Introduction

Migraine is a highly prevalent neurological disorder affecting 15% of the global adult population1. Moreover, migraine is the leading cause of years lived with disability worldwide among individuals between 15 and 50 years of age2. A migraine attack is characterized by recurrent episodes of moderate to severe headache, which is typically one-sided and is accompanied by nausea and photophobia3. Migraine is also typically aggravated by, or causes avoidance of, routine physical activity (e.g., walking or climbing stairs), which is also one of the diagnostic criteria3. In some cases, the headache is preceded by a so-called prodromal phase and succeeded by a postdromal phase3,4. Each phase is defined to last up to 48 h, in which fatigue is a frequently reported symptom3,4,5. Thus, physical activity is not only reduced during the headache phase, but also in the non-headache phase in adults with migraine6. Chronic migraine is a subclassification of migraine3, defined by ≥ 15 headache days per month, in which at least eight days are migraine over three consecutive months.

Over the past two decades, clinical trials of migraine treatment have frequently included improvement of different types of migraine-related disability as a secondary endpoint. So far, the Migraine Disability Assessment questionnaire has been used as a patient-reported outcome measure in the clinical trials of migraine treatment7. Recently, the Activity Impairment in Migraine Diary was proposed as a more comprehensive patient-reported outcome measure to monitor activity impairment in adults with migraine8. In the era of wearable health technologies, which are widely available, the time has come to move forward and consider the use of automatically recorded objective outcome measures of activity, which can be used as a supplement to the validated patient-reported outcome measures. Smartphones with built-in pedometer applications are widely available. The question is whether pedometric applications can be used to assess successful treatment effects in migraine patients. In the present study, we used step-per-day data extracted from smartphones as a surrogate marker of physical activity in adults with chronic migraine.

We hypothesized that successful treatment of chronic migraine with monoclonal antibodies (mAbs) against calcitonin gene-related peptide (CGRP) or its receptor would increase daily step counts, representing an increase in physical activity.

Methods

Study design and participants

The present study was a single-center, cross-sectional, retrospective study. The study was approved by the Ethical Committee of Copenhagen (J-19085557) and the local data protection department (P-2019-292). The study was conducted according to the regulations of the Danish Data Protection Agency. All participants provided written consent to participate after detailed written and oral information about the study in agreement with the Declaration of Helsinki of 1964, with later revisions.

We recruited participants from the outpatient clinic at the Danish Headache Center (Department of Neurology, Copenhagen University Hospital—Rigshospitalet) from 2022 August 27 to 2022 October 17. We approached adults with chronic migraine3 who were undergoing treatment with CGRP mAbs (i.e., erenumab or fremanezumab). Patients were eligible for inclusion if they experienced successful treatment with CGRP mAbs and any iPhone version with the Apple Health application activated for at least six months before and at least three months after treatment initiation. Exclusion criteria were change of iPhone within the data collection period, no collected data, change of job, new onset of serious illness, or major surgery during the observational period.

Successful treatment with CGRP mAbs was defined as ≥ 30% reduction in the monthly (28 days ± 2 days) migraine days (MMDs) at 3-month treatment compared to the baseline month (i.e., one month before treatment initiation). This definition of successful treatment was based on clinical guidelines at the center9.

Data extraction

The participants were instructed to access their Apple Health application and find the annual overview of the monthly average number of steps per day (Fig. 1). Subsequently, participants were asked to find the month in which they started on CGRP mAbs and provide the monthly average of steps per day for the 3 months before (i.e., baseline) and 3 months after the month of treatment start. Monthly headache days (MHDs) and MMDs were extracted from hospital medical records and activity data were collected by two investigators (FTJ and CKC).

Figure 1
figure 1

Data extraction from pedometer applications on patients’ smartphones. The average number of steps per day was extracted for each of the three months before and after the month the treatment was started. The figure shows a screenshot of the Health application on the Apple iPhone. Months (October to September) are displayed continuously on the X-axis, and number of daily average steps on the Y-axis. The daily average for a month is provided by clicking on the bar (e.g., 5815 steps in April 2019 in the shown case). D, day; W, week; M, month; 6M, six months; Y, year.

Full size image

Statistical analysis

The sample size calculation was based on detection of at least a 30% increase in steps per day (average of three months) between the baseline period and the successful treatment period, at 5% significance (2-tailed) with 80% power. We estimated 50% variation in the average steps per day. We calculated that at least 22 patients should be included.

All absolute values are presented as median with interquartile range (IQR). Percent changes are reported as mean with 95% confidence intervals (CI).

The primary endpoint was a difference in the steps per day (average of three months) before and after successful treatment with CGRP mAbs in adults with chronic migraine. The secondary endpoint was a correlation between the percentagewise increase in steps per day and percentagewise reduction in MMDs at Month 3.

We tested differences in the average daily steps per month between baseline and the 3-month treatment period with Wilcoxon signed rank test (two-tailed), as data were not normally distributed. Correlation analyses were performed using Spearman’s rho (two-tailed). We used SPSS (version 28·0) for all statistical analyses. A level of significance of 0.05 was accepted for each comparison.

Results

We included 22 (female, n = 20; male n = 2) participants (Fig. 2). Median age was 48.5 years (IQR: 45.0–57.0). Median baseline MHDs was 20.5 (IQR: 16.0–28.0). Median MMDs at Month 3 was 4.2 (IQR: 1.8–5.4) compared to baseline 14.0 (IQR: 10.0–19.3). All participants had at least 30% reduction in MMDs at month 3 after treatment initiation compared to baseline.

Figure 2
figure 2

Flowchart over the study.

