Abstract
Objective
Migraine is comorbid with restless legs syndrome (RLS). However, the temporal association between these two episodic disorders remains elusive. The current study investigated the temporal relationship between migraine and RLS attacks.
Methods
Migraine patients with RLS were recruited from a headache clinic. Patients with symptomatic RLS, RLS mimics, daily headaches, or daily RLS attacks were excluded. The patients recorded their headaches and RLS attacks for two weeks in a diary. The severity of each headache or RLS attack was rated on a four-point (0−3) Likert scale. Logit-normal, random-effects models were employed to estimate the odds ratios (ORs) for the temporal association between migraine and RLS attacks.
Results
Thirty migraine patients with RLS (28 F/2 M, mean age 35.5 ± 9.0 years) completed the study. On the basis of 420 daily diary records, migraine attacks were associated with subsequent RLS attacks occurring on the same and next nights (OR = 6.94, 95% confidence interval (CI) = 4.39–11.0 and OR = 3.00, CI = 1.92–4.68; both p < 0.001). RLS attacks were associated with subsequent migraine attacks only on Day 1 (OR = 1.97 (CI = 1.3–2.98; p = 0.01). Overall, the frequencies of migraine and RLS attacks in two weeks were correlated (Spearman’s correlation = 0.56, p = 0.001).
Conclusions
Our study results showed a bidirectional triggering association between migraine and RLS attacks. The association was stronger and lasted longer for migraine triggering subsequent RLS than that for vice versa.
Introduction
Migraine and restless legs syndrome (RLS) are chronic neurological diseases involving episodic attacks (1,2). Studying the associated comorbidities provides insight into the underlying pathophysiology and potential treatment strategies for application in clinical practice (3,4). In recent years, the comorbid relationship between migraine and RLS has been documented in clinical- and population-based studies (5–10). There are some similarities between these two disorders: e.g. female dominant, episodic attacks, fluctuations in levels of certain biogenic amines (i.e. dopamine and serotonin) during attacks postulated as crucial factors in the pathophysiology (11–13). However, clinically, the temporal relationship between migraine and RLS attacks has never been examined. The purpose of this study was to determine whether a migraine attack may trigger a subsequent RLS attack or vice versa.
Participants and methods
All migraine patients who visited the headache clinic at Lin-Shin Hospital, Taichung, Taiwan, were questioned about their RLS symptoms. One physician (Chen PK) interviewed, examined and diagnosed all patients. Migraine was diagnosed according to the International Classification of Headache Disorders, second edition (ICHD-2) (14). RLS was diagnosed in patients who met all four essential criteria proposed by the International RLS Study Group (15).
Migraine patients with comorbid RLS were invited to participate in the current study between April 1, 2012 and March 31, 2013. Patients were excluded if they experienced daily headaches or daily RLS in the past month (either condition can confound the potential temporal relationship between the disorders), were on preventive medication for migraine or RLS, refused to record their observations in the diary during the study period, or refused to join the trial. All patients underwent magnetic resonance imaging. Detailed medical histories were obtained and neurological examinations and blood tests were performed to exclude patients with intracranial lesions, symptomatic RLS, or RLS mimics (16).
Demographic data and clinical profiles of migraine and RLS were collected from the patients. During the first two-week study period, the patients were prohibited from taking any medications except for abortive treatments prescribed for acute headache attacks or medications for chronic medical diseases. They then recorded the onset and severity of headaches and RLS attacks on a daily basis in a diary. We contacted the patients by telephone to remind them to record their observations in the diary every day. Headache and RLS severity were individually rated on a four-point Likert scale, ranging from 0−3 (0-none, 1-mild, 2-moderate, and 3-severe). Moreover, headache profiles and painkiller usage were recorded, including headache duration, unilateral or bilateral location, pulsating quality, headache aggravation by or causing avoidance of routine physical activity, nausea, vomiting, photophobia and phonophobia. On the basis of their diary entries, we categorized the headache attacks into migraine according to the ICHD-2 criteria. Because headache characteristics may change after the use of abortive medications, we defined headaches not satisfying the ICHD-2 migraine diagnosis criteria as migraine attacks if the patients underwent abortive treatment (17).
