Clinical trials landscape¶
TL;DR — Migraine development has moved from repurposed oral preventives and vasoconstrictive acute drugs to CGRP-pathway agents, non-vasoconstrictive acute drugs, devices, biomarkers and strategy trials. Registration success is real but effect sizes remain bounded: pivotal preventive trials often add ~1–2 fewer monthly migraine days versus placebo, and acute gepant trials add ~7–10 percentage points in 2-hour pain freedom (Goadsby 2017, PMID 29171821; Croop 2019, PMID 31311674). TEMPLE now supplies direct atogepant-versus-topiramate evidence, while the registered frontier includes randomized antibody-plus-onabotulinumtoxinA combination, pediatric gepant programs, pregnancy exposure, menstrual migraine, wearables and deep phenotyping (Reuter 2026, PMID 42492556; NCT07040813; NCT05125302; NCT06150781; NCT06417775; NCT05755945). Status is volatile: the table is a live ClinicalTrials.gov v2 snapshot retrieved 2026-08-30, and “Unknown” means registry recency is inadequate, not that a study necessarily stopped. The largest gap is multi-class sequencing, withdrawal and controlled combination strategy evidence rather than another placebo-controlled member of an established class.
How to read the registry¶
| Field | Meaning | Frequent error |
|---|---|---|
| Recruiting | Site recruitment reported open | Assuming every listed country/site is currently open |
| Active, not recruiting | Intervention/follow-up continues | Calling results imminent |
| Completed | Study conduct ended | Assuming results were posted/published |
| Unknown | Last-known status not recently verified | Calling the trial terminated |
| Enrollment | Planned or actual according to record | Treating planned number as analyzed sample |
| Phase | Regulatory development convention | Applying drug phases to behavioural/device studies |
ClinicalTrials.gov registration verifies a study record and status, not trial quality or a positive result. Peer-reviewed publications are cited separately.
Landmark completed programs¶
| Program | Registration | Design/enrollment | Result anchor |
|---|---|---|---|
| Erenumab STRIVE | NCT02456740 | Phase 3, completed, n=955 | MMD −3.2/−3.7 vs −1.8 placebo (Goadsby 2017, PMID 29171821) |
| Fremanezumab episodic | NCT02629861 | Phase 3, completed, n=875 | Active–placebo difference −1.3 to −1.5 MMD (Dodick 2018, PMID 29800211) |
| Galcanezumab EVOLVE-1 | NCT02614183 | Phase 3, completed, n=862 | MMD −4.7/−4.6 vs −2.8 placebo (Stauffer 2018, PMID 29813147) |
| Eptinezumab PROMISE-2 | NCT02974153 | Phase 3, completed, n=1,121 | Chronic MMD −7.7/−8.2 vs −5.6 placebo (Lipton 2020, PMID 32209650) |
| Rimegepant prevention | NCT03732638 | Phase 2/3, completed, n=1,590 | Active–placebo MMD difference −0.8 (Croop 2021, PMID 33338437) |
| Erenumab vs topiramate | NCT03828539 | Phase 4, completed, n=777 | AE discontinuation 10.6% vs 38.9% (Reuter 2022, PMID 34743579) |
| Ubrogepant ACHIEVE II | NCT02867709 | Phase 3, completed, n=1,686 | 2-h pain freedom 21.8% vs 14.3% (Lipton 2019, PMID 31742631) |
| Rimegepant acute ODT | NCT03461757 | Phase 3, completed, n=1,811 | 2-h pain freedom 21% vs 11% (Croop 2019, PMID 31311674) |
| Zavegepant nasal | NCT04571060 | Phase 3, completed, n=1,978 | 2-h pain freedom 24% vs 15% (Lipton 2023, PMID 36804093) |
| Candesartan vs propranolol | NCT00884663 | Phase 2/3, completed, n=72 | Both superior to placebo; crossover trial (Stovner 2014, PMID 24335848) |
| Atogepant vs topiramate (TEMPLE) | NCT05748483 | Phase 3, completed, n=545 | AE discontinuation 12% vs 30%; ≥50% response 64% vs 39% (Reuter 2026, PMID 42492556) |
The completed set shows why absolute differences matter. Large within-group improvement and high placebo response coexist with statistically robust target validation.
Strategy trials versus molecule trials¶
| Question | Conventional registration design | Needed strategy design |
|---|---|---|
| Does drug work? | Drug vs placebo | Established |
| Which first? | Separate placebo programs | Head-to-head sequence |
| What after nonresponse? | Exclude prior failures or subgroup | Randomized switch/continue/add |
| Is combination additive? | Single-agent placebo trial | Stable-background add-on factorial |
| When to stop? | Open-label extension | Blinded withdrawal and retreatment |
| Does early treatment modify disease? | 12–24-week suppression | Multi-year post-withdrawal incidence |
HER-MES is a rare active comparator and found a large tolerability advantage for erenumab over topiramate, but it does not compare system-level first-line strategies or cost (Reuter 2022, PMID 34743579).
