Prospective Cohort Study of Routine Exercise and Headache Outcomes among Adults with Episodic Migraine

Abstract

Objective:

To evaluate the association of routine exercise with headache frequency, intensity, and duration among adults with episodic migraine.

Background:

A comprehensive management plan for episodic migraine must aim at reducing disability and cost of care; to do so requires optimizing acute and preventive medications, and behavior changes. Prophylactic medication use is associated with adverse events and contraindications with comorbidities. Aerobic exercise is reported to decrease migraine frequency. However, no study has evaluated a potential synergistic relation between regular exercise and preventive medication use among patients with episodic migraine.

Design and Methods:

This was a secondary analysis of data from a prospective cohort study of adults with episodic migraines. In that study, adults with episodic migraine (using International Classification of Headache Disorders–3 (ICHD-3) criteria confirmed by study physicians) were recruited from three academic medical centers in Boston, MA. At baseline, participants provided information on exercise, clinical and demographic characteristics, and lifestyle behaviors. We prospectively collected daily information on headaches and health behavior over at least 6 weeks using electronic questionnaires from 94 participants. We assessed the association between baseline self-reported moderate-vigorous exercise at least three times per week, at baseline, and prospectively recorded headache frequency, intensity, and duration. We further assessed whether these associations differed by the prevalent use of prophylactic migraine medication.

Results:

94 of 98 eligible participants were used in the analysis as 4 participants had missing data on routine exercise frequency or intensity at baseline. On average, patients who reported moderate-vigorous exercise at least three times per week at enrollment had 1.5 fewer headache days per month (−1.5 headache days/month; 95% confidence interval (CI) −3.1 to 0.1) compared to less exercise; this was not statistically significant (P = 0.066). The association between exercise and headache days per month varied by baseline use of migraine prophylaxis (P_interaction = 0.009). Among those who reported regular use of migraine prophylaxis, a report of moderate-vigorous exercise at least three times per week was associated with 5.1 fewer headache days (−5.1: 95%CI −8.2 to −2.0; P = 0.001) compared to those who reported lower levels of exercise. However, among those not using migraine prophylaxis, we observed only 0.4 fewer headache days per month (−0.4: 95%CI −2.2 to 1.3; P = 0.636) associated with moderate-vigorous exercise at least three times/week, a result that was not statistically significant. There were no association of self-reported moderate-vigorous exercise at least three times a week with headache intensity or duration.

Conclusion:

In patients with episodic migraine, those who reported moderate-vigorous exercise at least three times per week had fewer headache days per month, though not statistically significant. This association was significantly stronger in those who used prophylactic medication for migraines. Exercise appeared not to be associated with the severity or duration of headaches. Routine moderate-vigorous exercise may be an important adjunctive strategy for improving headache burden in patients eligible for migraine prophylaxis.

INTRODUCTION

Migraine is a common headache disorder that negatively impacts multiple facets of daily life, affecting about 14.4% of adults worldwide.¹ A comprehensive management plan must optimize acute and preventive treatments to reduce disability and cost of resource utilization.^2–4 Despite availability of efficacious prophylactic medications, their use is limited by side effects and contraindications.⁵ Therefore, identifying behavioral factors that may prevent headaches and reduce impairment is warranted.

Aerobic exercise is frequently recommended as an adjunctive prevention strategy.⁶ Cross-sectional studies have shown associations between low physical activity levels and higher prevalence of migraines and other headaches.^7–9 Findings from prospective studies in headache populations suggest routine exercise is associated with a lower frequency of migraine headache recurrence.^10–12 According to a systematic review of five randomized controlled trials and one controlled clinical trial, there was moderate-quality evidence that aerobic exercise decreases migraine frequency, but low-quality evidence of a small, and likely negligible, reduction in migraine intensity and duration.¹³

A management approach that appropriately maximizes a potential synergistic interaction between routine aerobic exercise and prophylactic medication could improve the functional capacity of patients. No study has evaluated a potential synergistic effect between regular exercise and the use of preventive medication in episodic migraines.¹³ Given the obstacles to medication optimization, there is a need for further evidence on whether exercise can be used as an adjunct to medications to decrease headache burden.