We adopted the method proposed by Lipton et al. to examine the temporal relationship between migraine and RLS (18). First, we tested whether the migraine attacks triggered subsequent RLS on the same or the following nights. We classified the triggering association into three categories of occurrence and defined them as follows: migraine-nocturnal RLS on Day 1: both migraine and nocturnal RLS attacks before bedtime, with the RLS attack occurring after the migraine attack; migraine-nocturnal RLS on Day 2: RLS attack on the second night after the migraine attack, and migraine-nocturnal RLS on Day 3: nocturnal RLS attack on the third night after the migraine attack. Furthermore, we examined the triggering association between RLS attacks and subsequent migraine. Because RLS attacks occur in the late evening or at night, we examined the relationship between nocturnal RLS and migraine attacks in the three days after the attack. We defined the relationships as follows: nocturnal RLS-migraine on the next day: nocturnal RLS attack and migraine attack on the day after the RLS attack; nocturnal RLS-migraine on Day 2: migraine attack before bedtime on the second day after the RLS attack; and nocturnal RLS-migraine on Day 3: migraine attack before bedtime on the third day after the RLS attack (Figure 1).
Association between migraine and restless legs syndrome (RLS) attacks. (a) Association (OR with 95% confidence intervals) between migraine and subsequent RLS attacks over the following three nights. (b) Association (OR with 95% confidence intervals) between RLS and subsequent migraine attacks over the following three days.
The participants were informed that their diary records would be used for evaluating the frequency and severity of their migraine and RLS attacks before treatment. However, they were not informed that these records would be used in investigating the temporal relationship between migraine and RLS. The Institutional Review Board of Lin-Shin Hospital, Taichung, Taiwan, approved the study protocol. Informed consent was obtained from all patients before including them in this study.
Statistical analyses
SAS Version 11 for Windows was used for statistical analyses. Migraine and RLS occurrence were modeled as binary outcomes. A logit-normal, random-effects model was employed to estimate subject-specific odds ratios (ORs) with 95% confidence intervals (CIs) for the triggering association between migraine attacks and subsequent nocturnal RLS attacks within specific time frames (the first (on the same day), second and third nights) (18). This model was also used for examining RLS attacks and the subsequent migraine attacks on Days 1, 2 and 3. Spearman’s rank correlation coefficient (rs) was used for determining the association between frequencies and the severity of migraine and RLS attacks. The chi-square test was used for testing differences in the categorical data. In this study, statistical significance was set at two-tailed p < 0.05.
Results
Patient demographics, headache profiles, and RLS symptoms.
RLS: restless legs syndrome; IRLSS: International RLS Study Group Rating Scale; F: female; M: male.
Mean ± standard deviations (range).
We collected 420 daily diary entries from the two-week patient diary records. Overall, these patients reported 182 headache days (43.3%), including 133 migraine (31.7%) and 171 RLS (40.7%) days.
After combining the various diary entries, we first examined the relationship between migraine attacks and subsequent RLS attacks in the following three nights and evaluated the OR (Figure 1(a)). For migraine-nocturnal RLS on Day 1, the results showed a significantly increased risk of nocturnal RLS occurrence following migraine (OR: 6.94, 95% CI: 4.39–11.00, p < 0.0001). Regarding migraine-nocturnal RLS on Day 2, the risk of nocturnal RLS attacks was still significantly increased, but the association was lower than that observed on Day 1 (OR: 3.00, 95% CI: 1.92–4.68, p < 0.0001). No significant association was observed between migraine and nocturnal RLS attacks on Day 3 (OR: 1.41, 95% CI: 0.90–2.20, p = 0.14).
We also analyzed the association between nocturnal RLS attacks and subsequent migraine attacks within three days, and the nocturnal RLS-migraine relationship on Day 1 was significant (OR: 1.97, 95% CI:1.30–2.98, p = 0.001). However, no such significance was observed for nocturnal RLS-migraine on Day 2 (OR: 1.09, 95% CI: 0.70–1.68, p = 0.71) or Day 3 (OR: 1.11, 95% CI: 0.70–1.74, p = 0.68) (Figure 1(b)).
According to the Spearman’s rank correlation analyses, the frequencies of migraine and RLS attacks in the recorded two weeks were positively correlated (rs = 0.56, p = 0.001). In addition, the severity of migraine attacks was correlated with the severity of RLS attacks on the same night (rs = 0.50, p < 0.0001). However, no such association was observed between nocturnal RLS and subsequent migraine attacks on the next day (rs = 0.23, p = 0.36). On the days of migraine attacks, the frequencies of RLS did not differ between patients who used abortive medications and those who did not (69.5% vs. 76.3%, p = 0.53).
Discussion
In the current study, our findings show that migraine attacks are likely to trigger nocturnal RLS attacks for two nights, particularly on the first night. Furthermore, the severity of migraine attacks was correlated with the severity of subsequent RLS. The triggering association declined as the time interval increased. However, the increased risk of subsequent migraine after RLS attacks was significant only on the first day and the association was weaker than that observed for migraine triggering RLS attacks.
