, Utley, White, and Hafeez); Division of Neurology, Duke University Medical Center, Durham, NC, USA (Dr. Bushnell).
Address all correspondence to Dr. Gretchen E. Tietjen, 3120 Glendale Avenue, RHC 1450, University of Toledo-Health Science Campus, Toledo, OH 43614, USA.
Objective.—To examine the headache characteristics of women with migraine and endometriosis (EM), and differences in the prevalence of comorbid conditions between female migraineurs with EM, without EM and nonheadache controls.
Background.—Migraine and EM are common conditions in women of reproductive age, and both are influenced by ovarian hormones. The comorbidity of migraine and EM is newly recognized, but reasons for the association are uncertain.
Methods.—This is a cross-sectional study of female headache outpatients and healthy controls conducted at University of Toledo and Duke University in 2005 and 2006. After a headache specialist determined headache frequency and diagnosis (based on criteria of the second International Classification of Headache Disorders), patients completed a self-administered electronic survey with information on demographics, headache-related disability, menstrual disorders, premenstrual dysphoric disorder (PMDD), vascular event risk, and comorbid conditions, including irritable bowel syndrome (IBS), fibromyalgia (FM), chronic fatigue syndrome (CFS), interstitial cystitis (IC), depression, and anxiety.
Results.—Study enrolled 171 women with migraine and 104 controls. EM was reported more commonly in migraineurs than in controls (22% vs 9.6%, P < .01). Frequency of chronic headache was higher in migraineurs with EM compared to without EM (P= .002) and median headache-related disability scores were also higher in the EM group (P= .025). Symptoms of PMDD were more common in migraineurs, but frequency did not differ by EM status. Migraineurs with EM reported more menorrhagia, dysmenorrhea, and infertility compared to the migraine cohort without EM and to controls. Depression, anxiety, IBS, FM, CFS, and IC were more common in migraine with EM group than in controls. Anxiety (OR = 2.2, 95% CI 1.0–4.7), IC (OR = 10.6, 95% CI 1.9-56.5), and CFS (OR = 3.6, 95% CI 1.1–11.5) were more common in migraine with EM group, than in the cohort with migraine without EM.
Conclusion.—Prevalence of EM is higher in women with migraine than in nonheadache controls. Migraineurs with EM have more frequent and disabling headaches, and are more likely to have other comorbid conditions affecting mood and pain, compared to migraineurs without EM.
premenstrual dysphoric disorder
chronic fatigue syndrome
irritable bowel syndrome
Among women of reproductive age, migraine1 and endometriosis (EM),2 hormonally influenced conditions with a wide spectrum of severity, are highly prevalent. The prevalence of migraine in women ranges between 15% and 18%.3 Endometriosis, a condition characterized by endometrial tissue growing outside the uterus, occurs in 5.5 million women in the United States alone, and is estimated to occur in 10% of women of reproductive age.4 Like migraine,5–9 EM has been demonstrated to have a greater than coincidental association with asthma, hypothyroidism, chronic fatigue syndrome, and fibromyalgia.10 Studies have suggested that migraine and EM are themselves comorbid,11,12 but there has been little research into characteristics of this overlap population. Although the nature of the linkage remains speculative, it has been suggested that the sensory fibers from ectopic endometrial implants produce central sensitization responsible for widespread hyperexcitability of neurons in the central nervous system.13
Our objectives in this study are to examine (1) the headache characteristics of women with migraine and EM compared to female migraineurs without EM and (2) the association of EM with other comorbid conditions in women with migraine, and in nonheadache controls.