Therefore, we conducted a secondary analysis of data from 94 of 98 patients with episodic migraines who enrolled in a prospective cohort study to investigate the associations between routine moderate-vigorous exercise at least three times per week reported at baseline, and prospective headache frequency, intensity, and duration. The prospective cohort study was designed to assess the associations between sleep and other health behaviors, and the occurrence and severity of migraine headaches during follow-up.^14,15

A priori, we hypothesized that frequent moderate-vigorous exercise is associated with lower frequency, intensity, and duration of headaches in the following six weeks compared with lower levels of exercise. Additionally, we hypothesized that this association would vary according to baseline use of prescription medication prophylaxis at the time of enrollment.

DESIGN AND METHODS

Study Sample

From March 2016 to October 2017, we recruited participants from three academic medical centers in Boston, MA - Beth Israel Deaconess Medical Center (BIDMC), Brigham and Women’s Hospital, and Massachusetts General Hospital. Participants met the criteria for migraine with or without aura following the International Classification of Headache Disorders (ICHD-3) confirmed by study physicians.¹⁶ Patients were excluded if they reported 15 or more headaches per month. Full details of the inclusion and exclusion criteria were described previously.¹⁵ A total of 131 individuals completed the screening visit, 101 met inclusion criteria, and 3 participants with less than 21 days of follow-up were excluded from all analyses (exclusion criteria), leaving 98 participants for all data analyses.¹⁵

At enrollment, each participant completed questionnaires on demographics, medical history, lifestyle factors, and psychosocial factors. Every morning and evening for the subsequent 6 weeks, participants completed online diaries about health behaviors and migraine headaches using the Research Data Capture (REDCap) system hosted by BIDMC.¹⁷ For this analysis, we further excluded participants with incomplete data on both exercise frequency and intensity, resulting in 94 participants in the analytic sample.

All participants provided written informed consent. The BIDMC Committee on Clinical Investigations approved the study.

Baseline Characteristics

At baseline, we collected information on age, sex, ethnicity, race, and employment. Participants completed the Headache Impact Test-6 (HIT-6)¹⁸ to assess the burden of disability by episodic migraine headaches. They reported their typical daily caffeine consumption, typical weekly alcohol consumption, medical history, depression symptoms (Center for Epidemiological Studies-Depression 20 (CES-D 20)),¹⁹ psychological stress (Perceived Stress Scale-10 (PSS-10)),²⁰ and their typical medication use. At the time of enrollment, participants reported whether they took prescription migraine prophylactic medications for episodic migraine (i.e., beta-blocker, topiramate, tricyclic antidepressant, selective serotonin-norepinephrine reuptake inhibitor)⁵. We did not include dietary supplements for migraine such as magnesium in our definition.⁵

Headache Characteristics

For at least six weeks after the baseline visit, participants completed morning and evening diaries to report the presence of any headache since their prior diary entry, including the time of onset and resolution, pain intensity, and medications used for the headache.¹⁵ When headache onset was reported within one calendar day of a prior headache’s resolution, it was considered a relapse, and the duration was calculated from the start of the first headache to the end of the last headache reported that day. For each headache, participants rated the maximum pain intensity on a horizontal visual analog scale (VAS, millimeters) from 0 (no pain at all) to 100 (worst pain in their life).

For every participant, we calculated the number of headache days per month by dividing their total headache days by the number of observation days and multiplying by 30. We also calculated the number of headaches per month, average pain intensity, and average duration of headaches.

Baseline Routine Exercise

At baseline, participants reported their: (1) routine exercise frequency as 0, 1 to 2, 3 to 4, or at least 5 times per week; and (2) typical intensity of exercise as light (normal walking, walking downstairs, yoga, gardening, etc.), moderate (brisk walking, lawn mowing, shoveling, dancing, etc.), or vigorous (jogging, running, cycling fast, kickboxing, etc.).

Exercise recommendations for people with migraine headaches and prior intervention studies examining the effect of exercise on headache have suggested that aerobic exercise for at least 40 minutes three times per week may be beneficial^10,21,22. In our primary analysis, we assessed self-reported baseline routine exercise as a binary exposure of moderate-vigorous exercise at least three times per week versus less than three times per week. The self-report of habitual exercise frequency mimics information collected in a typical clinical encounter. The dataset used for this study is available from the corresponding author on request.

