Skip to content

Migraine — overview

TL;DR — Migraine is a disabling recurrent neurological disorder, not simply severe headache. ICHD-3 distinguishes migraine with and without aura and defines chronic migraine by ≥15 headache days/month for >3 months with migraine features on ≥8 days (IHS 2018, PMID 29368949). Trigeminovascular/CGRP biology underpins new acute gepants and preventive antibodies, while cortical spreading depolarization explains aura more directly than headache (Ashina 2021, PMID 33773610; Fraser 2019, PMID 31847045). Triptans and NSAIDs remain acute anchors; conventional preventives, onabotulinumtoxinA and CGRP-pathway agents reduce attacks but rarely abolish them (VanderPluym 2021, PMID 34128998; Dodick 2010, PMID 20487038). No validated biomarker replaces the clinical diagnosis (Ashina 2021, PMID 33773610; van Dongen 2017, PMID 26888294).

Definition and diagnostic anchors

  • Migraine without aura requires recurrent 4–72-hour attacks with characteristic headache features plus nausea/vomiting or photophobia and phonophobia, after excluding a better diagnosis (IHS 2018, PMID 29368949).
  • Migraine with aura adds fully reversible visual, sensory, speech/language, motor, brainstem or retinal symptoms that usually develop gradually and precede, accompany or occur without headache (IHS 2018, PMID 29368949).
  • Chronic migraine is ≥15 headache days/month for >3 months, with migraine features on ≥8 days; episodic migraine falls below this threshold (IHS 2018, PMID 29368949).
  • Medication-overuse headache can coexist with chronic migraine; exposure and rising headache frequency can reinforce one another without proving a one-way cause (Ashina 2023, PMID 36732518).
  • Red flags—thunderclap onset, new neurological deficit, fever/meningism, pregnancy/postpartum onset, cancer/immunosuppression, positional or exertional pattern, papilledema or a major pattern change—require evaluation for secondary headache.
ICHD-3 axis Operational anchor Research implication
Migraine without aura ≥5 attacks, usually 4–72 h, characteristic pain plus nausea/vomiting or photo- and phonophobia A symptom-defined cohort, not a biomarker-defined disease
Migraine with aura ≥2 attacks with fully reversible focal symptoms and characteristic timing/evolution Requires explicit separation from TIA, seizure and retinal disease
Chronic migraine ≥15 headache days/month for >3 months; ≥8 migraine-feature days A threshold on a continuous burden distribution
Medication-overuse headache High-frequency headache plus class-specific regular acute-drug overuse May be coded together with chronic migraine

Diagnosis, frequency classification and secondary-headache exclusion are different tasks. A person with a long migraine history can still develop an unrelated urgent headache, and a normal scan cannot establish migraine (classification and diagnosis).

Epidemiology and burden

GBD 2016 ranked migraine among the highest causes of years lived with disability globally, especially in ages 15–49 (GBD 2016 Collaborators 2017, PMID 28919117). Burden is not captured by attack count alone: unpredictability, interictal anxiety, cognitive symptoms, presenteeism and family effects matter.

Migraine is more prevalent in women after puberty, and hormonal fluctuation contributes to menstrual attacks. Pregnancy may change frequency, but acute and preventive choices narrow because fetal and lactation safety data are uneven (Burch 2020, PMID 31579938).

Progression from episodic to chronic migraine is associated with baseline attack frequency, medication overuse, obesity, sleep and psychiatric comorbidity, though causal contribution differs across factors (Bigal 2009, PMID 19833063; May 2016, PMID 27389092).

Mechanism sketch

Migraine emerges from altered sensory-network excitability involving hypothalamic, brainstem, thalamic and cortical systems. Activation of trigeminal afferents releases CGRP and other neuropeptides around meningeal vessels; central sensitization contributes to allodynia and prolonged attacks (Ashina 2021, PMID 33773610).

Aura is linked to cortical spreading depolarization, a slowly propagating wave of neuronal/glial depolarization followed by suppressed activity. Aura mimics include TIA and seizure, making tempo and symptom evolution clinically important (Fraser 2019, PMID 31847045).

Migraine genetics is polygenic. Earlier reviews established familial and ion-channel contributions (Anttila 2018, PMID 29478595); a GWAS of 102,084 cases identified 123 risk loci and vascular and neuronal subtype signals, but no clinical genetic test for common migraine (Hautakangas 2022, PMID 35115687).

Treatment landscape

Treatment outcomes should be read as absolute responder differences as well as mean changes. In most trials, many placebo recipients improve and most actively treated participants do not become attack-free; a statistically significant group mean therefore does not imply a large individual benefit.