Study Sample.— A cross-sectional study of headache clinic patients and healthy controls was conducted after approval from the institutional review boards (IRB) at 2 different institutions (University of Toledo Medical Center, Toledo, OH and Duke University Medical Center, Durham, NC). All participants in this study were women. Cases were women with migraine, examined by a headache specialist (GT, CB, FH) in the clinic. During the clinic encounter, the physician determined the headache patient's eligibility for the study using the following inclusion criteria: women with migraine, as defined by the second International Classification of Headache Disorders (ICHD-II) criteria,14 18 years and older, willingness to complete a self-administered electronic questionnaire on a personal digital assistant (PDA), eg, the Palm® (Palm, Inc., Sunnyvale, CA, USA) handheld device. Exclusion criteria included the following: not physically well enough to complete an electronic questionnaire on a PDA, not literate in English. After consenting to participate in the study, the physician entered the primary headache diagnosis, based on the ICHD-II criteria, and the average monthly headache frequency over the prior 3 months (either <15 days, or ≥15 days). Per the instruction of the ICHD-II manual we had physicians diagnose antecedent migraine type in addition to other primary headache types. Given that the criteria for chronic migraine were under debate at the time of this study, we defined chronic as ≥15 days of headache per month, rather than use the “complications of migraine” criteria. Female control subjects were simultaneously recruited through advertising (via flyers and institution-wide email) to the institutions’ employees. Volunteers, contacting study personnel, who matched the age of the cases (based on 5-year group intervals) were screened for enrollment using the following questions to determine their eligibility for the nonmigraine control group: (1) has a doctor ever told you that you have migraine headaches? If “No” then, (2) have you had 5 or more headache lasting >4 hours? If “No,” the person was eligible. If “Yes,” the following 3 questions were asked: (i) has a headache limited your activities for a day or more? (ii) are you nauseated or sick to your stomach when you have a headache? (iii) does light bother you when you have a headache? A negative response to all 3 questions rendered the person eligible for participation as a control. All participants (cases and controls) were provided with an IRB-approved information sheet that described the purpose of the study, voluntary nature of the study, study aims, methods, and population, and were given opportunity to ask questions. The physician or the study personnel provided the participants with verbal instructions and a brief demonstration to complete the remainder of the survey. A written instruction sheet was also provided to the participants. All surveys were completed on the university campus in a quiet setting.
Study Questionnaire.— All participants completed the electronic questionnaire that collected information on age, race, education, menstrual abnormalities including menorrhagia, dysmenorrhea, premenstrual syndrome, interference of menstrual periods with activities of daily living, nonmenstrual bleeding symptoms, and family history of bleeding disorders. The questionnaire also inquired about the following self-reported physician-diagnosed conditions (“Have you ever been diagnosed by a doctor with ….”): hypertension, diabetes mellitus, elevated serum cholesterol, coronary artery disease, transient ischemic attack (TIA) or stroke, atherosclerosis, peripheral vascular disease, blood clot in arms, legs or lungs, Raynaud's disease, rheumatoid arthritis or systemic lupus erythematosus, irritable bowel syndrome (IBS), fibromyalgia (FM), chronic fatigue syndrome (CFS), interstitial cystitis (IC), depression, and anxiety. The response options included “yes,”“no,”“unknown,” or “I choose not to answer.” Participants were asked if they had been diagnosed with EM based on laparoscopy, and only positive responses were counted as having EM. Participants were also asked about cigarette smoking, including amount and duration. The questionnaire also inquired about use of nonsteroidal anti-inflammatory drugs (NSAIDs) and cyclooxygenase inhibitors. Participants with migraine were asked about age at onset of headache, frequency of headache, temporal relationship of headache with menstrual period, and headache-related disability.
Measures.— Headache-related disability was measured using the Headache Impact Test™ (HIT-6), a 6-item scale that correlates well with headache severity, and has been determined to be reliable and valid in evaluating the impact of headache on health-related quality of life in patients seeking primary and headache subspecialty care.15 HIT-6 produces a score ranging from 36 to 78. Severity of impact was graded as “little or no impact” for score ≤49, “some impact” for 50–55, “substantial impact” for 56-59, and “very severe impact” for a score ≥60.
Interference of menstrual periods with 7 activities of daily living was assessed using a Likert scale.16,17 A scale of 0 to 10 was used with “0” being that menses “does not interfere” and “10” being that menses “completely interferes” with the daily activity. A global menstrual interference score was derived and examined for differences between the groups. Women who were currently having menstrual cycle with periods were questioned if they experienced physical and emotional symptoms of premenstrual dysphoric disorder (PMDD) in the last year, during at least 5 of their menstrual periods. Based on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) criteria,18 PMDD was defined as having 5 (or more) of the 11 specified symptoms during the week before the onset of period and that then resolved within a few days of the beginning of the period (Table 4). Diagnosis is considered provisional if not confirmed prospectively by daily rating for at least 2 consecutive menstrual cycles.