Statistical Analysis

Since this is a secondary analysis of existing data, no statistical power calculation was conducted, and the sample size was based on the available data. We calculated the means, standard deviations (SDs), and proportions of participants with particular characteristics at baseline in the total sample and by moderate-vigorous exercise at least three times a week as reported at baseline. For participants who used prophylactic medication at the time of enrollment, we calculated the proportions that used each of the classes of medication.

In a complete-case primary analysis, we assessed the association between baseline self-reported moderate-vigorous exercise at least three times per week, and prospectively recorded headache frequency, intensity, and duration. The first model adjusted for age and sex. Additionally, we adjusted for menstrual status (premenopausal), PSS-10 score (upper quartile vs lower three quartiles), CES-D 20 score (≥16 indicating higher depressive symptoms), baseline HIT-6 score, weekly alcohol servings (continuous), daily caffeine servings (continuous), smoking history (ever vs never use), prophylactic medication, and baseline number of headaches per month. We evaluated whether the association between moderate to heavy exercise at least three times per week and headache days per month differed among those who did or did not use prophylactic medication by including an interaction term for self-reported exercise and prophylactic medication. We plotted expected marginal means of headache days per month by prophylactic medications and moderate-vigorous exercise simultaneously using the marginsplot command in Stata/MP 16. In a post hoc sensitivity analysis, we recategorized exercise behavior into an order of (1) at most light intensity exercise of any frequency in a week (2) moderate-vigorous exercise 1–2 times per week (3) moderate-vigorous exercise 3–4 times per week (4) moderate-vigorous exercise ≥5 times per week. We evaluated whether there was a dose-response relationship between frequency of moderate/vigorous exercise and headache days per month after adjusting for the covariates listed above We conducted all analyses using SAS Studio 3.8 University Edition (SAS Institute Inc., Cary NC) and Stata/MP 16 and used a two-tailed significance level of 0.05 as criteria for statistical significance.

RESULTS

At enrollment, three participants did not report both the frequency of exercise and intensity of exercise; one reported exercising at least five times a week but did not report intensity. These four participants were excluded from this analysis. Table 1 shows the baseline characteristics of the 94 participants. The mean age was 34.3 years (standard deviation (SD) 11.4). 82 of the 94 participants were women, of which 66 (70%) were premenopausal. 63% (59 of 94) of the analytic sample reported typically engaging in moderate to heavy exercise at least three times per week. On average, participants reported 5.0 (SD 3.5) headaches per month, HIT-6 score of 60.8 (SD 6.1); and 27% (25 patients) used prescription prophylactic medication.

Table 1.

Baseline characteristics of 94 participants with episodic migraine in the greater Boston area, Massachusetts, 2016–2017.

Characteristics	Totala(n=94)	Self-reported moderate-heavy exercise <3/weeka(n=35)	Self-reported moderate-heavy exercise ≥3/weeka(n=59)	P-value
Age, years	34.3 ± 11.4	35.7 ± 11.3	33.5 ± 11.5	0.362
Female	82 (87%)	32 (91%)	50 (85%)	0.525
Non-Hispanic White	68 (72%)	20 (57%)	48 (81%)	0.011
Premenopausal	66 (70%)	24 (69%)	42 (71%)	0.261
Employed	76 (81%)	26 (74%)	50 (85%)	0.213
Never-smoker	81 (86%)	26 (74%)	55 (93%)	0.014
Prophylactic medication use	25 (27%)	8 (23%)	17 (29%)	0.528
Beta blockers	6 (24%*)	2 (6%)	4 (7%)
Calcium channel blockers	1 (4%*)	0	1 (2%)
Topiramate	12 (48%*)	7 (20%)	5 (8%)
Tricyclic antidepressant	7 (28%*)	2 (6%)	5 (8%)
SSNRI	5 (20%*)	2 (6%)	3 (5%)
HIT-6 score (Range : 36–78)	60.8 ± 6.1	62.0 ± 5.6	60.0 ± 6.4	0.122
Baseline headaches per month	5.0 ± 3.5	5.6 ± 3.3	4.6 ± 3.6	0.183
CES-D 20 score (Range : 0–60)	9.7 ± 7.7	10.8 ± 8.9	9.1 ± 7.0	0.305
PSS-10 score (Range: 0–40)	14.5 ± 6.3	15.0 ± 6.7	14.2 ± 6.2	0.554

Abbreviations : CES-D 20 (Center for Epidemiological Studies - Depression 20); HIT-6 (Headache Impact Test-6); PSS-10 (Perceived Stress Scale 10); SSNRI (Selective Serotonin-Norepinephrine Reuptake Inhibitor)

^aMean ± standard deviation or N (%).

