Introduction

When a migraine treatment works in a trial, part of the benefit comes from the drug and part comes from everything around it: the act of taking something, the expectation of relief, the natural tendency of an attack to fade, and the way the study is run. Untangling those pieces matters, because it changes how we read the evidence behind every acute migraine medication.

A new systematic review and meta-analysis published in Neurology pooled 126 randomized controlled trials of acute migraine treatments to measure how often people improved after taking a placebo, and what made that response larger or smaller. The picture that emerges is useful for patients trying to make sense of their own experience and for clinicians trying to interpret trial results.

What "Placebo Response" Actually Means

It helps to be precise here, because the word placebo gets used loosely. In a trial, the placebo response is simply the improvement seen in the group that did not receive the active drug. That improvement is not one thing. It is a mix of several forces working at once.

This distinction is the reason a large placebo response is not evidence that migraine is "in your head." Migraine is a real neurologic disorder. A high placebo response reflects how sensitive migraine outcomes are to context, timing, and study design, not whether the condition is genuine.

Inside the Study

The researchers searched multiple databases and trial registers for double-blind, parallel-group phase 2 and phase 3 trials of acute migraine drugs, from the earliest records through November 2025. They pooled placebo response rates using random-effects models and then ran subgroup analyses and meta-regressions to see what pushed the response up or down.

The primary outcome was 2-hour pain freedom, meaning being completely free of head pain two hours after taking the placebo. This is the same benchmark regulators and headache specialists use to judge acute migraine drugs, so it is a fair yardstick.

The Key Numbers

Across all trials, the pooled 2-hour pain freedom rate on placebo was 11 percent, with a 95 percent confidence interval of 10 to 12 percent. The wider prediction interval ran from 4 to 27 percent, which tells you how much individual trials can vary.

A few of these deserve a second look. More than half of people on placebo still reached for rescue medication, a reminder that a placebo is far from a cure. At the same time, about a third reported meaningful pain relief and relief of their most bothersome symptom, which shows how much movement there can be even without an active drug.

What Drives the Response

The meta-regressions pointed to a handful of factors that consistently shaped how large the placebo response was.

Two patterns stand out. First, trials that enrolled fewer people with severe baseline headache saw greater placebo responsiveness, which fits the idea that milder attacks are more likely to ease on their own. Second, when participants had a lower probability of being assigned the real drug, the placebo response went up. Expecting that you are more likely to get something inactive should, in theory, lower response, so this finding underlines how design choices can quietly shift results.

Exploratory analyses also found higher placebo responses among non-Caucasian participants, in trials run in Asian countries, and in trials of gepants, the newer class of CGRP-targeting drugs. These were exploratory findings rather than firm conclusions, so they are best read as signals worth studying further.

Why the Response Keeps Rising

One of the most striking findings is that placebo response has grown steadily over time, and the trend held up even after adjusting for other variables. A trial run today tends to show a higher placebo response than a similar trial from the 1990s.

Several forces could be behind this. Newer trials often use more polished protocols, more frequent patient contact, and more elaborate delivery devices, all of which can raise expectation. Growing public awareness of migraine and its treatments may also lift how much people expect from any intervention they receive in a study. Whatever the cause, a rising placebo bar makes it harder for a genuinely effective drug to separate from placebo, which has real consequences for how new treatments are tested and approved.

Placebo response is a substantial and increasing component of outcomes in acute migraine trials, and better trial design, standardized communication, and transparent reporting can help preserve the ability to detect a true drug effect.

What It Means for You

For someone living with migraine, these numbers are worth holding onto in a practical way rather than a cynical one.

If a new treatment helps you, that benefit is real and worth keeping, whatever mix of drug and context produced it. The placebo response describes averages across trials, not a verdict on your personal experience. At the same time, understanding that roughly 1 in 9 people reach full pain freedom on placebo can make you a more informed reader of headlines. When a drug reports, say, a 30 percent pain freedom rate, the number that matters is how far it rises above placebo, not the raw figure on its own.

The context effects also cut in a helpful direction. The same forces that inflate placebo responses, such as clear expectations, a calm setting, and a delivery method you trust, are part of why unhurried, supportive care tends to work better. That is an argument for good communication with your clinician, not for abandoning proven treatments.

Important Caveats

A meta-analysis is only as good as the trials inside it, and there are limits worth keeping in view.

• Trial data, not clinic data. These findings describe controlled trials with strict entry criteria and close monitoring. Everyday care looks different, so the exact percentages will not map perfectly onto your own treatment.
• Response is not pure placebo effect. Much of the improvement reflects attacks resolving naturally and regression to the mean, not the power of suggestion alone.
• Exploratory findings need confirmation. The differences by ethnicity, region, and drug class are early signals, not settled facts.
• Averages hide variation. The wide prediction interval, from 4 to 27 percent, shows how much any single trial can differ from the pooled estimate.

Key Takeaways