Table 4.—. Differences in the Symptom Prevalence of Premenstrual Dysphoric Disorder in Women With Currently Having Menstrual Cycle With Periods*
| 1.||Marked depression, or feelings of hopelessness or worthlessness||51 (48)†||6 (43)||44 (51)†||15 (22)|
| 2.||Marked anxiety, tension, or feelings of being “keyed up” or “on edge”||71 (67)†||10 (71)†||57 (66)†||22 (32)|
| 3.||Feeling suddenly sad or tearful or increased sensitivity to rejection||71 (67)†||10 (71) ||57 (66)†||33 (48)|
| 4.||Persistent and marked anger or irritability or increased conflict with others||71 (67)†||10 (71)†||58 (67)†||22 (32)|
| 5.||Decreased interest in usual activities (eg, work, school, friends, hobbies)||55 (52)†||11 (79)‡||42 (48)†||13 (19)|
| 6.||Difficulty concentrating||56 (53)†||8 (57)†||46 (54)†||15 (22)|
| 7.||Marked lack of energy and easy fatigability||74 (70)†||11 (79)†||61 (70)†||35 (51)|
| 8.||Marked change in appetite, over eating, or specific food craving||76 (72) ||10 (71) ||62 (72) ||43 (62)|
| 9.||Changes in sleep pattern, oversleeping, or difficulty sleeping||59 (56)†||10 (71)†||47 (55)†||25 (36)|
|10.||Sense of being overwhelmed or out-of-control||53 (50)†||10 (71) ||42 (48)†||15 (22)|
|11.||Physical symptoms, such as breast tenderness or swelling, joint muscle pain, sensation of bloating, or weight gain||90 (85) ||13 (93) ||73 (84) ||54 (78)|
|Provisional diagnosis of premenstrual dysphoric disorder§||77 (47)†||12 (86)†||63 (72)†||27 (39)|
Data Collection.— Data from the 2 study centers were gathered using handheld computer devices. The electronic questionnaire was designed with Pendragon® Forms 3.2 software (Pendragon Software Corporation, Libertyville, IL, USA). No personal health identifiers were included in the questionnaire. The questions were formatted as forced-response items (each item requires a response to proceed to the next item) to minimize data entry errors and eliminate missing items. For certain questions such as race, education, etc, an option of “I choose not to answer” was provided for individuals who wanted to withhold this information. For questions on comorbid diagnoses, an option of “unknown” was also included for subjects who were not sure whether they had received a diagnosis by a physician. The average time taken for completion of the survey was 15 minutes. The physician or the study personnel entered a security code to indicate completion of the survey and for secure data transmission. Data were uploaded to secure central database by synchronizing the PDA's with an internet-enabled computer using Pendragon® SyncServer computer software (Pendragon Software Corporation). The database was maintained with technical support at the University of Toledo, the primary center for the study. A total of 171 migraine patients completed the surveys (106 surveys from the University of Toledo, and 65 surveys from the Duke University Medical Center). One hundred and four nonheadache female university based controls completed the survey.
Statistical Analysis.— Analysis of categorical data was done using χ2 or Fisher's exact test and continuous data were analyzed using independent sample t-test. For data with deviation from normality, including the menstrual interference and HIT-6 scores, nonparametric tests were utilized. Logistic regression analysis was performed with models adjusted for significant covariates. Odds ratios (OR) with 95% confidence intervals (CI), were used to measure the associations between the comorbid conditions and the groups (migraine with EM, migraine without EM, and controls). Significance of the ORs was examined using Wald's χ2 statistic and Hosmer and Lemeshow test19 was used to assess the fit of the regression models. All statistical hypotheses were tested at 0.05 level of significance, and the analysis was performed using SAS version 9.1 (SAS Institute, Inc., Cary, NC, USA).