Active pharmacologic trials¶
| Registration | Intervention/question | Status, phase, enrollment (2026-08-30) |
|---|---|---|
| NCT06401642 | Zavegepant acute use during CGRP-targeted prevention | Recruiting; phase 4; n=200 |
| NCT06047444 | AbobotulinumtoxinA (Dysport) for chronic-migraine prevention | Active, not recruiting; phase 3; n=759 |
| NCT06602479 | MEDI0618 versus placebo in episodic migraine | Recruiting; phase 2; n=488 |
| NCT07645924 | Elismetrep acute treatment | Recruiting; phase 3; n=1,800 |
| NCT07674654 | Elismetrep long-term acute safety | Recruiting; phase 3; n=1,000 |
| NCT07304518 | PRT-064040 nasal spray for acute migraine | Recruiting; phase 2; n=456 |
| NCT07301008 | Rimegepant preemptive use for predictable triggers | Recruiting; phase 4; n=60 |
| NCT06417775 | Ubrogepant in menstrual migraine | Active, not recruiting; phase 3; n=496 |
The preemptive-trigger and menstrual programs test timing/phenotype rather than only another average acute attack, a more clinically informative direction.
Combination and difficult-to-treat disease¶
The Nordic phase-3 trial directly compares CGRP antibody plus onabotulinumtoxinA against antibody monotherapy in chronic migraine (recruiting, n=450; NCT07040813). This closes a major evidence gap created by observational combination use.
Other registered questions include nerve block versus onabotulinumtoxinA (recruiting, n=64; NCT06684249), greater-occipital nerve block in MOH (not yet recruiting, n=130; NCT07568496), and abobotulinumtoxinA as a potential alternative toxin formulation (NCT06047444).
| Design risk | Control needed |
|---|---|
| Visible forehead weakness reveals toxin | Allocation guess and active comparator |
| Block duration shorter than preventive | Time-aligned repeat protocol |
| Regression from frequency peak | Baseline diary and concurrent control |
| Add-on selected after partial response | Stable background and randomized add-on |
Pediatrics¶
| Registration | Question | Status/enrollment |
|---|---|---|
| NCT05125302 | Ubrogepant acute efficacy/safety ages 6–17 | Recruiting phase 3; n=1,059 |
| NCT05127954 | Ubrogepant pediatric long-term extension | Enrolling by invitation phase 3; n=1,200 |
| NCT04743141 | Rimegepant pediatric acute long-term safety | Recruiting phase 3; n=600 |
| NCT05337033 | Cannabis for chronic headaches in adolescents | Recruiting phase 2; n=20 |
Pediatric trials must overcome high placebo response, developmental pharmacology and school-centered outcomes. Long-term extensions expand exposure denominators but lack placebo comparison and preferentially enroll prior-study participants.
The earlier pediatric chronic-headache CBT trial record remains “Unknown” in the registry despite published results (NCT00389038; Powers 2013, PMID 24368463), demonstrating why registry status and publication reality must be reconciled rather than read in isolation.
Pregnancy, reproductive and menstrual evidence¶
The Aimovig pregnancy exposure registry is recruiting with planned enrollment 2,842 (NCT06150781). Exposure registries are essential for uncommon outcomes but face voluntary enrollment, exposure timing and comparator selection.
External trigeminal nerve stimulation in pregnancy is recruiting 550 participants (NCT06788977), addressing demand for non-systemic options. Ubrogepant menstrual-migraine phase 3 (NCT06417775) separates a predictable high-burden phenotype from general acute trials.
No pregnancy trial should be interpreted only through malformations: pregnancy loss, growth, prematurity, hypertensive disease, neonatal adaptation and maternal disability all matter.
Biomarkers and prediction¶
| Registration | Biomarker domain | Design/status |
|---|---|---|
| NCT03231241 | Chronification: imaging, deep phenotyping, proteomics | Recruiting observational; n=700 |
| NCT06692088 | Longitudinal amylin under CGRP therapy | Not yet recruiting; n=216 |
| NCT05755945 | Wearables to predict attacks | Active, not recruiting; n=20 |
| NCT06633484 | MRI biomarker program (BIOMIGA) | Active, not recruiting; n=219 |
| NCT07395908 | CGRP/PACAP-38 around interventions | Active, not recruiting; n=100 |
Prediction studies need locked external validation and a clinical decision comparison. A wearable that detects an attack after symptoms begin is not a pre-attack predictor; a marker associated with response in one treatment arm is not necessarily treatment-predictive.
REFORM uses clinical, blood, MRI and provocation measures around erenumab response and provides a prospective framework for multimodal prediction (Karlsson 2023, PMID 37303034).
