Financial Support: The American Migraine Prevalence and Prevention (AMPP) Study was funded through a research grant to the National Headache Foundation from McNeil-Janssen Scientific Affairs LLC, Raritan, NJ. The AMPP database was donated by McNeil-Janssen Scientific Affairs LLC to the National Headache Foundation for use in various projects. Additional funding and support for this manuscript was provided through a grant to the National Headache Foundation by ENDO Pharmaceuticals, Chadds Ford, PA.
Conflict of Interest: Dawn C. Buse, PhD, has received grant support and honoraria from Allergan Pharmaceuticals, MAP Pharmaceuticals, Novartis and NuPathe. Elizabeth W. Loder, MD, MPH, reports receiving salary support from the British Medical Journal in exchange for services as a clinical editor. Jennifer A. Gorman, MS, has nothing to disclose. Walter F. Stewart, PhD, received grants from Allergan Pharmaceuticals, Endo Pharmaceuticals, GlaxoSmithKline, Minster, Merck, Inc., Neuralieve, Novartis, OrthoMcNeil, Pfizer, the National Headache Foundation, and the National Institutes of Health. Michael L. Reed, PhD, has received funding for research and data analysis from Allergan Inc., CoLucid, Endo Pharmaceuticals, GlaxoSmithKline, MAP Pharmaceuticals, Merck, Novartis, NuPathe, Ortho-McNeil, and the National Headache Foundation. Kristina M. Fanning, PhD, has received research support from Allergan Inc., CoLucid, Endo Pharmaceuticals, GlaxoSmithKline, MAP Pharmaceuticals, Merck & Co., Inc., NuPathe, Novartis, Ortho-McNeil, and the National Headache Foundation. Daniel Serrano, PhD, has received research support from Allergan Inc., CoLucid, Endo Pharmaceuticals, GlaxoSmithKline, MAP Pharmaceuticals, Merck & Co., Inc., NuPathe, Novartis, Ortho-McNeil, and the National Headache Foundation. Richard B. Lipton, MD, receives research support from the NIH [PO1 AG03949 (Program Director, Project and Core Leader), RO1AG025119 (Investigator), RO1AG022374-06A2 (Investigator), RO1AG034119 (Investigator), RO1AG12101 (Investigator), K23AG030857 (Mentor), K23NS05140901A1 (Mentor), and K23NS47256 (Mentor)], the National Headache Foundation, and the Migraine Research Fund; serves on the editorial board of Neurology, has reviewed for the NIA and NINDS, holds stock options in eNeura Therapeutics; serves as consultant, advisory board member, or has received honoraria from: Allergan, American Headache Society, Autonomic Technologies, Boehringer-Ingelheim Pharmaceuticals, Boston Scientific, Bristol Myers Squibb, Cognimed, Colucid, Eli Lilly, ENDO, eNeura Therapeutics, GlaxoSmithKline, Merck, Novartis, NuPathe, Pfizer, and Vedanta.
Address all correspondence to D. Buse, Montefiore Medical Center – Montefiore Headache Center, 1575 Blondell Avenue Suite#225, Bronx, NY 10471, USA, email: email@example.com
The strikingly higher prevalence of migraine in females compared with males is one of the hallmarks of migraine. A large global body of evidence exists on the sex differences in the prevalence of migraine with female to male ratios ranging from 2 : 1 to 3 : 1 and peaking in midlife. Some data are available on sex differences in associated symptoms, headache-related disability and impairment, and healthcare resource utilization in migraine. Few data are available on corresponding sex differences in probable migraine (PM) and other severe headache (ie, nonmigraine-spectrum severe headache). Gaining a clear understanding of sex differences in a range of severe headache disorders may help differentiate the range of headache types. Herein, we compare sexes on prevalence and a range of clinical variables for migraine, PM, and other severe headache in a large sample from the US population.
This study analyzed data from the 2004 American Migraine Prevalence and Prevention Study. Total and demographic-stratified sex-specific, prevalence estimates of headache subtypes (migraine, PM, and other severe headache) are reported. Log-binomial models are used to calculate sex-specific adjusted prevalence ratios and 95% confidence intervals for each across demographic strata. A smoothed sex prevalence ratio (female to male) figure is presented for migraine and PM.
One hundred sixty-two thousand seven hundred fifty-six individuals aged 12 and older responded to the 2004 American Migraine Prevalence and Prevention Study survey (64.9% response rate). Twenty-eight thousand two hundred sixty-one (17.4%) reported “severe headache” in the preceding year (23.5% of females and 10.6% of males), 11.8% met International Classification of Headache Disorders-2 criteria for migraine (17.3% of females and 5.7% of males), 4.6% met criteria for PM (5.3% of females and 3.9% of males), and 1.0% were categorized with other severe headache (0.9% of females and 1.0% of males). Sex differences were observed in the prevalence of migraine and PM, but not for other severe headache. Adjusted female to male prevalence ratios ranged from 1.48 to 3.25 across the lifetime for migraine and from 1.22 to 1.53 for PM. Sex differences were also observed in associated symptomology, aura, headache-related disability, healthcare resource utilization, and diagnosis for migraine and PM. Despite higher rates of migraine diagnosis by a healthcare professional, females with migraine were less likely than males to be using preventive pharmacologic treatment for headache.
In this large, US population sample, both migraine and PM were more common among females, but a sex difference was not observed in the prevalence of other severe headache. The sex difference in migraine and PM held true across age and for most other sociodemographic variables with the exception of race for PM. Females with migraine and PM had higher rates of most migraine symptoms, aura, greater associated impairment, and higher healthcare resource utilization than males. Corresponding sex differences were not observed among individuals with other severe headache on the majority of these comparisons. Results suggest that PM is part of the migraine spectrum whereas other severe headache types are not. Results also substantiate existing literature on sex differences in primary headaches and extend results to additional headache types and related factors.