NB: The total number of participants using the different classes of prophylactic medication is greater than 25 because some participants were on more than one class of medication.

^(*): percentage of prophylactic medication users who use the medication type.

Fisher’s exact test used for sex, menstrual status, and smoking history. Chi-square test for race/ethnicity, employment status, and prophylactic medication use. Two-sample t-test for age, HIT-6, baseline headaches per month, CES-D 20 score, and PSS-10 score.

Missing data: ethnicity (4).

Participants had on average 46.6 days of observation (SD 7.3; Range 38 to 84 days). During the 4219 days of prospectively collected follow-up data, participants reported 804 discrete headaches, representing an average of 8.6 headaches per person. On average, participants reported 7.6 (SD 4.4) headache days per month, duration of 8.9 (SD 8.2) hours, and headache intensity (VAS) of 45.8 (SD 14.8).

After adjusting for potential confounding, participants reporting moderate-vigorous exercise at least three times a week had on average 1.5 fewer headache days per month (−1.5 headache days/month; 95%CI −3.1 to 0.1) compared to less active participants (Table 2). This association was not statistically significant (P = 0.066). The association between exercise and headache days per month varied by baseline use of migraine prophylaxis (P_interaction = 0.009). Among those who reported regular use of migraine prophylaxis, report of moderate-vigorous exercise at least three times per week was associated with 5.1 fewer headache days (−5.1 headache days per month: 95%CI −8.2 to −2.0; P = 0.001) compared to report of lower levels of exercise. However, among those not using migraine prophylaxis, we observed only 0.4 fewer headache days per month (−0.4 headache days per month: 95%CI −2.2 to 1.3; P = 0.636) associated with moderate-vigorous exercise at least three times a week, a difference that was not statistically significant. Figure 1 shows a plot of the predictive margins (average headache days per month) for the exercise-prophylaxis interaction. In a weakly-powered exploratory post-hoc analysis, we did not observe a statistically significant dose-response relationship between frequency of moderate-vigorous exercise and headache days per month (p=0.172). The associations between exercise and average headache intensity and duration were generally null (Table 2).

Table 2.

Association between self-reported routine exercise performance (≥3/week) and headache outcomes

Headache outcome	Mean difference (95% confidence interval)
	Age/sex adjusted	Fully adjusteda
Average headache days per month	−1.8 (−3.4, −0.1)	−1.5 (−3.1, 0.1)
Average duration of headaches (hours)	−0.9 (−4.4, 2.7)	0.3 (−3.5, 4.0)
Average headache intensity (VAS, mm)	−4.6 (−10.6, 1.5)	−0.4 (−6.4, 5.7)

Abbreviation: VAS – visual analogue scale (100 mm)

^aAdjusted for age (years), sex, menstrual status, PSS-10 (upper quartile vs lower quartiles), CES-D20 (≥ 16), alcohol, smoking (ever use), caffeine, prophylactic medication, HIT-6 score, and baseline headaches per month.

Figure 1. A plot of expected headache days per month ( 95% confidence intervals) according to frequency of moderate-vigorous exercise and prophylactic migraine medication use.

The estimates are adjusted for observed values of age, sex, menstrual status, age (years), sex, menstrual status, PSS-10 (upper quartile vs lower quartiles), CES-D20 (≥ 16), alcohol, smoking (ever use), caffeine, HIT-6 score, and baseline headaches per month.

DISCUSSION

In this prospective cohort study of 94 patients with episodic migraine followed for at least six weeks, baseline report of moderate-vigorous exercise at least three times per week was associated with 1.5 fewer headache days per month. Medication prophylaxis was associated with an even greater reduction to about five fewer headache days per month. There were no association of moderate-vigorous exercise at least three times per week with average headache intensity or duration. Our analysis addresses the clinically relevant question of how a physician’s knowledge of a patient’s routine health behavior at a single time-point informs short-term prognosis.