Devices and behavioural trials¶
| Registration | Intervention | Status/enrollment |
|---|---|---|
| NCT06623188 | Auricular neuromodulation for episodic migraine | Recruiting; n=106 |
| NCT07508449 | Exercise in migraine | Active, not recruiting; n=140 |
| NCT06450444 | RECLAIM implant/device study | Recruiting; n=110 |
| NCT07599228 | VR mindfulness | Not yet recruiting; n=160 |
| NCT07694063 | Online Hatha yoga add-on | Recruiting; n=60 |
These studies broaden outcomes beyond drug response, but sham credibility, adherence and intervention dose require the same rigor as pharmacology.
Vascular mechanism and PFO¶
A phase-3 trial of PFO closure versus medicine for migraine is active but not recruiting (n=440; NCT04946734). Because prior randomized closure trials were negative on primary outcomes, new work should prespecify aura phenotype, antiplatelet effects and multiplicity rather than rely on post-hoc responders (Elbadawi 2019, PMID 29914296).
Common design pathologies¶
| Pathology | Consequence | Repair |
|---|---|---|
| Single treated attack | No consistency estimate | Multi-attack repeated randomization |
| Mean MMD only | Responder heterogeneity hidden | Responder and functional distributions |
| Healthy/low-comorbidity sample | Safety/effectiveness overestimated | Pragmatic active-comparator cohort |
| Inadequate sham sensation | Expectancy bias | Allocation-guess reporting |
| Short double-blind phase | Rare/long-term harms missed | Registry and active surveillance |
| Last-observation imputation | Attrition can inflate benefit | Estimand/missing-data sensitivity |
| Multiple secondary endpoints | Selective positive narrative | Prespecified hierarchy and full results |
| Sponsor-only pipeline | Publication/analysis dependence | Data sharing and independent replication |
Acute-endpoint review found wide variation across pain relief, pain freedom and associated-symptom outcomes, limiting comparison (García-Azorin 2018, PMID 30242571).
Trial priorities¶
- first-line strategy trials with cost and sustained function;
- randomized combination/add-on and component withdrawal;
- pregnancy and pediatric long-term safety;
- prevention of chronification after treatment withdrawal;
- locked biomarker/wearable validation against simple clinical models;
- low-resource implementation trials using essential formularies;
- patient-weighted outcomes including interictal burden and consistency.
Open questions¶
- Will dual CGRP-antibody/onabotulinumtoxinA therapy show additive phase-3 benefit and acceptable cost? (NCT07040813)
- Can pediatric gepant trials demonstrate consistent multi-attack benefit above high placebo response? (NCT05125302; NCT04743141)
- Do wearables predict enough lead time to improve preemptive treatment outcomes? (NCT05755945; NCT07301008)
- Can pregnancy registries recruit representative exposed and disease-matched comparator cohorts? (NCT06150781)
- Which completed studies remain unpublished or have outdated “Unknown” status, and what results would alter synthesis?
Related pages¶
- Acute treatment — pivotal endpoint interpretation.
- Preventive treatment — completed registration programs.
- Biomarkers — validation requirements for prediction studies.
- Guidelines — how new evidence changes sequencing.
- Open questions — prioritized trial-ready gaps.
References¶
- Goadsby PJ, et al. Controlled trial of erenumab for episodic migraine. N Engl J Med. 2017. PMID 29171821
- Dodick DW, et al. Fremanezumab for episodic migraine prevention. JAMA. 2018. PMID 29800211
- Stauffer VL, et al. Galcanezumab EVOLVE-1. JAMA Neurol. 2018. PMID 29813147
- Lipton RB, et al. Eptinezumab PROMISE-2. Neurology. 2020. PMID 32209650
- Croop R, et al. Rimegepant preventive trial. Lancet. 2021. PMID 33338437
- Reuter U, et al. Erenumab versus topiramate. Cephalalgia. 2022. PMID 34743579
- Lipton RB, et al. Ubrogepant ACHIEVE II. JAMA. 2019. PMID 31742631
- Croop R, et al. Rimegepant ODT acute trial. Lancet. 2019. PMID 31311674
- Lipton RB, et al. Zavegepant nasal phase 3. Lancet Neurol. 2023. PMID 36804093
- Stovner LJ, et al. Candesartan versus propranolol. Cephalalgia. 2014. PMID 24335848
- Powers SW, et al. CBT plus amitriptyline for pediatric chronic migraine. JAMA. 2013. PMID 24368463
- Karlsson WK, et al. REFORM methodology and baseline characteristics. J Headache Pain. 2023. PMID 37303034
- Elbadawi A, et al. PFO closure for migraine: meta-analysis of randomized trials. Acta Cardiol. 2019. PMID 29914296
- García-Azorin D, et al. Endpoints in symptomatic primary-headache trials. J Headache Pain. 2018. PMID 30242571
- Reuter U, et al. Atogepant versus topiramate for migraine prevention: TEMPLE. Lancet Neurol. 2026. PMID 42492556
All 40 NCT records in this page were retrieved from the ClinicalTrials.gov v2 API on 2026-08-30; their live statuses may change.