Migraine: treatment and what CGRP blockers show

What is migraine and how many people does it affect?

Migraine is a neurological disease marked by recurrent attacks of headache, often pulsating and on one side, accompanied by nausea, sensitivity to light and sound, and in some patients aura (transient neurological symptoms, mostly visual)¹³. It is not simply "a bad headache": it is a disorder of sensory processing involving the trigeminovascular system, the brainstem and the thalamus³⁷.

The burden is enormous. The Global Burden of Disease study estimated that almost 3 billion people had migraine or tension-type headache in 2016: 1.04 billion with migraine and 1.89 billion with tension-type headache². Because migraine carries a much higher disability weight, it caused 45.1 million years lived with disability (YLDs), against 7.2 million for tension-type headache². It was most burdensome in women aged 15 to 49 years, in whom it caused 20.3 million YLDs, 11.2% of all YLDs in that age and sex group².

Symptoms and how migraine is diagnosed

Diagnosis is clinical and rests on the International Classification of Headache Disorders, third edition (ICHD-3)¹. Migraine without aura requires at least five attacks lasting 4 to 72 hours, with at least two of four pain features (unilateral, pulsating, moderate or severe intensity, aggravated by physical activity) and either nausea or both photophobia and phonophobia¹. Chronic migraine is defined as headache on 15 or more days a month for more than 3 months, with at least 8 days having migraine features¹.

There is no blood test or scan that confirms migraine. Neuroimaging is reserved for red flags (sudden-onset headache, persistent neurological deficit, abrupt change in pattern), because the aim is to exclude secondary causes³. A headache diary helps count pain days and days of acute-medication use, two data points that change management⁴.

Standard treatment of migraine today

Treatment has three parts: general measures (identifying and reducing triggers, regular sleep), treatment of the acute attack and, when attacks are frequent or disabling, preventive treatment³⁴. For the acute attack, the mainstays are nonsteroidal anti-inflammatory drugs and triptans, which activate serotonin receptors. A Cochrane overview of sumatriptan included data from 52,236 participants: the subcutaneous route took pain from moderate or severe to none at 2 hours in 59% versus 15% with placebo (number needed to treat 2.3), and the oral route did so in 28% versus 11% (number needed to treat 6.1)⁵.

The 2021 American Headache Society consensus statement notes that since the previous version no evidence has emerged to alter the established principles of acute or preventive treatment, and it adds the gepants ubrogepant and rimegepant, the 5-HT1F agonist lasmiditan, the intravenous antibody eptinezumab and neuromodulation as new options⁴. In 2024 the same society updated its position on prevention: it concluded that CGRP-targeting therapies (erenumab, fremanezumab, galcanezumab, eptinezumab, rimegepant and atogepant) should be considered a first-line option alongside earlier preventive treatments, without requiring prior failure of other classes²⁴. A systematic review and network meta-analysis for the American College of Physicians (61 studies, 20,680 patients, 16 treatments) concluded that no preventive treatment for episodic migraine was shown superior to another with high-certainty evidence, although anti-CGRP antibodies probably caused fewer discontinuations for adverse events than topiramate (risk difference −16.2%)²².

CGRP: the peptide that became a target

Calcitonin gene-related peptide (CGRP) is a 37-amino-acid vasodilator neuropeptide found in trigeminal fibers. In 1990, Goadsby, Edvinsson and Ekman measured neuropeptides in venous blood from patients during a migraine: they found a substantial elevation of CGRP in the external jugular vein but not in cubital blood, and the rise was larger in migraine with aura than in common migraine; the other peptides measured (neuropeptide Y, vasoactive intestinal peptide and substance P) did not show the same pattern⁶. That observation opened the therapeutic path described in later pathophysiology reviews⁷.

Proof of concept came in 2004 with olcegepant (BIBN 4096 BS), an intravenous CGRP receptor antagonist: in the acute attack, the selected dose achieved a 66% response versus 27% with placebo (P=0.001), with an adverse-event rate of 25% versus 12% with placebo⁸. Its route of administration prevented clinical development, but it showed that blocking CGRP could stop an attack. Two strategies followed: monoclonal antibodies (against the peptide or its receptor) and oral small-molecule antagonists, the gepants⁴⁷.

What anti-CGRP antibodies have shown in prevention

Erenumab is a human antibody that blocks the CGRP receptor. In STRIVE (955 patients with episodic migraine, 8.3 migraine days a month at baseline), the reduction in months 4 to 6 was 3.2 days with the lower dose and 3.7 with the higher, versus 1.8 with placebo; a reduction of at least 50% was achieved in 43.3% and 50.0%, versus 26.6% with placebo⁹. It was also tested in chronic migraine in a randomized, placebo-controlled phase 2 trial¹⁰.

Fremanezumab, an antibody against the ligand, was studied in chronic migraine (1,130 patients, about 13 headache days a month at baseline): the mean reduction was 4.3 days with the quarterly regimen, 4.6 with the monthly one and 2.5 with placebo (P<0.001); a reduction of at least 50% occurred in 38%, 41% and 18%, respectively¹¹. Galcanezumab, also anti-ligand, reduced migraine days in EVOLVE-1 by 4.7 and 4.6 days depending on dose, versus 2.8 with placebo (858 patients), with discontinuation for adverse events in under 5%¹². Eptinezumab, the only intravenous antibody, reduced monthly migraine days by 3.9 to 4.3 days (higher doses) versus 3.2 with placebo in PROMISE-1 (888 patients)¹³.