Furthermore, a positive correlation was observed between migraine and RLS attack frequencies (rs = 0.56). This finding supports the results of a population-based study, which reported that RLS was a comorbidity in patients with active migraine but not past migraine (8). A similar RLS attack occurrence was observed among patients with chronic and episodic migraine in our previous study (7). Therefore, migraine chronicity may determine the frequency but not the presence or absence of RLS.
The exact mechanisms underlying migraine triggering subsequent RLS attacks are not clearly understood. Our results provide a clue that migraine attacks may induce certain changes in the brain that are associated with RLS attacks. RLS is associated with a relative decline in intracranial dopamine levels at night compared with an abnormal increase in the dopamine levels during the day (19). A positron-emission tomography (PET) study reported that serotonin synthesis increased during migraine attacks (13). It is also reported that serotonin and dopamine share an antagonistic relationship and exert opposite effects (20). Therefore, we postulate that the increased extracellular serotonin levels after a migraine attack may be absorbed into the dopamine terminals, thus hampering dopamine secretion (21). These actions of serotonin may partially account for the imbalance in dopamine levels and thus explain RLS occurrence following migraine attacks. Conversely, the mechanism of RLS triggering migraine also remains unknown. RLS-related sleep disturbance might be a possible cause of this relationship. The RLS epidemiology, symptoms, and treatment (REST) study reported that more than 40% of RLS patients presented with sleep-related symptoms (7,22). In migraine patients, sleep disturbance was one of the most common triggers of migraine attacks, particularly in the morning of the next day (23,24). Of note, we provide potential explanations for the bilateral triggering association between migraine and RLS attacks. However, we do not understand why the triggering association was stronger from migraine to RLS attacks, which deserves future research.
Our study has several treatment implications. First, prophylactic treatment for migraine may be effective in treating RLS attacks in migraine patients. This intervention may be similar to those in uremic patients with comorbid RLS, of which the symptoms improve after renal transplantation (25). Many migraine prophylactics (i.e. tricyclic antidepressants, dopamine antagonists, mirtazapine, selective serotonin reuptake inhibitors, and serotonin-norepinephrine reuptake inhibitors) might worsen idiopathic RLS and suggest prohibiting their use (2). However, if RLS attacks in migraine patients are mainly triggered by migraine attacks, these migraine-preventive medications, which can reduce migraine attacks, may have a chance to decrease rather than increase the RLS attacks as observed in idiopathic RLS patients. Nevertheless, these medications should be prescribed with caution. Second, in addition to other secondary causes, migraine history should be investigated in patients with a chief complaint of RLS. As indicated by the findings of our study, a patient diary may facilitate evaluating the presence of a triggering association. In addition, the data in our study indicate that acute abortive treatment did not alleviate RLS symptoms.
Our study has limitations. First, the patient diary was limited to a two-week period, and the observations were recorded by the patients at home. However, our study had a prospective design involving reminder calls, and data accuracy was confirmed when the patients recorded their attacks in the diaries. Second, this study aimed to investigate the relationship between migraine and RLS. Therefore, only patients with “episodic” but not “daily” attacks were recruited. This is our original study design; however, selection bias might have been introduced by excluding those with daily headache or daily RLS. Third, our patients reported a high frequency of headaches. This finding is not surprising because our patients were recruited from a headache clinic. Therefore, our findings may not be generalizable to other headache populations. Third, our findings were observed in migraine patients with comorbid RLS; therefore, they may not be extrapolated to RLS patients with comorbid migraine recruited from RLS clinics.
In conclusion, our study results demonstrated a bidirectional triggering association between migraine and RLS attacks in migraine patients with comorbid RLS; however, the association of migraine triggering RLS was stronger and lasted longer than that of RLS triggering migraine.
Clinical implications
A bidirectional triggering association existed between migraine and restless legs syndrome (RLS) attacks. The relationship was stronger and lasted longer for migraine triggering subsequent RLS than that for vice versa. In migraine patients comorbid with RLS, some migraine-preventive medications, i.e. tricyclic antidepressants, dopamine antagonists, selective serotonin reuptake inhibitors, mirtazapine, and serotonin-norepinephrine reuptake inhibitors, may decrease rather than increase the RLS attacks as observed in idiopathic RLS patients.
Footnotes
Acknowledgment
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of this manuscript.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
This study was supported in part by grants from the National Science Council of Taiwan (103-2321-B-010-017, 102-2321-B-010-030, 100-2314-B-010-018-MY3, and 99-2314-B-075-036-MY3), Taipei Veterans General Hospital (VGHUST102-G7-6-1, V102C-118, V102E9-001, V103E9-006, V103C-080, and VGHUST103-G7-4-1), NSC support for the Center for Dynamical Biomarkers and Translational Medicine, National Central University, Taiwan (NSC 101-2911-I-008-001, 102-2911-I-008-001).