The study enrolled 171 women with migraine, 60% of whom had episodic headache, and 104 women without migraine. From the headache clinic population, approximately 10% of those asked did not participate, usually because of acute headache, or lack of time following the visit. The female control group included faculty, students, and professional and nonprofessional staff, who contacted study personnel in response to advertisement. Approximately 50% of respondents did not meet eligibility criteria, and were not enrolled. Very few persons (<1%) who started the survey did not complete it. Endometriosis was reported by 22% of the women in the migraine group and 9.6% in the control group (P < .01). Information on EM was not available in 8 migraineurs (patients marked either “unknown” or “I choose not to answer” in the questionnaire), but was available in all the controls. The study population characteristics are presented in Table 1. There was no significant difference in the age, race, or the level of education between migraine and the control groups. The level of education in women with migraine with EM (migraine/EM+) was lower compared to the level in women with migraine without EM (migraine/EM–) group. More women with migraine currently smoke compared to the control group, but there was no difference between the 2 migraine subgroups based on EM diagnosis. The use of NSAIDs and cyclooxygenase inhibitors was higher in both migraine subgroups compared to the control group. There were no significant differences across the groups (migraine/EM+ vs migraine/EM– vs controls) in the frequencies of cardiovascular conditions (hypertension: 19% vs 21% vs 16%, diabetes: 17% vs 6% vs 7%, angina/myocardial infarction: 6% vs 2% vs 2%, stroke/TIA: 3% vs 5% vs none, hyper-cholestremia: 29% vs 21% vs 24%), Raynaud's disease (none vs 6.3% vs 5.8%), and rheumatoid arthritis or systemic lupus erythematosus (9% vs 5% vs 1%).
Table 1.—. Study Population Characteristics*
| Mean ± SD||37.6 ± 10||38.5 ± 8||37.6 ± 11||40.6 ± 12|
| Caucasian||133 (79)||31 (86)||98 (78)||81 (78)|
| African American|| 22 (13)||3 (8)||18 (14)||16 (15)|
| Other||14 (8)||2 (6)||9 (7)||7 (7)|
| Not a high school graduate|| 5 (2)||2 (6)||2 (2)|| 1 (<1)|
| High school graduate|| 53 (31)||11 (31)||40 (31)||23 (22)|
| Undergraduate|| 72 (43)||19 (53)||48 (38)||54 (52)|
| Masters or postgraduate|| 39 (23)||3 (9)||36 (28)||26 (25)|
| Current|| 39 (24)‡|| 8 (23)||28 (23)||8 (8)|
| Ever|| 64 (38)||14 (40)||46 (37)||34 (49)|
|NSAIDs|| 142 (83)‡|| 30 (83)‡||107 (84)‡||61 (59)|
|Cyclooxygenase inhibitors|| 38 (22)‡|| 11 (30)‡|| 25 (20)‡||9 (9)|
Table 2 illustrates the differences in the headache profiles between migraine/EM+ and migraine/EM– groups. The mean age at headache onset was similar between the groups. Frequency of chronic headache was higher in the migraine/EM+ group as was headache-related disability. Similar proportions of migraineurs with and without EM reported to have headaches that occurred regularly with their menstrual periods.
Table 2.—. Differences in Headache Characteristics*
|Age at headache onset (mean ± SD), y||22 ± 9||20 ± 10|
| <15 Days/month||13 (36)||82 (65)|
| ≥15 Days/month||23 (64)||45 (35)|
|HIT 6 score (median, range)‡||66, 46–78||64, 42–78|
| Very severe disability (score >60)||29 (81)||94 (74)|
| Headaches regularly with menstrual periods||22 (65)||72 (61)|
Table 3 illustrates the differences in gynecological history between migraine/EM+, migraine/EM–, and control groups. Of the women currently having menstrual periods, the impact of the menses on their lives was greater in the migraine/EM+ group than in the migraine/EM– group or the control group. Each of the activities that make up the global menstrual interference score were more affected in the migraine/ EM+ group than the migraine/EM– cohort: general activity (P < .001), ability to work or go to school (P < .001), participation in normal family activities (P < .001), sleep (P < .001), ability to enjoy life (P < .001), mood (P= .005), and overall quality of life (P < .001).
