The migraine mystery: The prime suspect has been freed

Serena Orr, MD, MSc, Director Pediatric Headache Program, Alberta Children’s Hospital and Alec Mian, PhD, CEO Curelator, Inc.

Serena Orr MD, MSc
Alec Mian PhD, CEO

Stress is the most commonly suspected trigger of migraine. Multiple studies over the last few decades have implicated a link between stress and the start of an attack of migraine. For example, in one commonly cited study, about 80% of 1,750 patients surveyed reported perceived stress as the most common trigger1. Studies have also found that stress is the most commonly suspected trigger of attacks of migraine in children and adolescents.2,3

But is stress really a trigger of migraine - or do people with migraine have stress because they have migraine? The long held belief that stress triggers attacks of migraine is based on studies, like those cited above, where people were asked in retrospect what they thought their migraine triggers were. The problem with this approach is that we, as humans, are subject to error (the so-called “recall bias”) when it comes to making associations about things that have happened in the past.

Perhaps instead of asking individuals to guess what triggers their attacks of migraine, a better approach might be to first let individuals report their day-to-day levels of perceived stress and then simply observe when attacks of migraine occur. If we then look at the data on an individual level, we can objectively establish an individual’s stress-migraine relationship. We can make observations about a single individual and ask: what percentage of that individual’s attacks are preceded by changes in perceived stress? What percentage of that individual’s headache-free days are preceded by the same changes in perceived stress? This individual relationship is important. Most studies focus on establishing conclusions that are true for the average person. However, very few people should be concerned about the average person because very few people are in fact average4.

We are now able to take this more precise and accurate approach to studying what causes an attack of migraine to start. We owe this new approach to advances in our ability to digitally collect and analyze high-resolution data reported by individuals with migraine on a daily basis for a long time. This way of studying migraine triggers, which focuses on the individual, forms the basis of “personalized” or “precision” medicine. If we can understand key individual differences in migraine triggers, then we can help each person to focus on treatment approaches that will be most successful for that person.

When we used this personalized approach to study the relationship between perceived stress and attacks of migraine, we were surprised by the results5. While we were expecting that the majority of attacks would be triggered by stress, what we found was quite different. We actually found that only 7.1% of individuals had a majority of their attacks associated with increasing levels of perceived stress prior to the start of their attacks of migraine. A smaller number, 3.4%, had the majority of their attacks associated with decreasing levels of perceived stress, and the largest, 61.5%, proportion of individuals had no change in the level of perceived stress before the start of most of their attacks of migraine. The other 28% of individuals in the study each had a mix of increasing, decreasing, or unchanging patterns of stress before their attacks. When looking at all attacks across individuals, only 24% of those fell into the “stress as a trigger” pattern, where attacks were preceded by increasing perceived stress levels.

Results of study- Increasing stress levels, flat stress levels, decreasing stress levels and mixed stress levels. Results of study- Increasing stress levels, flat stress levels, decreasing stress levels and mixed stress levels.
Only 7.1% of individuals had 50% or more of their attacks preceded by increased levels of perceived stress, while most individuals, 61.5%, had at least 50% with no changes in their stress levels preceding their attacks. Only 3.4% of people had attacks after decreasing levels of stress. The remaining portion of individuals, 28%, had a non-dominant mixture of increasing, flat and/or decreasing levels of stress preceding their attacks.

What is happening here? These results seem contrary to our collective wisdom, experience, and to prior research. However, the study itself is very large and rigorously conducted. In fact, historically this is the largest observational cohort study to examine the relationship between perceived stress and migraine to date: both in number of participants (n= 351) as well as study duration (daily data collected for three months). During the course of the study, participants had a total of 2,115 migraine attacks, each of which were analyzed with respect to individuals’ daily stress levels before, during, and after the attack. With this large, high resolution, long daily collection data set, we took an individual-focused approach and learned that previous beliefs about stress as a migraine trigger are overstated. Most individuals do not have the start of their attacks triggered by stress. However, there is a small proportion of individuals for which this may in fact be true, and we were able to identify them.

Why is this important? We believe that taking the type of approach used in this study to understand an individuals’ stress-migraine relationship will help to drive a personalized approach to treatment. At present, the same clinical advice around stress management is typically offered to all patients regardless of the patient’s individual profile. However, a more precise and beneficial approach could be taken if we knew more about each individual’s profile. For example, for those 7.1% of individuals who do show the “stress as a trigger” pattern, a greater focus on investing time and resources into evidence-based treatments that target coping strategies or stress reduction, such as mindfulness-based stress reduction, biofeedback, or relaxation techniques, is likely warranted. Conversely, perhaps those strategies would not yield the same benefits in individuals who do not show a stress-migraine relationship in their individual data.

We hope for a future where individuals with migraine can use their personal data and analysis to discover their most effective coping and treatment strategies, based on their own unique profile. This approach has the potential to significantly improve treatment outcomes for individuals with this disabling chronic disease.

Go back to articles


  1. Kelman, L., The triggers or precipitants of the acute migraine attack. Cephalalgia, 2007, 27, 394–402
  2. Neut D, Fily A, Cuvellier JC, Vallee L. The prevalence of triggers in paediatric migraine: A questionnaire study in 102 children and adolescents. J Headache Pain 2012;13(1):61–5.
  3. Solotareff L, Cuvellier J-C, Duhamel A, Vallée L, Nguyen The Tich S. Trigger factors in childhood migraine: a prospective clinic-based study from north of france. J Child Neurol 2017;32(8):754–8.
  4. Mian, A., MacGregor, A,. Curelator blog, 2017, Will the Real Mr. Average Please Stand Up?
  5. Marina Vives-Mestres, PhD ; Amparo Casanova, MD, PhD; Dawn C. Buse, PhD; Stephen Donoghue, PhD; Timothy T. Houle, PhD; Richard B. Lipton, MD; Alec Mian, PhD; Kenneth J. Shulman, DO; Serena L. Orr, MD, MSc. Patterns of perceived stress throughout the migraine cycle: a longitudinal cohort study using daily prospective diary data. Headache 2021;61(1):90-102.