International Classification of Headache Disorders
Migraine Disability Assessment Questionnaire
With few exceptions, it is well established that the majority of primary headache disorders have a higher prevalence in females than males. A review of global population estimates of primary headache subtypes of 107 studies from 6 continents reported prevalence of 42% for tension-type headache, 11% for migraine, and 3% chronic daily headache (3%). Although the report found differences in the prevalence of headache across continents, all three of these headache types were more prevalent among females compared to males on every continent. Female to male sex prevalence ratios (PRs) are most dramatic in migraine and chronic daily headache but also exist in tension-type headache. In fact, the only primary headache types that have not demonstrated a female preponderance are the trigeminal autonomic cephalalgias. The majority of these headache types are more common in men, especially cluster headache, which has female to male sex prevalence estimates ranging from 1 : 3.5 to 1 : 7.
Elevated female to male sex PRs in migraine have been reported from studies around the world with a variety of samples and methodologies.3-23 The female to male sex PR for migraine has consistently varied across the lifespan ranging from 3 or 4 to 1 in midlife and lowering to 2 to 1 or less at both ends of the age spectrum. In addition to the female preponderance in migraine prevalence, some studies have reported that females may experience greater symptomology and headache-related disability.[3, 4, 8, 19, 24, 25] Sex differences have also been observed in the prevalence of probable migraine (PM), although the direction is not always consistent.[5, 9, 26, 27] Few data are available on sex differences in associated symptomology, frequency, and disability in PM. (See MacGregor et al, 2011, Smitherman et al, 2013, and Merikangas, 2013 for detailed reviews of sex-related differences in migraine and other headache types.)
Several large scale studies have reported sex prevalence differences in migraine, including the American Migraine Study (AMS) I and II[7, 8] and the American Migraine Prevalence and Prevention (AMPP) Study. In 1989, a self-administered questionnaire was sent to 15,000 households as part of the AMS I. Questionnaires collected data on sociodemographics, headache symptomology, frequency, and related disability among other topics. Of 20,468 respondents, 17.6% of females and 5.7% of males were found to have one or more migraine headaches per year (a 3 to 1 female to male sex PR). Researchers also found that females with migraine had more frequent attacks than males but the sexes did not differ substantially in terms of headache-related disability. In 1999, 20,000 households were surveyed as part of the AMS II.[7, 8] Of 29,727 respondents, the prevalence of migraine was 18.2% among females and 6.5% among males. Although the reported frequency of severe headache pain was similar for female and male migraineurs, females were somewhat more likely to report “severe impairment” during migraine, longer duration of impairment, and were more likely to report photophobia, phonophobia, unilateral pain, nausea, vomiting, blurred vision, and aura associated with headache.
In 2004, the AMPP Study collected data from 120,000 US households and assessed headache symptomology, frequency, headache-related disability, and other data. Surveys asked about “severe headache” and second edition of International Classification of Headache Disorders (ICHD-2) criteria, which were applied to determine the 3 most severe headache types experienced by respondents. Data were received from 162,756 individuals aged 12 and older to determine the consistency of sex-specific patterns across 3 defined subgroups of “severe” headache including migraine, PM, and other (ie, nonmigraine spectrum) severe headache. Previous analyses of AMPP Study data have revealed sex differences in migraine and PM prevalence.[27, 31] The prevalence of migraine was found to be 17.1% in females and 5.6% in males, and PM was 5.1% in females and 3.9% in males. However, data on sex differences in symptomology, aura, headache-related disability, health resource utilization, and comorbidities for individuals meeting criteria for migraine, PM, and other types of “severe” headache have not yet been systematically reported from the 2004 AMPP Study sample.
In the current study, sex-specific patterns on multiple variables were examined in the 2004 AMPP Study population overall as well as by major racial subgroups, age groups, and by other sociodemographic variables. We report the sex-specific prevalence of migraine, PM, and other severe headache. We also report the sex-specific occurrence of headache symptoms, reported aura, headache-related disability, healthcare resource utilization (eg, consulting behavior, emergency department/urgent care clinic visits for headache), and healthcare professional (HCP) assigned headache diagnoses, among other variables, by headache subtype. In addition, we will examine prevalence and female to male PRs by age, race, and annual household income for each of the 3 types of severe headache.
Study Design and Sample.—
The AMPP Study is a 2-phase, longitudinal, population-based investigation. The AMPP Study was modeled on the methods of the AMS I and II.[7, 8, 20] In 2004, a self-administered questionnaire was mailed to a stratified random sample of 120,000 US households, drawn from a 600,000 household nationwide panel maintained by TNS, a survey sampling firm. TNS maintains panels comprised of sampling blocks, each containing 5000 households, which are constructed to represent the US population in terms of census region, population density, age of head of household, household income, and number of household members.
Sociodemographic and census information is obtained from each household during initial contact and is routinely updated by the survey company. In phase 2, a random sample of 24,000 adults (≥18 years of age) was selected from the 28,261 respondents to the 2004 survey with complete data and who reported experiencing at least one “severe” headache in the preceding year. These participants were asked to complete annual surveys from 2005 to 2009. The current analyses are derived from the 2004 survey data as this population-based sample allows for the calculation of prevalence estimates. Respondents aged ≥12 who endorsed having at least one “severe” headache in the preceding year were included in analyses. This study was approved by the Albert Einstein College of Medicine Institutional Review Board.
2004 AMPP Study Survey.—
The 2004 AMPP Study survey instructed the head of household to report the total number of household members and the number of household members who “suffer from severe headaches.” Each household member (up to a total of 3) with “severe headaches” was asked to complete the remainder of the questionnaire, which included sociodemographic characteristics (age, sex, height, and weight), headache frequency, associated symptoms, and other information necessary to assign an ICHD-2 diagnosis, headache-related disability, average pain of “severe headaches,” age of onset of severe headache, headache-related emergency department/urgent care clinic visits, acute and preventive headache medication use, and physician diagnosis of headache subtype.