Table 3.—. Differences in the Gynecological History*
|Currently having menstrual periods||106 (62)||14 (39)†||87 (69)||73 (65)|
|Hysterectomy history||43 (25)||17 (47) ||26 (21)||27 (26)|
|Menorrhagia history||116 (68)‡||33 (92)†|| 75 (59)‡||45 (43)|
|In previously menstruating women|| 53 (31)‡||22 (61)†||28 (22)||21 (20)|
|Dysmenorrhea history|| 91 (54)‡||33 (92)†|| 55 (44)‡||24 (23)|
|Menstrual bleeding >7 days history|| 81 (47)‡||26 (72)†|| 48 (38)‡||18 (17)|
|Infertility history||39 (24)||14 (42)†||24 (24)||14 (19)|
|Global score for impact of menses (mean ± SD, median), in those currently having menstrual periods||26 ± 19, 22‡||40 ± 20, 41†||21 ± 16, 18‡||16 ± 15, 12|
Women with migraine were more likely to endorse symptoms consistent with premenstrual dysphoric disorder than controls, with 9 of 11 symptoms being more frequent (Table 4). Within the migraine cohort, only the symptom of “decreased interest in usual activities (eg, work, school, friends, hobbies)” was more common in the women with EM.
Migraineurs with EM more frequently reported diagnoses of comorbid conditions compared to the control group (Table 5). In separate logistic regression models, the migraine/EM+ group was more likely to report interstitial cystitis (OR = 10.6, 95% CI 1.9-56.5), chronic fatigue syndrome (OR = 3.6, 95% CI 1.1-11.5), and anxiety (OR = 2.2, 95% CI 1.0-4.7).
Table 5.—. Association of Migraine and Endometriosis With Comorbid Conditions*
|Irritable bowel syndrome||12 (33)†||4.8 (1.9–12.4)||29 (23)†||2.7 (1.2–5.7)||11 (11)||1.0|
|Fibromyalgia|| 5 (15)†||4.6 (1.2–17.9)||11 (9) ||2.5 (0.8–7.7)||5 (5)||1.0|
|Interstitial cystitis|| 5 (15)‡||17.8 (1.9–158.1)|| 2 (1.6)|| 1.7 (0.7–19.1)||1 (1)||1.0|
|Chronic fatigue syndrome|| 6 (18)‡||15.8 (2.7–90.9) ||7 (6)|| 4.1 (0.8–20.9)||2 (2)||1.0|
|Depression||19 (56)†||6.9 (2.9–16.6)||60 (48)†||5.2 (2.7–9.9)||16 (15)||1.0|
|Anxiety||19 (54)‡||12.6 (4.8–32.8) ||43 (35)†|| 5.7 (2.6–12.6)||9 (9)||1.0|
|PMDD§||12 (86)†||9.8 (2.0–48.0)||63 (72)†||4.5 (2.3–9.1)||27 (39)||1.0|
While our findings underscore established relationships of migraine with various psychiatric conditions (anxiety, depression),20–22 and syndromatic medical disorders (FM, IBS, CFS),5,8,9,23 we found that these associations are more robust in the migraine subgroup with EM. Premenstrual syndrome is reported to occur in two-thirds of women with migraine,24 yet our finding that that PMDD is comorbid with migraine and affects 86% of migraineurs with EM is novel. There are only a few studies of the comorbidity of headache and interstitial cystitis.25,26 In our study the relationship appears to hinge entirely on the presence of EM. There were also differences between the migraine subgroups for chronic fatigue syndrome and anxiety. The EM subgroup was less educated, despite similar age of headache onset, suggesting that the differences in educational levels may be associated with EM or with other comorbid conditions. A large cross-sectional survey of women with EM demonstrated that compared to the general US female population, women with EM were better educated, more likely to be Caucasian, and had higher household income.10 Migraine was not surveyed in that study, but the authors report an association of EM with fibromyalgia and chronic fatigue syndrome.