Acute treatment

Systematic review supports triptans, NSAIDs, acetaminophen, gepants and lasmiditan over placebo, with effect size, recurrence and adverse-event differences (VanderPluym 2021, PMID 34128998). A 2024 network meta-analysis found triptans generally among the most effective oral monotherapies, while newer drugs offer options when triptans are ineffective or contraindicated (Karlsson 2024, PMID 39293828).

Ubrogepant improved 2-hour pain freedom and freedom from the most bothersome symptom versus placebo in ACHIEVE II (Lipton 2019, PMID 31742631). Gepants do not vasoconstrict, but cost, access and long-term comparative data shape use.

Prevention

Traditional options include topiramate, beta-blockers, candesartan, valproate, amitriptyline and others, chosen around comorbidity and safety. CGRP-targeted treatments shifted prevention toward mechanism-specific therapy:

Intervention Anchor evidence
Erenumab STRIVE episodic-migraine RCT (Goadsby 2017, PMID 29171821)
Fremanezumab Episodic-migraine RCT (Dodick 2018, PMID 29800211)
Galcanezumab EVOLVE-1 RCT (Stauffer 2018, PMID 29813147)
Eptinezumab PROMISE-2 chronic migraine and DELIVER after preventive failures (Lipton 2020, PMID 32209650; Ashina 2022, PMID 35716692)
Rimegepant Oral preventive phase 2/3 trial (Croop 2021, PMID 33338437)
OnabotulinumtoxinA PREEMPT chronic-migraine program (Dodick 2010, PMID 20487038)

In active comparisons, erenumab and atogepant each produced fewer adverse-event discontinuations and higher ≥50% response than topiramate. HER-MES tested erenumab in 777 participants; TEMPLE tested atogepant in 545, with a −1.8-day MMD contrast (95% CI −2.5 to −1.0) over months 4–6 (Reuter 2022, PMID 34743579; Reuter 2026, PMID 42492556). Sponsor context and enrolled populations still constrain transportability.

Clinical problem Evidence-supported classes Main unresolved comparison
Acute attack NSAID/acetaminophen, triptan, gepant, lasmiditan Head-to-head effectiveness under real-world timing and prior failure
Episodic prevention Conventional oral preventive, CGRP antibody, preventive gepant Which first-line sequence maximizes sustained response and access
Chronic migraine Oral preventive, CGRP-pathway therapy, onabotulinumtoxinA Monotherapy versus rational combination and de-escalation
Medication overuse Education/withdrawal plus preventive strategy Which components are necessary for which drug class
Non-drug care Behavioral therapy and device-specific neuromodulation Additive benefit under credible controls

Behavioral and device approaches

CBT meta-analysis found reductions in headache frequency and disability, though study sizes and methods varied (Bae 2021, PMID 35056352). Sleep regularity, exercise, meal regularity and trigger management should avoid turning normal life into exhaustive avoidance. Non-invasive neuromodulation offers options with modest and device-specific evidence; this field needs sham integrity and comparative trials.

Pediatric acute and preventive recommendations differ from adult evidence, with high placebo response and fewer proven preventives (Oskoui 2019, PMID 31413171; preventive guideline PMID 31413170).

Vascular and reproductive questions

Migraine—especially with aura—is associated with increased ischemic stroke risk in observational meta-analysis, but absolute risk in most young individuals is low and confounding by smoking, estrogen exposure and vascular factors matters (Zhang 2022, PMID 35451664). Association does not mean that migraine prophylaxis prevents stroke.

Menstrual migraine can be more prolonged and treatment-resistant. Pregnancy and lactation require drug-specific evidence review rather than assuming class safety (Burch 2020, PMID 31579938).

The same distinction applies to vascular evidence: an epidemiological association, a contraindication derived from pharmacology and evidence that treating migraine changes vascular outcomes are three separate propositions. Current evidence is strongest for association and weakest for treatment-mediated risk modification (Zhang 2022, PMID 35451664).

How the evidence base should be read

Evidence type What it can establish Recurrent limitation in migraine
Placebo-controlled acute RCT Short-term efficacy at prespecified time points Trial dosing may be earlier and more standardized than routine use
Preventive RCT Mean monthly-day change and responder proportion over months Long-term persistence, sequencing and comparative effectiveness remain sparse
Active-comparator trial Relative efficacy/tolerability in the enrolled population Blinding and differential dropout can alter interpretation
Observational cohort Natural history, burden and uncommon safety outcomes Treatment selection and disease severity confound associations
Meta-analysis Precision and cross-study synthesis Endpoint and phenotype heterogeneity can make pooled rankings unstable
Biomarker study Candidate association or mechanistic signal Few locked, external, clinically representative validations

Migraine trials use several non-equivalent denominators: attacks, participants, monthly migraine days and monthly headache days. “Pain relief,” “pain freedom,” “most-bothersome-symptom freedom” and ≥50% preventive response answer different questions. Cross-page tables retain the original endpoint rather than translating unlike outcomes into a common adjective.