Associated Symptoms, Headache Frequency, and Assignment of Headache Diagnosis.—
Respondents were asked to report the average number of days that they experienced headache in a week, month, or year. They were also asked if they experienced pain-free intervals between attacks. Respondents were asked if they experienced symptoms with “severe” headache including nausea; vomiting; unilateral head pain; pulsating or throbbing pain; sensitivity to light; sensitivity to noise; blurring of vision in association with headaches; presence of shimmering lights, circles, other shapes, or colors before the eyes before the start of the headache; and presence of numbness of lips, tongue, fingers, or legs before the start of the headache. Respondents were asked to report average pain intensity of severe headaches as: extremely severe pain, severe pain, moderately severe pain, or mild pain. Responses to these items were used to assign headache type based on the ICHD-2. Use of these items to assign a diagnosis was validated in a population sample of subjects with migraine and other types of headache. The items exhibited a sensitivity of 100% and specificity of 82.3% for the diagnosis of migraine. Although this diagnostic module was not revalidated using ICHD-2 criteria, the migraine criteria remained essentially unchanged relative to ICHD-1 criteria.
Migraine and PM diagnoses were derived by applying modified ICHD-2 criteria. Respondents satisfied Criterion 1 if they reported one or more of the following associated with headache: severe or extremely severe pain, unilateral headache, or pulsatile or throbbing pain. Respondents satisfied Criterion 2 if they reported one or more of the following associated with headache: nausea or vomiting, photophobia and phonophobia, or visual or sensory aura. A migraine diagnosis was assigned if a respondent met both Criteria 1 and 2. A diagnosis of PM was assigned if a respondent met either Criterion 1 or 2. If neither of these criteria were met, respondents were assigned with “other severe headache.” This group may logically capture cluster headache, tension-type headache, and other nonmigrainous forms of headache that respondents subjectively rated as “severe.”
Headache-related disability was assessed with the Migraine Disability Assessment Questionnaire (MIDAS). The MIDAS is a self-administered 5-item questionnaire that assesses days of missed activity or substantially reduced activity due to headache in the preceding 3 months in 3 domains: schoolwork/paid employment, household work or chores, and nonwork (family, social, and leisure) activities. Responses are summed and fall into 1 of 4 grades of headache-related disability: little or none (0-5), mild (6-10), moderate (11-20), or severe (21-40). Respondents were also asked how they are “usually affected by severe headaches” with the following response options: able to work/function normally, working ability or activity impaired to some degree, working ability or activity severely impaired, and bed rest required. Additionally, respondents were asked “Each time you have a severe headache, how long are you unable to work or undertake normal activities?” with options ranging from 0 days to ≥6 days.
Medication Use and Healthcare Resource Utilization.—
Respondents were asked how they usually treat their severe headaches, with the following response options: (1) take nonprescription medications; (2) take prescription medications; (3) take both prescription and nonprescription medications; or (4) take no medications. Respondents were asked if they have ever taken a prescription medication on a daily basis for headache prevention, and if so, if they are currently taking it or when they had taken it in the past. Finally, respondents were asked if they were taking any other daily medications, with the following response options: (1) water pill or prescription diuretic for high blood pressure; (2) prescription medicine (other than water pill) for high blood pressure; (3) prescription medicine for seizures, epilepsy, or fits; (4) prescription medicine for diabetes; (5) prescription medicine for cholesterol; (6) prescription medicine for depression or anxiety. Responses to this item were used to estimate comorbid conditions.
Respondents were asked to report prior headache diagnoses from a physician or other HCP from the following response options: tension headaches, sinus headaches, cluster headaches, stress headaches, “sick” headaches, and migraine headaches. Respondents were asked to report age of first migraine diagnosis; belief that their most severe headache was migraine; number of household members with migraine; and usual headache treatment with response options including prescription medication, nonprescription medication, both and neither. Respondents were asked if they had ever used a prescription preventive medication for headache on a daily basis and if they had ever used the emergency department or urgent care facility for severe headache.
Sociodemographic data including race, region of the country, population density, and annual household income were collected by TNS during enrollment. Age, sex, height, and weight were gathered from self-report in the questionnaire. For the majority of analyses, respondents were divided into 6 age categories (12-17, 18-29, 30-39, 40-49, 50-59, ≥60), 4 racial groups (Caucasian, African American, “other,” or unknown if data were missing), 4 population density groups (<100,000, 100,000-499,999, 500,000-1,999,999, ≥2,000,000), 5 annual household income groups (<$22,500, $22,500-$39,999, $40,000-$59,999, $60,000-$89,999, ≥$90,000), and 9 US regions (New England, Middle Atlantic, South Atlantic, East North Central, West North Central, East South Central, West South Central, Mountain, and Pacific). Average headache frequency (reported as the number of days with headache) in a week, month, or year was obtained from respondents in free text. For the purposes of analyses, headache frequency was categorized as days per month within the following categories: <1 per month, 1-4 month, 5-9 per month, and ≥10 per month.
One hundred twenty thousand households, containing a total of 257,339 household members, were contacted to participate in the AMPP Study survey. Surveys were returned by 77,879 households (64.9% response rate) yielding data for 162,756 individual household members aged ≥12 years old (Table 1). Respondents were primarily female (52.6%, N = 85,571) and Caucasian (86.6%, N = 140,948). Response rates did not differ substantially between males (62%) and females (64%), but were higher in Caucasians (65%) than in African Americans (56%, P < .01) and in those aged ≥50 years old (P < .01). Response rates did not differ significantly by geographic region, population density, or annual household income. For additional details on comparisons between respondents and nonrespondents, see Lipton et al. It was not possible to assess reasons for nonparticipation.
Table 1. Total Population Surveyed and Respondents (2004 AMPP Study Survey) by Sociodemographic Characteristics
Total Population Surveyed (N = 257,339)
Total Population Surveyed (%)
Survey Respondents (N = 162,756)
Response Rate (%)
Asian, Pacific Islander
East North Central
West North Central
East South Central
West South Central
Thirty thousand seven hundred twenty-one respondents reported experiencing “severe headache,” of whom 430 were missing sociodemographic data and were not included in the AMPP Study cohort (ie, they were not included in these or any other analyses from the AMPP Study data set). This resulted in 30,291 respondents with “severe headache” of whom 28,261 reported experiencing “severe headache” in the preceding year. Of these, 19,189 respondents (11.8%) met criteria for migraine, 7485 (4.6%) met criteria for PM, and 1587 (1.0%) reported experiencing “severe” headache that did not meet criteria for migraine or PM (ie, other severe headache).
