Skip to content

Epidemiology and burden

TL;DR — Migraine affects more than one billion people and concentrates disability in adolescence through midlife rather than shortening life directly (Simmonds 2023, PMID 38043969; GBD 2023 Collaborators 2025, PMID 41240916). Older population summaries estimated annual prevalence near 6% in men and 15–18% in women; contemporary estimates vary with case definition, survey method, geography and whether probable migraine is included (Lipton 1997, PMID 9058393). GBD 2023 modelled 1.2 billion people with migraine in 2023 and made time-symptomatic adjustments rather than treating all prevalent days as disabled days (GBD 2023 Collaborators 2025, PMID 41240916). The burden is largely invisible in mortality statistics: attacks remove productive, social and family time, and chronic migraine is associated with worse health, lower income and unemployment (Leonardi 2023, PMID 38043966). Diagnosis and migraine-specific treatment remain incomplete even in well-resourced settings (Katsarava 2018, PMID 29392600).

The denominator problem

Prevalence is not a single transportable constant. Estimates change when studies use clinician interview, validated questionnaire, administrative diagnosis or self-report; when the recall period is lifetime versus one year; and when probable migraine is counted. Clinic samples over-represent severity: fewer than 5% of affected people reached specialists in older epidemiological work (Silberstein 1993, PMID 8272178).

Design choice Direction of likely effect Reason
Include probable migraine Raises prevalence Captures attacks missing one ICHD feature
Claims/diagnosis codes Usually lowers detected prevalence Requires access, recognition and coding
Headache-clinic sampling Inflates severity and chronicity Referral selects difficult disease
Web-panel screening Broad reach but selection-sensitive Participation and internet access shape sample
One-year recall More comparable than lifetime recall Reduces distant-event memory error
Diary ascertainment Better frequency resolution Higher respondent burden and shorter observation

GBD does not conduct one global survey. It synthesizes population-based sources through Bayesian meta-regression, estimates missing age-sex-country cells and applies disability weights; uncertainty therefore includes sampling and modelling assumptions. The 2010 nonfatal-burden analysis established the sequela/YLD framework later releases extended, while a 1990–2017 US state analysis illustrates how the same modelling architecture can be resolved below national level (GBD 2010 Collaborators 2012, PMID 23245607; GBD 2016 Collaborators 2017, PMID 28919117; GBD Headache Collaborators 2018, PMID 30353868; Feigin 2021, PMID 33136137).

Global scale

Estimate Year/data Population and method Result
Global migraine prevalence 2023 GBD 2023; population sources + hierarchical Bayesian meta-regression 1.2 billion people; age-standardised prevalence higher in females (GBD 2023 Collaborators 2025, PMID 41240916)
Global headache burden 2016 GBD 2016, 195 countries Migraine among the leading YLD causes; burden concentrated in working ages (GBD Headache Collaborators 2018, PMID 30353868)
All-cause nonfatal burden 2016 328 causes, DisMod-MR 2.1 Established the comparative YLD ranking used for migraine (GBD 2016 Collaborators 2017, PMID 28919117)
Neurological burden 2016 15 neurological categories Migraine was a major contributor to neurological DALYs through disability rather than mortality (GBD Neurology Collaborators 2019, PMID 30879893)
GBD 2021 trend re-analysis 1990–2021 Extracted GBD estimates and ARIMA projection Cases rose 58.15%, from 732.56 million to 1.16 billion, largely reflecting population growth and changing demography (Dong 2025, PMID 39661241)
African community prevalence Studies 1970–2014 21 studies; 137,277 participants; random-effects meta-analysis Demonstrated substantial heterogeneity and evidence gaps rather than an intrinsically low African burden (Woldeamanuel 2014, PMID 24814950)

The 2023 GBD revision is methodologically important: prevalence is not identical to time disabled. The analysis estimated the proportion of time symptomatic by frequency state, making disability more responsive to monthly attack burden (GBD 2023 Collaborators 2025, PMID 41240916). Comparisons with older GBD releases must therefore distinguish true trend from model change.

Sex, age and life course

Migraine prevalence is similar by sex before puberty, then becomes substantially higher in females. The older recurring population estimate—6% of men and 15–18% of women annually—remains a useful orientation but not a universal constant (Lipton 1997, PMID 9058393). Across social and geographical groups, women carry higher prevalence at nearly every adult age (Simmonds 2023, PMID 38043969).

Prevalence peaks during the years in which education, employment, caregiving and family formation overlap. GBD age curves and population studies place the greatest YLD burden below age 50; this is why migraine ranks higher by YLD than by DALY and essentially disappears from death-based rankings (GBD Headache Collaborators 2018, PMID 30353868; Leonardi 2023, PMID 38043966).

Life stage Epidemiological feature Measurement caveat
Childhood Prevalence rises with age; phenotype can be shorter and bilateral Adult-shaped questionnaires may miss cases
Puberty/adolescence Female excess emerges Menarche and hormonal exposure are correlated with other developmental changes
Ages 25–55 Highest population prevalence and work impact Employment-based samples miss non-employed people
Pregnancy Attack frequency often changes Medication changes and recall complicate causal inference
Later life Prevalence generally falls New aura-like or headache symptoms require stronger secondary-cause exclusion

Sex is not a sufficient causal explanation. Hormonal fluctuation, genetic liability, pain processing, exposure, social roles and recognition may all contribute; epidemiological ratios cannot apportion them.