Research frontier

  • Predicting which CGRP-pathway agent or conventional preventive will work before serial trials.
  • Defining when early effective prevention alters chronification rather than temporarily suppressing attacks (May 2016, PMID 27389092).
  • Biomarkers: CSF studies show inconsistent candidate differences, with no validated clinical assay (van Dongen 2017, PMID 26888294; Ashina 2021, PMID 33773610).
  • Long-term cardiovascular and pregnancy safety for newer therapies.
  • Separating drug efficacy from expectation in visible-device and injection trials.

Open questions

  • Can early targeted prevention stop chronification? (Bigal 2009, PMID 19833063; May 2016, PMID 27389092)
  • Which marker predicts response to CGRP blockade versus a conventional preventive? (Ashina 2021, PMID 33773610; Reuter 2022, PMID 34743579)
  • Does successful migraine prevention change vascular risk, or are migraine and stroke linked by shared causes? (Zhang 2022, PMID 35451664)
  • What are the long-term reproductive safety profiles of gepants and CGRP antibodies? (Burch 2020, PMID 31579938)
  • Which behavioral and neuromodulation benefits survive rigorous sham and implementation studies? (Bae 2021, PMID 35056352)

References

  1. Headache Classification Committee of the IHS. International Classification of Headache Disorders, 3rd edition. Cephalalgia. 2018. PMID 29368949
  2. GBD 2016 Collaborators. Global incidence, prevalence and YLDs for 328 diseases, 1990–2016. Lancet. 2017. PMID 28919117
  3. Ashina M, et al. Migraine: disease characterisation, biomarkers and precision medicine. Lancet. 2021. PMID 33773610
  4. Fraser CL, et al. Migraine aura: pathophysiology, mimics and treatment. Semin Neurol. 2019. PMID 31847045
  5. Anttila V, et al. Genetics of migraine. Handb Clin Neurol. 2018. PMID 29478595
  6. Hautakangas H, et al. GWAS of 102,084 migraine cases identifies 123 loci. Nat Genet. 2022. PMID 35115687
  7. Bigal ME, et al. Migraine chronification: concept and risk factors. Curr Pain Headache Rep. 2009. PMID 19833063
  8. May A, et al. Chronic migraine: risk factors, mechanisms and treatment. Nat Rev Neurol. 2016. PMID 27389092
  9. VanderPluym JH, et al. Acute treatments for episodic migraine in adults. JAMA. 2021. PMID 34128998
  10. Karlsson WK, et al. Comparative acute migraine drugs: network meta-analysis. BMJ. 2024. PMID 39293828
  11. Lipton RB, et al. ACHIEVE II ubrogepant trial. JAMA. 2019. PMID 31742631
  12. Goadsby PJ, et al. Controlled trial of erenumab for episodic migraine. N Engl J Med. 2017. PMID 29171821
  13. Dodick DW, et al. Fremanezumab for prevention of episodic migraine. JAMA. 2018. PMID 29800211
  14. Stauffer VL, et al. EVOLVE-1 galcanezumab trial. JAMA Neurol. 2018. PMID 29813147
  15. Lipton RB, et al. PROMISE-2 eptinezumab in chronic migraine. Neurology. 2020. PMID 32209650
  16. Ashina M, et al. DELIVER eptinezumab after preventive failures. Lancet Neurol. 2022. PMID 35716692
  17. Croop R, et al. Rimegepant for preventive treatment. Lancet. 2021. PMID 33338437
  18. Dodick DW, et al. PREEMPT onabotulinumtoxinA pooled results. Headache. 2010. PMID 20487038
  19. Reuter U, et al. Erenumab versus topiramate. Cephalalgia. 2022. PMID 34743579
  20. Bae JY, et al. CBT for migraine. Medicina. 2021. PMID 35056352
  21. Oskoui M, et al. Acute pediatric migraine guideline. Neurology. 2019. PMID 31413171
  22. Oskoui M, et al. Pediatric migraine prevention guideline. Neurology. 2019. PMID 31413170
  23. Zhang S, et al. Migraine and stroke risk. J Neurol. 2022. PMID 35451664
  24. Burch R. Menstrual migraine and migraine during pregnancy/lactation. Headache. 2020. PMID 31579938
  25. van Dongen RM, et al. CSF biomarkers in migraine. Cephalalgia. 2017. PMID 26888294
  26. Ashina S, et al. Medication overuse headache. Nat Rev Dis Primers. 2023. PMID 36732518
  27. Reuter U, et al. Atogepant versus topiramate for migraine prevention: TEMPLE. Lancet Neurol. 2026. PMID 42492556