Unadjusted, Sex-Specific Prevalence of Migraine, PM, and Other Severe Headache by Sociodemographic Variables.—
23.5% of females and 10.6% of males reported experiencing “severe headache” in the preceding year (Table 2). The unadjusted prevalence of migraine and PM was higher among females than males, whereas the prevalence of other severe headache was similar between sexes. 17.3% of females and 5.7% of males met criteria for migraine, 5.3% of females and 3.9% of males met criteria for PM, and 0.9% of females and 1.0% of males reported headache which was classified as other severe headache. Prevalence of migraine and PM were highest in midlife for both sexes. Among those aged 30-39, the unadjusted prevalence of migraine was 28.4% in females and 9.1% in males. In the same age group, unadjusted prevalence of PM was 6.8% in females and 5.2% in males. Prevalence of other severe headache was fairly consistent across the lifespan, ranging from 0.4% during adolescence to 1.2% among persons age ≥60 for both sexes.
Table 2. Unadjusted Sex-Specific Prevalence of Migraine, Probable Migraine, and Other Severe Headache by Sociodemographic Variables
Total Sample N = 162,756
Females (N = 85,571; 52.6%)
Males (N = 77,185; 47.4%)
“Severe headache”: No
“Severe Headache”: Yes
Other Severe Headache (N = 805, 0.9%)
Probable Migraine (N = 4493, 5.3%)
Migraine (N = 14,775, 17.3%)
Other Severe Headache (N = 782, 1.0%)
Probable Migraine (N = 2992, 3.9%)
Migraine (N = 4414, 5.7%)
Annual Household Income
Within race, the unadjusted prevalence of migraine was higher than PM for all races in both sexes with one exception. The prevalence of PM was slightly higher than migraine prevalence among African American males (Table 2). Between races, unadjusted prevalence rates for migraine were highest in females among the “other” racial category (ie, not Caucasian or African American) (19.3%) followed by Caucasian females (17.5%), whereas unadjusted prevalence rates of PM were highest among African American females (7.6%) compared with 5.0% of Caucasian females. The same pattern held true for males. Rates of migraine were highest in the “other” racial category (6.9%), and rates of PM were highest among African American males (4.9%) compared with Caucasian males (3.7%). The combined prevalence of migraine and PM was similar for Caucasians and African Americans for both sexes (females: Caucasians 22.6%, African Americans 21.6%; males: Caucasians 9.5%, African Americans 9.2%), demonstrating that the total migraine-spectrum (including PM) prevalence is similar between the 2 groups.
Unadjusted prevalence of migraine and PM was inversely related to annual household income for both sexes (Table 2) and the number of family members living in a household (data not shown). Prevalence of migraine was highest for both females (20.6%) and males (9.1%) among individuals who live in households with an annual average income of ≤$22,500 per year. For both sexes, prevalence decreased as household income increased. In households with an annual income ≥$90,000, 13.6% of females and 4.2% of males met criteria for migraine. A similar pattern was observed in the prevalence of PM for both sexes.
Adjusted, Sex-Specific PRs of Migraine, PM, and Other Severe Headache by Sociodemographic Variables.—
Compared with persons aged 12-17 (the reference group), PRs for migraine were highest for both males and females in the 30 to 39-year-old age group. Females in this age group were 3.8 times more likely (PR = 3.80, 95% CI = 3.47-4.15), and males were 1.7 times more likely (PR = 1.72, 95% CI = 1.53-1.94) to have migraine compared with teenage respondents (Table 3). Individuals aged ≥60 were significantly less likely to have migraine than those in adolescence (females: PR = 0.77, 95% CI = 0.70-0.85; males: PR = 0.36, 95% CI = 0.31-0.42). A similar pattern was observed for PRs of PM by age for both sexes. Individuals in their 30s and 40s had the highest rates of PM. Other severe headache was more likely at all ages compared with the 12-17 year age group for both sexes and generally increased over the lifespan. However, absolute differences with age were small.
Table 3. Sex-Specific Adjusted Prevalence Ratios by Headache Subtype†
Females (N = 85,571)
Males (N = 77,185)
Other Severe Headache PR (95% CI)
Probable Migraine PR (95% CI)
Migraine PR (95% CI)
Other Severe Headache PR (95% CI)
Probable Migraine PR (95% CI)
Migraine PR (95% CI)
†Adjusted prevalence ratios and 95% confidence limits were estimated using log-binomial models predicting migraine status (within sex). Adjusted for age, race, region of the country, population density, annual household income, and household size. Results are shown for age, race, and annual household income.
Annual Household Income
Within sex by race, adjusted PRs for African Americans (compared with Caucasians as the reference group) were well below 1.0 for migraine for both sexes (female: PR = 0.69, 95% CI = 0.65-0.73; male: PR = 0.65, 95% CI = 0.56-0.74), but significantly greater than 1.0 for PM for both sexes (female: PR = 1.38, 95% CI = 1.26-1.51; male: PR = 1.20, 95% CI = 1.05-1.37) (Table 3). Thus, African Americans of both sexes are less likely to have migraine but more likely to have PM than Caucasians. African Americans had higher risk for other severe headache compared with Caucasians, although this difference was only significant for females (PR = 1.39, 95% CI = 1.12-1.72).
Adjusted PRs for average annual household income were similar between sexes. Using the lowest annual household income group as the reference, both females and males in the highest income group were significantly less likely to have migraine (female: PR = 0.54, 95% CI = 0.51-0.57; male: PR = 0.45, 95% CI = 0.41-0.50) and PM (female: PR = 0.64, 95% CI = 0.58-0.70; male: PR = 0.48, 95% CI = 0.43-0.54) (Table 3). When compared with the lowest income level, the PRs for migraine and PM decreased as household income increased for both sexes. Household size revealed a similar pattern for both sexes as those in households with more members had lower risk of migraine or PM (data not shown).