Episodic, chronic and medication-overuse burden

Chronic migraine is less prevalent but produces disproportionate disability, health-care use and economic loss. CaMEO used modified ICHD criteria in a large US web panel and repeatedly assessed headache frequency, disability and treatments over a year, showing movement between episodic and chronic categories rather than fixed membership (Adams 2015, PMID 25304766).

CaMEO-I screened 90,613 respondents across Canada, France, Germany, Japan, the UK and US and documented substantial cross-country gaps in diagnosis and care (Adams 2023, PMID 37314231). Because web participation, modified criteria and self-report all enter the denominator, the survey is strongest for within-study comparisons and care gaps, not as a replacement for door-to-door prevalence studies.

Medication overuse is concentrated in high-frequency disease. In CaMEO, acute-medication overuse varied by drug class and was associated with greater headache burden; cross-sectional data cannot by themselves decide how much is cause versus marker of severe disease (Schwedt 2021, PMID 34476122). Opioid use was similarly enriched among respondents with higher disability and comorbidity (Lipton 2020, PMID 32527971).

Disability beyond headache days

Domain Observed burden Source and boundary
Work Absence plus reduced performance while present Societal-impact synthesis; methods vary by wage and recall assumptions (Leonardi 2023, PMID 38043966)
Career/finances More adverse effects in chronic than episodic migraine 19,891-person CaMEO sample; self-reported associations (Buse 2019, PMID 31407321)
Partner/family Missed activities, altered roles and relationship strain Paired CaMEO family module (Buse 2016, PMID 27132088)
Adolescent children Social, academic and emotional effects reported when a parent has migraine Adolescent CaMEO module (Buse 2018, PMID 29355924)
Sleep Poor sleep and high sleep-apnoea risk more frequent with chronic migraine Cross-sectional CaMEO association; not proof of direction (Buse 2019, PMID 30381821)
Interictal life Planning, fear of the next attack and activity restriction Impact synthesis; often absent from attack-frequency endpoints (Leonardi 2023, PMID 38043966)

The IMPAC scale was developed from literature, clinician input and patient focus groups to quantify impact on partners and adolescent children, formalizing burden outside the affected individual (Lipton 2017, PMID 28185239). This matters because cost-effectiveness models centered only on pain hours and health-care use miss redistributed household labour and lost family participation.

Economic burden

Direct costs include consultations, emergency care, investigations and medicines. Indirect costs include absence, presenteeism, reduced hours, career limitation and unpaid-care displacement. Older work already found indirect costs exceeded direct costs because prevalence peaks in working ages (Lipton 1997, PMID 9058393); newer synthesis retains that pattern across countries (Leonardi 2023, PMID 38043966).

Cost estimates are not portable across years or health systems. Currency year, purchasing-power conversion, wage method, employed-only denominators, inclusion of informal work and whether comorbidity is attributed to migraine can change totals more than sampling error.

Cost-study field that must be reported Why it matters
Price/currency year Inflation can dominate longitudinal comparison
Perspective Payer, employer and societal totals include different resources
Frequency category Chronic migraine has much higher per-person cost
Productivity instrument and recall Presenteeism is sensitive to self-attribution
Comorbidity adjustment Depression, anxiety and sleep disorders share costs
Treatment status Successful prevention can change both pharmacy and productivity costs

Diagnosis and treatment gaps

In Eurolight, many people meeting migraine criteria had not consulted a physician or used a triptan, demonstrating underdiagnosis and undertreatment across ten European countries (Katsarava 2018, PMID 29392600). The HURT questionnaire was developed to help primary care detect under-response across diagnosis, disability, medication use and treatment satisfaction, acknowledging that specialist-only service models cannot cover population need (Steiner 2018, PMID 29445880).

In CaMEO, only 22.9% of 13,624 respondents reported current acute prescription use; 12.6% had discontinued and 64.5% reported none (Hutchinson 2020, PMID 32247344). Among discontinuers, reasons included inadequate efficacy and tolerability, but survey data cannot tell whether access, diagnosis, prescriber decisions or patient preference was primary (Lipton 2019, PMID 31544244).

Disparity can arise at each step:

  1. symptoms are normalized or stigmatized;
  2. care is inaccessible;
  3. headache is not recognized as migraine;
  4. an acute medicine is never offered or is unaffordable;
  5. prevention is not discussed despite disability;
  6. follow-up does not measure response;
  7. specialist referral is delayed after failure.

Interpretation rules

  • Report both percentage and denominator; “women with migraine” and “all women” answer different questions.
  • Do not average conflicting prevalence estimates across criteria or methods without a model designed for that heterogeneity.
  • Separate prevalence growth from population growth and changes in GBD methodology (Dong 2025, PMID 39661241).
  • Avoid interpreting an administrative diagnosis rate as disease prevalence.
  • Treat chronic migraine as a high-burden state with transitions, not an immutable person-level trait (Adams 2015, PMID 25304766).
  • Do not infer that a comorbidity causes migraine from a cross-sectional odds ratio.