Full size image

Change in daily step counts between the baseline and the treatment periods

The average steps per day between baseline (4421, IQR: 2847–6336) and 3-month (5241, IQR: 3709–7983) were increased (P = 0.039) (Fig. 3). The mean percentagewise increase in daily steps was 21.3% (95% CI 0.5–42.1) in the total study population.

Figure 3
figure 3

Steps per day before and after treatment start. Average steps per day (SPD) across 3 months before (blue) and 3 months after (red) initiation of treatment with monoclonal antibodies against calcitonin gene-related peptide or its receptor was investigated in patients (n = 22) with successful treatment effect (≥ 30% reduction in the monthly migraine days). There was an increase in SPD after successful treatment compared to before treatment initiation (P = 0.039). The whisker plot shows medians (before: 4421; after: 5241), and interquartile range (before: 2847–6336; after: 3709–7983) of STP with minimum (before: 1889; after: 1435) and maximum (before: 10,111; after: 10,763) SPD values, and outliers (before: 12,869 SPD; after: 15,511 SPD).

Full size image

Correlations between daily step count and treatment efficacy

We found a positive correlation between responder rate (RR) and the percentagewise change in average steps per day (correlation coefficient 0.521, P = 0.013). In explorative analyses, we found a difference between the subgroup of participants with < 70% RR and the subgroup with ≥ 70% RR (P = 0.041, Mann–Whitney U, [2-tailed]).

Discussion

This study used an objective tool to assess the treatment effect on chronic migraine-related disability. The main finding was that the average steps per day increased noticeably (21.3%) after initiation of successful treatment with CGRP mAbs compared to baseline in adults with chronic migraine. Moreover, an increase in average steps per day was positively correlated with responder rates.

The steps per day count represents the level of physical activity. The recommendation for young adults between 35 and 50 years of age is 7000 steps per day10. A large-scale study of 717,527 individuals from 111 countries collected pedometric data from smartphone applications. The study demonstrated that the average number of steps per day for most European Union countries was at least 5000–550011. Based on similar smartphone data from 3924 individuals (median age of 33 years), this study reported mean 5263 steps per day in a background population in Denmark11. Our study population was a representative sample of Danish adults with chronic migraine treated at the Danish Headache Center who fulfilled the Danish criteria for successful treatment with mAb and owned a smartphone9. This may limit the generalization drawn from these results, although 85% of individuals between 16 and 85 years of age in Denmark report owning a smartphone12. Compared to the background population, our population with chronic migraine had substantially reduced level of physical activity at baseline. The average at baseline was 4421 steps per day before treatment, which increased to 5241 steps per day over a 3-month successful treatment period in our study. This finding merits further investigation in a more generalized migraine population.

Migraine is, contradictory to other common primary headache types, characterized by the aggravation of symptoms by routine physical activity. Consequently, adults with migraine may avoid physical activity during attacks. Reduced physical activity is associated with cognitive symptoms, risk of depression and anxiety13, which are also comorbidities associated with chronic migraine1. Our study demonstrated that successful treatment of migraine increases the level of physical activity in adults with chronic migraine. Interestingly, successful treatment of chronic migraine may not only reduce pain days, but may most probably also prevent inactivity-related risk of comorbidities. We also demonstrate a positive correlation between reduction of monthly migraine days and increase in the average steps per day. The correlation coefficient was 0.521, which is considered a moderate positive correlation. We therefore suggest implementation of adjunct use of pedometric applications to monitor treatment response in patients with chronic migraine. Further, pedometric measurement may be a simpler and more objective patient-reported outcome than counting migraine pain days per month. Thus, this may be a supplementary way to monitor treatment effects, as pain days do not consider the disability during or outside of attacks, which is a major issue in adults with migraine2. Moreover, implementation of an automatic monitoring tool may also save patients from recording data in headache diaries and patient-reported outcomes (e.g., on days with severe migraine attacks or postdromal fatigue) which may lead to better compliance.

Strengths and limitations

The major strength of our study was that none of the participants knew they would be asked to show their smartphone data retrospectively. Some participants were unaware that their daily steps were recorded by their smartphone app. Individuals who track their steps take more steps per day than those who do not, which could have biased the results14. Another strength of the study was that the daily step count was recorded automatically in the Apple Health app and thus did not depend on manual entries, which could have caused missing data.

One limitation is that we were not able to extract data on the daily step count but as average steps per day for each month. Moreover, we only had data on the summary of monthly headache and migraine days. Consequently, we were not able to perform correlation analysis between daily pain intensity and step count, which could have shed light on the potential inactivity on the day before and/or after a pain day. Moreover, it was a limitation that we did not know the patients’ smartphone use pattern. We do not believe such habits changed over the three months unless there was an obvious reason (e.g., change of job or serious illness), which is why we included such potential covariates in our exclusion criteria for the study. The step count was not assessed beyond three months of treatment in the pilot study. In future studies, long-term effects on daily steps must be evaluated. We recommend that the study is verified with a more precise and repeatable means of step measuring (e.g. wearable pedometers) while using prospective data collection. Lastly, this study did not include a control group of patients who were treated during the same period with no effective response to treatment. This would have revealed if walking habits were generally changed in the study population. This aspect may be investigated further in future studies as well. We are confident that a repeated study with a control group would confirm our conclusions, as there was a positive correlation between increase in daily steps and the percentagewise treatment response.

Conclusion

This study evaluated physical activity as a surrogate marker of successful treatment response to CGRP mAbs in adults with chronic migraine. Physical activity was assessed as daily steps, which were below the expected number at baseline before treatment initiation. Successful treatment with at least 30% reduction in monthly migraine days at 3 months was positively correlated with 21% increase in the average steps per day. Automatically registered steps per day on wearable pedometers may be a useful marker of treatment response in adults with chronic migraine in future studies and clinical settings.