Adjusted Female to Male PRs of Migraine, PM, and Other Severe Headache.—
Females had higher prevalence of migraine than males at all ages, although the differences varied across the lifespan. Female to male adjusted PRs for migraine peaked at 3.25 (95% CI = 3.00-3.52) among those aged 18-29. Prevalence of migraine was still higher among females at both ends of the age spectrum although the difference was not as pronounced, with a female to male PR during ages 12-17 of 1.48 (95% CI = 1.30-1.69) to 2.91 (95% CI = 2.62-3.22) among those aged ≥60 (Table 4). Female to male adjusted PRs for PM showed a similar pattern across the lifespan, with a peak of 1.53 (95% CI = 1.35-1.73) among those aged 18-29 to 1.35 (95% CI = 1.21-1.50) among those aged ≥60 and 1.28 (95% CI = 1.04-1.56) during adolescence. Other severe headache revealed a different pattern with a heightened prevalence among females during adolescence (female to male PR = 1.24, 95% CI = 0.73-2.11) but a male preponderance among those aged ≥18 (Table 4). The Figure shows a graph of smoothed female to male PRs for migraine and PM by age group.
Table 4. Adjusted Sex Prevalence (Female to Male) Ratios (95% CI) by Headache Subtype†
Other Severe Headache PR (95% CI)
Probable Migraine PR (95% CI)
Migraine PR (95% CI)
†Adjusted sex prevalence ratios and 95% CIs estimated using log-binomial models predicting migraine status with main covariate sex (within sociodemographic category). Adjusted for age, race, region of the country, population density, annual household income, and household size.
Annual Household Income
Adjusted female to male PRs for migraine were approximately 3 to 1 among all 3 racial groups: Caucasian female to male PR = 2.97 (95% CI = 2.88-3.08), African American female to male PR = 3.02 (95% CI = 2.62-3.49), “other” racial group female to male PR = 2.71 (95% CI = 2.31-3.19). Significantly elevated adjusted female to male PRs were also seen in PM among Caucasians (PR = 1.30, 95% CI = 1.23-1.36) and African Americans (PR = 1.48 (95% CI = 1.27-1.72). Among persons with other severe headache, the adjusted female to male PR was 0.84 (95% CI = 0.75-0.94) for Caucasians indicating a male preponderance, and nonsignificant for other racial groups indicating the lack of a significant difference between sexes. Adjusted female to male PRs for the 3 headache types are also presented by annual household income in Table 4.
Headache Symptoms and Frequency by Sex for Migraine, PM, and Other Severe Headache.—
Females with migraine were significantly more likely than males to report all ICHD-2 criteria and most symptoms commonly associated with migraine including nausea, vomiting, unilateral head pain, pulsing or throbbing pain, photophobia, phonophobia, blurred vision, and visual aura, but not sensory aura (Table 5). Females with PM were also significantly more likely than males to report most of these symptoms including nausea, vomiting, photophobia, phonophobia, and visual aura. Among those with other severe headache, significant sex differences were only seen in rates of photophobia.
Table 5. Sex-Specific Prevalence and Female to Male Prevalence Ratios of Headache Symptoms, Frequency, and Pain Intensity by Headache Subtype†
“Which statement best describes the pain of your severe headaches?”
Extremely severe pain
Moderately severe pain
Among those with migraine, the majority of respondents in both sexes (48.8% of females and 45.3% of males) reported 1-4 days of headache per month (Table 5). Females with migraine were less likely than males to endorse a frequency of <1 day with headache per month although rates only differed by 2.1% (23.5% vs 25.6%, female to male PR = 0.92, 95% CI = 0.87-0.97). The majority of individuals of both sexes with PM also endorsed 1-4 days of headache per month (50.4% of females and 47.1% of males), and females were less likely to endorse a frequency of <1 day with headache per month (21.1% vs 23.8%, female to male PR = 0.89, 95% CI = 0.81-0.96). High-frequency (HF) migraine (headache ≥10 days per month) was more common among males, occurring in 16.7% of males with migraine and 14.9% of females with migraine (female to male PR = 0.90, 95% CI = 0.83-0.97). PM shows the same pattern; 16.4% of males and 13.8% of females reported headache on ≥10 days per month.
Reports of average pain associated with severe headache did not form a clear pattern of sex differences for migraine or PM. The majority of respondents of both sexes with migraine endorsed “severe pain” associated with headache. Males with migraine were slightly more likely to endorse “extremely severe pain” whereas females were more likely to endorse “severe pain,” although absolute percentages varied by only 2%. Respondents with PM showed similar results. The majority of females with PM endorsed “moderately severe” pain and the majority of males endorsed “severe pain” associated with headache. Males with PM were slightly more likely to endorse “extremely severe pain” than females although absolute rates were only 2% different (12.8% males vs 10.8% females, female to male PR = 0.84, 95% CI = 0.74-0.95).
Headache-Related Disability and Impact by Sex for Migraine, PM, and Other Severe Headache.—
Females with migraine were 1.34 times more likely than males (12.4% vs 9.3%, 95% CI = 1.21-1.48) to have the highest level of headache-related disability (MIDAS Grade 4) (Table 6). Females were more likely than males to have moderate (PR = 1.46, 95% CI = 1.31-1.63) or mild (PR = 1.46, 95% CI = 1.33-1.60) headache-related disability whereas males were significantly more likely to report no headache-related disability (PR = 0.84, 95% CI = 0.82-0.86). Among those with PM, there was not a significant sex difference among those with severe headache-related disability; however, females with PM were significantly more likely to have moderate (PR = 1.52, 95% CI = 1.24-1.87) or mild (PR = 1.47, 95% CI = 1.25-1.72) levels of headache-related disability than males and were less likely to report no headache-related disability (PR = 0.93, 95% CI = 0.91-0.95). Examination of individual MIDAS items reveal that females with migraine and PM were significantly more likely than males to report inability to do household work on at least 1 day due to headache, work or school productivity reduced by at least 50% on at least 1 day due to headache, and missed family or social activities on at least 1 day due to headache.