Open questions

  • How much of the apparent rise in global migraine burden reflects population growth, ascertainment and modelling revisions versus changing age-specific risk? (Dong 2025, PMID 39661241; GBD 2023 Collaborators 2025, PMID 41240916)
  • Which low- and middle-income regions still lack representative ICHD-based surveys, and how does evidence sparsity widen GBD uncertainty? (Woldeamanuel 2014, PMID 24814950)
  • Can population cohorts measure unpaid work, interictal restriction and household spillover without double-counting disability? (Buse 2016, PMID 27132088; Leonardi 2023, PMID 38043966)
  • Which care-cascade intervention—recognition, first prescription, preventive access or structured follow-up—recovers the most disability per resource? (Katsarava 2018, PMID 29392600; Steiner 2018, PMID 29445880)
  • What proportion of chronic-migraine economic burden is reversible within a year of effective prevention? (Adams 2015, PMID 25304766)

References

  1. GBD 2023 Headache Collaborators. Global, regional, and national burden of headache disorders, 1990-2023. Lancet Neurol. 2025;24:1005-1015. PMID 41240916
  2. Simmonds L, et al. Epidemiology of migraine. Handb Clin Neurol. 2023;198:31-38. PMID 38043969
  3. Lipton RB, et al. Prevalence and impact of migraine. Neurol Clin. 1997;15:1-13. PMID 9058393
  4. GBD 2016 Collaborators. Global, regional, and national incidence, prevalence, and YLDs for 328 diseases and injuries, 1990-2016. Lancet. 2017;390:1211-1259. PMID 28919117
  5. GBD 2016 Headache Collaborators. Global, regional, and national burden of migraine and tension-type headache, 1990-2016. Lancet Neurol. 2018;17:954-976. PMID 30353868
  6. GBD 2016 Neurology Collaborators. Global, regional, and national burden of neurological disorders, 1990-2016. Lancet Neurol. 2019;18:459-480. PMID 30879893
  7. Dong L, et al. The global burden of migraine: a 30-year trend review and future projections. Pain Ther. 2025;14:297-315. PMID 39661241
  8. Vos T, et al. Years lived with disability for 1160 sequelae of 289 diseases and injuries, 1990-2010. Lancet. 2012;380:2163-2196. PMID 23245607
  9. Silberstein SD. Epidemiology of migraine. Neuroepidemiology. 1993. PMID 8272178
  10. Woldeamanuel YW, et al. Prevalence of migraine headache and its weight on neurological burden in Africa. J Neurol Sci. 2014;342:1-15. PMID 24814950
  11. Leonardi M, et al. Societal and personal impact of migraine. Handb Clin Neurol. 2023;198:23-29. PMID 38043966
  12. Katsarava Z, et al. Poor medical care for people with migraine in Europe: evidence from the Eurolight study. J Headache Pain. 2018;19:10. PMID 29392600
  13. Steiner TJ, et al. The HURT questionnaire, an outcome measure to guide follow-up in primary care. J Headache Pain. 2018;19:15. PMID 29445880
  14. Adams AM, et al. The impact of chronic migraine: CaMEO methods and baseline results. Cephalalgia. 2015;35:563-578. PMID 25304766
  15. Adams AM, et al. CaMEO-International: methods and multi-country baseline findings for diagnosis rates and care. Cephalalgia. 2023;43:3331024231180611. PMID 37314231
  16. Buse DC, et al. Sleep disorders among people with migraine: CaMEO. Headache. 2019;59:32-45. PMID 30381821
  17. Buse DC, et al. Life with migraine: effects on relationships, career, and finances from CaMEO. Headache. 2019;59:1286-1299. PMID 31407321
  18. Hutchinson S, et al. Characterization of acute prescription migraine medication use: CaMEO. Mayo Clin Proc. 2020;95:709-718. PMID 32247344
  19. Buse DC, et al. Adolescent perspectives on the burden of a parent's migraine: CaMEO. Headache. 2018;58:512-524. PMID 29355924
  20. Buse DC, et al. Impact of migraine on the family: perspectives of people with migraine and partners in CaMEO. Mayo Clin Proc. 2016. PMID 27132088
  21. Lipton RB, et al. Family impact of migraine: development of the IMPAC scale. Headache. 2017;57:570-585. PMID 28185239
  22. Lipton RB, et al. Characterizing opioid use in a US population with migraine: CaMEO. Neurology. 2020;95:e457-e468. PMID 32527971
  23. Schwedt TJ, et al. Medication overuse and headache burden: CaMEO. Neurol Clin Pract. 2021;11:216-226. PMID 34476122
  24. Lipton RB, et al. Discontinuation of acute prescription medication for migraine: CaMEO. Headache. 2019;59:1762-1772. PMID 31544244
  25. Feigin VL, et al. Burden of neurological disorders across the US from 1990-2017. JAMA Neurol. 2021;78:165-176. PMID 33136137