Our results are consistent with prior research and extend the knowledge base. For example, in a non-randomized trial among patients with episodic migraine, Narin et al.¹¹ compared baseline frequency of headache to headache frequency over 8 weeks of follow-up among participants assigned to 60 minutes of exercise three times per week versus usual care. They found a greater reduction in migraine days for those assigned to exercise (7.4 days to 3.6 days) compared to those assigned to usual care (8.9 days to 7.0 days). In a randomized trial of patients with chronic migraine, Santiago et al.²² assigned participants to aerobic exercise 3 times per week with amitriptyline or to amitriptyline alone for 12 weeks. They found a significantly greater decrease in headache days per month in those assigned to aerobic exercise and amitriptyline (23 days to 5 days), compared to those on amitriptyline alone (25 days to 13 days). Our results are also consistent with a 3-arm randomized trial comparing the effects of aerobic exercise three times a week, topiramate, or relaxation for three months.¹⁰ Those in the exercise group reported 1.98 (SD 0.55) fewer migraine days per month. Statistically, the reduction in migraine days per month was similar across the three arms of treatment. None of the prior studies assessed the potential synergistic effect of exercise and prophylactic medication.

The mechanisms underlying the observation that routine aerobic exercise is associated with a reduction in headache frequency may involve both direct and indirect pathways. Exercise may yield benefits via neurovascular and neuroinflammatory processes involved in migraine headaches.²³ Exercise may also lead to changes in nociception and central processing of pain information (neuromodulation) which can improve symptomatology.¹² Additionally, exercise may impact psychological and behavioral factors influencing headache and perception of impairment.^23–25

Like any observational study, ours is not without limitations. We defined exposure according to exercise reported at enrollment, therefore we cannot distinguish between people who changed their habits shortly before study enrollment from people who may have been active or sedentary for several years. It is possible that an individual’s headache characteristics and use of medication prophylaxis at baseline may influence their frequency and intensity of exercise and subsequent experience of headaches; however, our findings were robust to adjustment for baseline headache frequency as recalled by the participants. However, our baseline questions reflect a typical set of questions that could be used in a clinical encounter when assessing lifestyle factors among patients with episodic migraine. These findings suggest that the response to this simple screening question is associated with headache frequency in the subsequent six weeks. Also, the potential of exposure misclassification is likely non-differential for the prospectively collected outcome data and, at worst, may have led to some loss of statistical precision and a bias toward not observing a statistically significant association between routine exercise and headache frequency, intensity, and duration. As in all observational studies, there may be other factors confounding the association between exercise and migraine. However, we did adjust for factors that may be predictors of exercise and headaches, such as age, menstrual status, sex, and baseline headache burden in our multivariable models. The finding that the association varies according to migraine prophylaxis may also be due to other differences between these groups. The study participants were predominantly premenopausal women of European ancestry, without significant medical and psychiatric comorbidities which may limit the generalizability of our findings. Our study also had significant strengths including prospectively collected detailed daily diary information on migraine headache occurrence, duration, and intensity for at least six weeks in each participant.

CONCLUSION

In patients with episodic migraine, those who report moderate-vigorous exercise at least three times per week have fewer headache days per month, though not statistically significant. This association is significantly stronger in those who use prophylactic medication for migraines. Routine exercise appeared not to have an association with the severity or duration of headaches. Routine moderate-vigorous exercise may be an important adjunctive strategy for improving headache burden in patients eligible for migraine prophylaxis. Based on prior studies and these findings, a two-by-two factorial randomized trial is warranted to definitively evaluate the synergistic effect of exercise and prophylactic medication on migraine headaches.

Abbreviations:

ICHD-3

International Classification of Headache Disorders–3

BIDMC

Beth Israel Deaconess Medical Center

REDCap

Research Data Capture

HIT-6

Headache Impact Test-6

CES-D 20

Center for Epidemiological Studies-Depression 20

PSS-10

Perceived Stress Scale-10

VAS

Visual analogue scale (millimeters)Standard deviation (SD)

95%CI

95% confidence interval

P_interaction

P-value of interaction

SSNRI

Selective serotonin-norepinephrine reuptake inhibitor