Table 6. Sex-Specific Prevalence and Female to Male Prevalence Ratios of Headache-Related Disability and Impairment by Headache Subtype†
Other Severe Headache N = 1587
Probable Migraine N = 7485
Migraine N = 19,189
Sex Prevalence Ratio (female to male), 95% CI
Sex Prevalence Ratio (female to male), 95% CI
Sex Prevalence Ratio (female to male), 95% CI
†Denominator excludes “unknown” and missing data.
Headache-Related Disability: MIDAS Grade
Grade 1: None
Grade 2: Mild
Grade 3: Moderate
Grade 4: Severe
Headache-Related Disability: Individual MIDAS Item “Because of your headaches on how many days in the last 3 months … ?”
Missed at least 1 day of work/school
Work/school productivity reduced by at least 50% due to headache on at least 1 day
Did no household work due to headache on at least 1 day
Household productivity reduced by at least 50% due to headache on at least 1 day
Missed family or social activity due to headache on at least 1 day
“Which best describes how you are usually affected by severe headaches?”
Able to work/function normally
Working ability or activity impaired to some degree
Working ability or activity severely impaired
Bed rest required
“Each time you have a severe headache, how long are you unable to work or undertake normal activities?”
1-< 3 days
3-< 6 days
When asked how they were usually affected by their “severe” headaches, females with both migraine and PM were significantly more likely than males to report requiring bed rest during an attack, whereas males with migraine and PM were more likely to report being able to work and function normally (Table 6). When asked how long after a headache attack they were unable to work or undertake normal activities, females with migraine were more likely than males to be impaired for 3-<6 days, whereas males with migraine were significantly more likely to report being impaired for 0 or <1 day. Females with PM were significantly more likely than males to be impaired 1-<3 days whereas males with PM were significantly more likely to report no impairment following attacks.
Diagnosis, Acute and Preventive Pharmacologic Treatment and Emergency Department/Urgent Care Visits by Sex for Migraine, PM, and Other Severe Headache.—
Females who met ICHD-2 criteria for migraine at the time of the AMPP Study survey were significantly more likely than males who met these criteria to have been diagnosed with migraine by a HCP (69.8% vs 46.2%; PR = 1.29, 95% CI = 1.25-1.34), but also significantly more likely to have been diagnosed with tension headache (33.2% vs 25.5%; PR = 1.30, 95% CI = 1.23-1.38), sinus headache (40.7% vs 33.8%; PR = 1.21, 95% CI = 1.15-1.26), and “stress” headaches (30.2% vs 23.7%; PR = 1.27, 95% CI = 1.20-1.35) (Table 7). Females were significantly less likely than males with migraine to have been diagnosed with cluster headache (9.8% vs 10.9%; PR = 0.90, 95% CI = 0.82-0.99). A similar pattern was seen in PM; females who met criteria for PM were more likely than males with PM to have been diagnosed with migraine (24.0% vs 15.1%; PR = 1.59, 95% CI = 1.44-1.76), tension headache (27.1% vs 21.5%; PR = 1.26, 95% CI = 1.16-1.37), sinus headache (35.9% vs 31.3%; PR = 1.15, 95% CI = 1.07-1.23), and “stress headaches” (23.9% vs 18.2%; PR = 1.31, 95% CI = 1.20-1.44), and less likely to have been diagnosed with cluster headache (4.0% vs 5.0%; PR = 0.81, 95% CI = 0.66-1.00). Females with other severe headache were significantly more likely than males to have been diagnosed with every type of headache assessed.
Table 7. Sex-Specific Prevalence and Female to Male Prevalence Ratios of HCP Diagnosis of Headache Subtype, Emergency Department/UCC Use for Headache, and Medication Use by Headache Subtype†
Other Severe Headache N = 1587
Probable Migraine N = 7485
Migraine N = 19,189
Sex Prevalence Ratio (female to male), 95% CI
Sex Prevalence Ratio (female to male), 95% CI
Sex Prevalence Ratio (female to male), 95% CI
†Denominator excludes “unknown” and missing data.
‡Respondents were instructed to endorse all that applied.
Diagnosis received from an HCP
Current pharmacologic treatment of severe headache
Take non-prescription (OTC) medications only
Take prescription medications only
Take both prescription and non-prescription medications
Water pill or other diuretic for high blood pressure
Other medication for high blood pressure (not water pill)
Prescription medication for high cholesterol
Prescription medication for “seizures/epilepsy/fits”
Prescription medication for diabetes
Ever used emergency department/urgent care clinic for headache
Females with migraine were also significantly more likely than males to use prescription medications only for headache (PR = 1.33, 95% CI = 1.23-1.43) and to report taking both prescription and nonprescription medications for headache (PR = 1.22, 95% CI = 1.15-1.29) (Table 7). Females with migraine were significantly less likely than males to use only nonprescription medications for headache (PR = 0.83, 95% CI = 0.80-0.86) and also less likely than males to report not taking any medications for headache (PR = 0.65, 95% CI = 0.52-0.80). Similar patterns were seen for medication use by males and females with PM. There were no significant differences between the sexes for current preventive medication use among persons with migraine or PM. However, females with migraine or PM were significantly more likely to have taken a preventive medication previously, whereas males with either migraine or PM were more likely to have never used a preventive medication for headache.
Females with migraine were significantly more likely than males to be currently taking a prescription medication for depression or anxiety or to be taking a “water pill or prescription diuretic for high blood pressure” (Table 7). Females with PM followed a similar pattern. These data suggest higher rates of these conditions among females compared with males. Males with migraine were significantly more likely to be taking a prescription medication for high cholesterol or epilepsy, and males with PM were significantly more likely to be taking prescription medication for high blood pressure, high cholesterol, epilepsy, and diabetes, suggesting higher rates of comorbidity for these conditions among males with migraine or PM.
Females with migraine were significantly more likely to have visited an emergency department or urgent care clinic for “severe headache” than males (32.4% vs 24.7%; PR = 1.31, 95% CI = 1.24-1.39). Females with PM were also significantly more likely to have visited an emergency department or urgent care clinic for headache than males (13.2% vs 10.0%; PR = 1.33, 95% CI = 1.16-1.52). These results may be indicative of financial barriers or other obstacles faced by females in receiving optimal care.