In absolute terms the effects are moderate: on average, roughly 1 to 2 fewer migraine days a month than placebo. What matters most is the proportion of patients who respond (over 40% halved their migraine days in STRIVE, versus 26.6% with placebo) and tolerability⁹²².

Gepants: oral CGRP receptor antagonists

Gepants are small molecules, not peptides, that block the CGRP receptor. The FDA identifies eight CGRP or CGRP-receptor blockers approved and marketed in the United States, counting antibodies and gepants together²¹. Ubrogepant was evaluated in 1,672 participants during an attack: pain freedom at 2 hours was 11.8% with placebo, 19.2% with the lower dose and 21.2% with the higher, and freedom from the most bothersome symptom was 27.8%, 38.6% and 37.7%; the most common adverse events were nausea, somnolence and dry mouth¹⁴. Rimegepant (1,186 patients) achieved pain freedom at 2 hours in 19.6% versus 12.0% with placebo (absolute difference 7.6 percentage points; 95% CI, 3.3 to 11.9) and freedom from the most bothersome symptom in 37.6% versus 25.2%¹⁵. It was also tested for prevention in a phase 2/3 trial¹⁶.

Atogepant was studied only for prevention (ADVANCE; 910 patients enrolled and 873 in the analysis): monthly migraine days fell by 3.7, 3.9 and 4.2 days with the three doses, versus 2.5 days with placebo; the differences from placebo were −1.2, −1.4 and −1.7 days, with constipation a frequent adverse event¹⁷. In absolute terms, acute gepants add about 7 to 9 percentage points of pain freedom at 2 hours over placebo; the improvement is real and not magic¹⁴¹⁵.

Beyond CGRP: the peptide PACAP

The next peptide under study is PACAP (pituitary adenylate cyclase-activating polypeptide), another neuropeptide involved in migraine pathophysiology: infusing it into people with migraine can provoke attacks, and an antibody directed against its receptor showed no efficacy in a placebo-controlled trial¹⁹. In the phase 2 HOPE trial, a single intravenous infusion of a monoclonal antibody against PACAP (Lu AG09222) reduced migraine days over weeks 1 to 4 by 6.2 days, versus 4.2 with placebo (difference −2.0 days; 95% CI, −3.8 to −0.3; P=0.02), in 237 participants with a baseline mean of 16.7 migraine days a month¹⁸. This is a phase 2 result, not an approval: the primary endpoint was assessed with the 750 mg dose, and in this small, short study the endpoint of a 50% or greater reduction in migraine days was not met¹⁸¹⁹.

To place these molecules in perspective, the history of therapeutic peptides is in A century of therapeutic peptides, another disease where a peptide was a well-characterized target is in Pulmonary arterial hypertension: treatment and sotatercept, and neuropeptides for sleep and cognition, with much weaker evidence, are in Sleep and cognition peptides.

Safety and contraindications of CGRP blockers

The safety of CGRP blockade is the point that needs most care, because this peptide also regulates vascular tone outside the brain. A prevention trial of telcagepant, a first-generation gepant, was terminated on the recommendation of the safety board: 13 patients, all in the telcagepant groups, had an alanine aminotransferase elevation of at least 3 times the upper limit of normal, and 7 also had it for aspartate aminotransferase, with two symptomatic cases that resolved after the drug was stopped²⁰. Current gepants have not shown that signal in the trials described, but the precedent explains the monitoring.

Hypertension is the most recent signal. An FDA pharmacovigilance analysis published in 2026 identified 106 cases of hypertension with a probable or possible causal association with approved CGRP blockers other than erenumab (which had already led to a label amendment), 74% with antibodies and 26% with gepants²¹. A systematic review of spontaneous reporting data also covers post-marketing safety for the class²³. In trials, the most frequent adverse events were constipation (atogepant), nausea and somnolence (ubrogepant), injection-site reactions and upper respiratory tract infections^11131417.

Summary by agent

Frequently asked questions

Are anti-CGRP drugs peptides? No. CGRP is a peptide, but the antibodies (erenumab, fremanezumab, galcanezumab, eptinezumab) are large proteins and the gepants are small molecules that block the peptide or its receptor⁴⁷.

Are CGRP blockers approved for migraine? Yes. The FDA identifies eight CGRP or CGRP-receptor blockers approved and marketed in the United States, and the 2021 American Headache Society statement lists the gepants ubrogepant and rimegepant and the antibody eptinezumab among the new treatments⁴²¹.

How much do they reduce migraine days? In prevention trials, 1.4 to 2.1 days a month more than placebo; for example, 4.7 days with galcanezumab versus 2.8 with placebo, and 4.3 with quarterly fremanezumab versus 2.5¹¹¹².

What is a triptan and when is it used? It is a serotonin receptor agonist for the acute attack; in the Cochrane review of sumatriptan, the oral route cleared pain at 2 hours in 28% versus 11% with placebo⁵.

Are other peptides being studied for migraine besides CGRP? Yes. An antibody against PACAP reduced migraine days in a phase 2 trial, but it is not yet approved¹⁸¹⁹.