Some of the differences between the migraine subgroups were expected, including the higher proportion of dysmenorrhea, infertility, prolonged, disabling menses, and menorrhagia in those with EM. The reason for increased frequency of menorrhagia in migraine is uncertain, but our previous work in a female migraine population suggested a systemic bleeding problem with increased bruisability, rectal bleeding, and excessive bleeding from minor trauma, and a trend toward history of epistaxis and surgically related bleeding.11 The role of the platelet has been well studied in migraine, largely with regards to aggregation and serotonin metabolism,27 but studies reporting increased bleeding risk are limited,11,28 and mechanisms thereof remain uncertain. Migraineurs with EM also reported more frequent and disabling headaches. The reports of menstrual related migraine were similar between the groups, and in keeping with other reports.29
The underlying reason for the comorbidity of migraine with depression, anxiety, PMDD, FM, IBS, and CFS remains uncertain, although dysfunction of the serotonergic system has been proposed.30 Recognizing the response of these conditions to serotonin-influencing antidepressants, this complex of syndromes has been referred to as the affective spectrum disorder.31 These conditions are all clinically defined by symptoms related to pain and mood, and still lack objective biomarkers. Endometriosis, a painful condition associated with migraine in our study, has underlying objective pathological characteristics, thus offering an opportunity for further study of underlying pathophysiological links of disorders in the complex.
Although the pathogenesis of EM is poorly understood, Sampson's theory of retrograde menstruation into the peritoneal cavity is well accepted.32 Menorrhagia, which occurs with increased frequency in migraine,11 might result in an increase of cells reaching the peritoneum. Alterations in the immunological milieu allowing for adherence, implantation and proliferation of endometrial cells on the peritoneal surface are likely also necessary. Angiogenic cytokines, including TGFB-133,34 and VEGF35 are hypothesized to play a critical role in the pathogenesis of EM, possibly by stimulating matrix metalloproteinases (MMPs),36 enzymes synthesized in and released from the vascular endothelium. MMPs mediate normal tissue turnover, including endometrial breakdown at menstruation. Menstruating tissue is rich in MMP, and MMP gene expression in ectopic endometrium has been found to be much greater than in eutopic endometrium37 and may facilitate implantation. In EM, these substances have been identified locally as well as in the systemic circulation.38 The same is true in migraine,39 although effects on distant organ systems have not been studied. The co-occurrence of EM, migraine, and pain conditions support the hypothesis that these are linked through the phenomenon of central sensitization, which may be precipitated or augmented by EM.13
Other evidence of a neuro-immunoendocrine link between EM and migraine, fibromyalgia, irritable bowel syndrome, chronic fatigue syndrome, interstitial cystitis, and mood disorder involves the finding of increased mast cell activation.40–43 Mast cell activation without allergic degranulation has been documented to occur in response to stress44,45 and lead to painful sterile inflammatory states.46
Our study has certain limitations. The study was performed in headache clinics and the results may not be generalized outside a subspecialty clinic population. Migraine was diagnosed by a headache specialist, but the comorbid conditions were self-reported, and there may be errors related to this, including recall bias. We cannot exclude reporting bias, whereby measurement of 1 variable influences the reporting of a second outcome variable independent of a true association. Although we attempted to enroll consecutive patients, availability of research personnel determined the sampling. The diagnosis of PMDD is provisional given that it was not confirmed prospectively by daily rating for at least 2 consecutive menstrual cycles. As a cross-sectional study we cannot ascertain influence of one condition on another over time. Controls were from the university community rather than from a clinic population. The prevalence of EM in our controls was, however, similar to that reported in the general population of women of childbearing age.4 Lastly, this is a small study and we may have lacked statistical power to identify between group differences.
In conclusion, there is an increased prevalence of EM in migraine, and the subgroup of migraineurs with EM is more likely to have other comorbid conditions affecting mood and pain. Migraineurs with EM have more frequent and disabling headaches. Better understanding of the migraine–endometriosis association may contribute to our understanding of the pathogenesis of both conditions.
Conflict of Interest: None