This study compared the prevalence and other features of migraine, PM, and other (nonmigraine spectrum) severe headache by sex within a large population sample. These data add to the existing global body of literature on sex differences in primary headache. The prevalence of migraine reported in this study both overall and by sex is consistent with results of 2 previous population-based US prevalence studies, the AMS I and AMS II[7, 8, 20] demonstrating that the roughly three-to-one female to male sex PR has remained relatively stable in the United States over the past 30 years. Although rates vary to some degree from reports both within the United States and from other countries,1,3-30 the female preponderance in migraine is consistent. Variations in prevalence may be due to true differences in prevalence or differences in methodology and sampling strategy. The prevalence of PM reported in this study, both overall and by sex, varies more from other US and global estimates, which again may be a reflection of true prevalence or sampling and methodological issues, yet the female preponderance remains consistent.[5, 9, 26]
Our findings add to a growing body of research showing that migraine and PM are not only more prevalent in females than males, but also associated with greater symptomology, higher headache-related disability and impact, and greater healthcare resource utilization.[3, 4, 8, 19, 24, 25] Among individuals meeting criteria for migraine, females reported experiencing all migraine symptoms and visual aura at higher rates than males, which is consistent with other published reports.[34, 35] Females also reported more prescription and nonprescription medication use for headache and greater use of emergency departments and urgent care centers for headache compared with males. This is not surprising as many studies have reported that females are more likely to consult for headache than males.36-40 Although a report from the AMS found that 68% of females and 57% of males had ever consulted an HCP for headache, a recent examination of barriers to diagnosis and treatment of migraine among persons with EM with at least moderate headache-related disability from the AMPP Study database found that rates of consulting an HCP for headache within the preceding year were similar among males (46.4%) and females (45.4%). However, among consulters, diagnosis was almost 3 times more likely (odds ratio [OR] = 2.8, 95% CI = 1.34-6.00) and using guideline-specific acute treatment was almost twice as likely (OR = 1.8, 95% CI = 0.86-3.70) in females than males.
Sex differences in diagnosis and care have been extensively studied in pain, cardiovascular care, and other areas of medicine, and many sex-related disparities have been identified.41-44 Our findings suggest that this disparity might also be true for migraine. In this study, females who met ICHD-2 criteria for migraine or PM were more likely than males to have been correctly diagnosed with migraine, but were also more likely to have been diagnosed with other headache subtypes such as tension, sinus, or “stress” headache. Conversely, males in both groups were more likely to have been diagnosed with cluster headache. There are several possible explanations for this outcome. Misdiagnosis might be more likely to occur in females. Females might be more likely to seek medical consultation for headache and therefore receive more diagnoses, or females might report more symptoms than males. The beliefs of HCPs about headache epidemiology and headache causation may also play a role. HCPs might be more likely to consider a diagnosis of cluster headache in males and migraine headache in females based on existing knowledge of sex differences for those conditions. The higher prevalence of “stress” and “tension” headache diagnoses among females may also reflect cultural stereotypes of females as more likely to experience psychologically influenced headaches.
We did not find sex differences in the current use of prescription preventive headache medication, although males were slightly more likely to have never taken a preventive medication for headache. In both males and females, the proportion of respondents using preventive medication was low. Previous publications from the AMPP Study have shown that only a small fraction of subjects judged by experts to be candidates for preventive therapy received it. Females with migraine or PM were more likely than males to have gone to an emergency department or urgent care clinic for their headaches, which may indicate poorly controlled headaches, nonoptimized treatment, or greater willingness among females to seek medical care for headache.
Finally, we found that females with migraine or PM were more likely than males with these diagnoses to report using medications typically used for depression or anxiety, suggesting higher rates of these conditions. Although we did not directly assess these conditions in 2004, subsequent AMPP Study annual surveys included depression and anxiety measures. As suspected, rates of clinical depression and anxiety were significantly higher among females. Furthermore, females with migraine and PM were less likely than males to have used medications to treat either epilepsy or high cholesterol, which likely reflects the underlying population prevalence for these diseases by sex.
Although males and females reported similar subjective average headache pain severity levels and headache frequency within headache types, females reported greater headache-related disability compared with males for all headache types. Considering individual MIDAS items that focus on specific aspects of disability, females with migraine and PM were more likely than their male counterparts to report missing at least 1 day of housework and having their ability to perform housework impaired by greater than 50%, as well as reporting missing family or social activity due to headache on at least 1 day. These findings confirm much of the existing data on sex differences in headache-related disability.1,8,19,25,35,46-50 These findings may have multiple explanations. The former may be explained by greater household responsibilities on the part of females compared with males. On the other hand, females also missed more social activities, which may be engaged in equally by both sexes and possibly indicates greater impairment on the part of female migraineurs. It is also possible that females engage in more social activities and therefore reported more missed activities than males. Females may have more severe disease than males or they may be more likely than males to report symptoms and seek care. In addition, some studies have suggested that menstrual migraine is more severe and associated with greater disability, which may be reflected in these results.[51, 52] These are complementary alternative hypotheses rather than competing explanations.
Differences between sexes in migraine are likely due to a combination of biologic and psychosocial influences.53-55 Hypothesized biologic explanations have focused on fluctuations in sex hormones and receptor binding as well as the exploration of genetic factors; however, underlying mechanisms are poorly understood. Most convincing evidence for underlying gender dependent morphological and functional changes in migraine has come from recent imaging studies. High-field magnetic resonance imaging was performed in individuals with and without migraine (interictally for migraineurs). Female migraineurs were found to have thicker posterior insula and precuneus cortices compared with male migraineurs and healthy controls of both sexes. Maleki et al. also observed differential functional responses to heat and concurrent functional differences by sex among migraineurs. They conclude that these findings support a “sex phenotype” in migraine and note that sex differences involve both brain structure and function. Despite a growing awareness of sex differences in migraine, over the past 10 years nearly 80% of animal studies published in Pain included only male subjects, and only 4% were designed to test for sex differences.[55, 56] A consensus report from 2007 urged testing hypotheses on both sexes and noted the invalidity of generalizing conclusions from male-only studies to females. Researchers have also examined a variety of psychosocial factors of note in migraine expression including gender and social role expectations, differences in coping styles, and psychological differences. In a review of this topic, Smitherman and Ward (2011) warn that sex differences in the manifestation of migraine and related factors (eg, headache-related disability, comorbidities) have not been consistently replicated among treatment seeking populations and results may be moderated by a range of factors that are not consistently measured or reported.
Because the striking female preponderance in migraine is one of its most distinctive characteristics, comparing sex differences across migraine, PM, and other severe headache can assist in more clearly defining the border between migraine and other headache types. A manuscript derived from the AMS presented sex prevalence results for migraine and other severe headache but included PM in the other severe headache group. In the current study, we separated PM from other severe headache. We found a clear female to male preponderance for PM but not for other severe headache. These results suggest that PM may be part of the migraine spectrum rather than a distinct entity. Although elevated, we did not find the female to male PR for PM to be as dramatic as that of migraine. This lower female to male PR may be due to miscategorization of individuals in the PM group. Alternatively, the female preponderance in migraine may be driven by disease severity. This hypothesis is partially supported by sex differences observed in headache-related disability. That is, the sex PR increased as MIDAS grade increased for all 3 headache groups (Table 6). The distribution of headache-related disability differed by headache type. Individuals with migraine were more likely to be in the highest disability grades (MIDAS Grades III or IV) compared with those with PM or other severe headache. However, increased female to male sex ratios were also associated with MIDAS grade within each headache type as well. In general, females experienced greater headache-related disability within each headache type.
Minor differences exist in the prevalence estimates in this manuscript compared with 2 previous reports from the AMPP Study due to small differences in the number of participants included in analyses. In the earlier reports, respondents who met criteria for migraine and had headache on ≥30 days per month were not included in analyses of migraineurs. The inclusion of those respondents in the present analyses results in minor changes in prevalence estimates. For example, the estimated prevalence of migraine in previous AMPP Study manuscripts was 11.7% (17.1% in females and 5.6% in males) compared with 11.8% herein (17.3% in females and 5.7% in males). The total prevalence of PM was estimated as 4.5% (5.1% in females and 3.9% in males) compared with 4.6% (5.3% in females and 3.9% in males) herein.
This study has several limitations. First, these data are based on self-report. Healthcare records, pharmacy, or other objective data were not obtained. However, the use of self-reported sociodemographic information, symptoms, and medical information is common practice in large epidemiological studies. Second, important limitations arise from multiple statistical comparisons. Multiple statistical comparisons increase the chance that significant findings are due to chance. Readers should be aware of this limitation when interpreting the clinical significance of the findings. Although our findings are generally consistent with those of other studies, no adjustment was made to the nominal alpha level. Thus, these results are best viewed as descriptive and hypothesis-generating rather than conclusive. Interpretation of the results should emphasize the width of the CIs rather than their corresponding P values. Third, we did not correct for potential correlations stemming from the possibility that multiple members of a single household contributed data to our analyses. However, this is unlikely to have an effect on prevalence statistics or other main findings. Fourth, although some of the results are presented as both unadjusted and adjusted for sociodemographic variables, for the sake of parsimony, only unadjusted results are presented in Tables 5-7. It is possible that some of the outcomes such as healthcare resource utilization are affected by socioeconomic variables in addition to sex. These relationships have been explored in other AMPP Study based manuscripts and are also targets for future analyses. Finally, the analysis was cross-sectional in design, which limits our ability to examine causal relationships or longitudinal trajectories. Strengths of this study include its large sample size, population-based format, and symptom items that allow for assignment of ICHD-2 headache diagnoses. In addition, several validated instruments were used including the MIDAS questionnaire.
These findings extend previous research showing that migraine and PM are not only more prevalent in females, but also more disabling and associated with more symptoms and greater healthcare resource utilization. For the most part, we did not find corresponding sex differences in other (nonmigraine spectrum) severe headache. Future research into sex differences in migraine and other severe headache types should continue to explore both biologic and psychosocial hypotheses. It is imperative that females are included in both basic science and clinical studies of sex differences in headache biology and expression. Greater understanding of biologic and psychosocial factors will shed light on observed sex differences in prevalence, treatment seeking, diagnosis and treatment of migraine and nonmigraine spectrum headache and will create opportunities to improve care and outcomes for both sexes.
The authors would like to thank C. Mark Sollars, MS, and Jelena M. Pavlovic, MD, PhD, for editorial support, Christa A. Bruce, MS, for editorial support and project management and Michael T. Lynch for assistance with graphics.
Statement of Authorship
(a)Conception and DesignDawn C. Buse, Elizabeth W. Loder, Jennifer A. Gorman, Walter F. Stewart, Michael L. Reed, Kristina M. Fanning, Daniel Serrano, Richard B. Lipton
(b)Acquisition of DataMichael L. Reed, Richard B. Lipton
(c)Analysis and Interpretation of DataDawn C. Buse, Elizabeth W. Loder, Jennifer A. Gorman, Walter F. Stewart, Kristina M. Fanning, Daniel Serrano, Richard B. Lipton
(a)Drafting the ManuscriptDawn C. Buse, Elizabeth W. Loder, Jennifer A. Gorman, Kristina M. Fanning, Daniel Serrano, Richard B. Lipton
(b)Revising It for Intellectual ContentDawn C. Buse, Elizabeth W. Loder, Jennifer A. Gorman, Walter F. Stewart, Michael L. Reed, Kristina M. Fanning, Daniel Serrano, Richard B. Lipton
(a)Final Approval of the Completed ManuscriptDawn C. Buse, Elizabeth W. Loder, Jennifer A. Gorman, Walter F. Stewart, Michael L. Reed, Kristina M. Fanning, Daniel Serrano, Richard B. Lipton