Abstract
In order to evaluate a possible association between migraine and restless legs syndrome (RLS), we performed a case-control study on the comorbidity of RLS and migraine. Patients with migraine (n = 411) and 411 sex- and age-matched control subjects were included. Migraine was diagnosed according to International Headache Society criteria, RLS according to the criteria of the International Restless Legs Syndrome Study Group. Furthermore, all patients had to fill out a self-assessment test performance on depression [Beck's Depression Inventory (BDI)]. RLS frequency was significantly higher in migraine patients than in control subjects (17.3% vs. 5.6%, P < 0.001; odds ratio 3.5, confidence interval 2.2, 5.8). In our sample, there was no significant association between migraine and depression as defined by the BDI score (9.6% in migraine vs. 4.0% in control subjects, P = 0.190). Depression was, however, not significantly more frequent in migraine patients with RLS (13.6%) than in migraine patients without RLS (8.7%). In addition, migraine patients with RLS had a significantly higher BDI score. RLS features did not differ significantly between migraine patients with RLS and control subjects with RLS. There is an association between RLS and migraine and, in addition, a co-association with depression. The underlying mechanism, however, remains undetermined and might be related to a dysfunction of dopaminergic metabolism in migraine.
Introduction
Migraine has been associated with some sleep disorders and sleep disturbances such as prolonged rapid eye movement (REM) sleep latency and REM sleep time (1), decreased arousal index (2), somnambulism (3) and narcolepsy (4), although conflicting results have been obtained for the latter association (5). For movement disorders in sleep, preliminary data have been published suggesting an association between migraine and periodic or restless limb movements (6, 7). In addition, recurrent limb pain in children has also been associated with migraine (8).
Restless legs syndrome (RLS) is a sensorimotor disorder with a prevalence ranging between 2.5% and >10% in the White population (6), affecting women twice as often as men (9, 10). The prevalence increases with age until about 65 years and then decreases again (9, 11, 12). The diagnosis of RLS is based on the clinical criteria of the International Restless Legs Syndrome Study Group (13). These criteria comprise the urge to move because of uncomfortable and/or unpleasant sensations in the legs, relief of symptoms by movement, increase of symptoms in rest or inactivity, and increase of symptoms in the evening and at night.
The pathophysiology of RLS is still unknown. A dysfunction of the dopaminergic neurotransmitter system is very likely to play a major role in the occurrence of RLS (14, 15). However, there is also evidence that peripheral mechanisms such as C-fibre neuropathy and spinal lesions can contribute to RLS (16). Metabolic disorders such as iron deficiency, diabetes mellitus, renal insufficiency (in particular with dialysis) and polyneuropathy can cause a secondary or symptomatic form of RLS (11, 17). A positive family history is obvious in about 50% of the patients with idiopathic RLS. The hereditary type can be found in particular in families with early onset of symptoms. Three gene loci have been described to date (on chromosomes 12q, 9p and 14q) (18).
Several studies have described an involvement of dopamine in the pathophysiology of migraine attacks (19, 20). It is assumed that there is a relation between prodromal, accompanying and postdromal symptoms and the activation of the dopamine system (20). The headache phase may also be influenced by the dopaminergic system, either in the nociception mechanism or the regulation of cerebral blood flow. Recent pharmacological findings have supported this theory and the view that hypersensitivity of peripheral and central dopaminergic receptors is a specific migraine trait. Finally, a high density of lymphocytic D5 receptors has been found in migraineurs. Therefore, it might be that up-regulation of the dopaminergic system is important for certain subtypes of migraine.
Although no studies on the association between migraine and RLS are available, there are some common features that point to possible comorbidity or overlap. Several studies have described a correlation between psychiatric disorders and migraine or RLS. Patients with migraine have a threefold higher risk of developing depression than healthy controls, especially in the subgroup with migraine with aura; on the other hand, patients with depression have a threefold higher risk of developing migraine (21–23). In patients with RLS, a higher prevalence of anxiety and depressive symptoms has also been described (24, 25), suggesting some common mechanisms in both migraine and RLS. Furthermore, dopaminergic dysfunction has also been described in migraine (19, 20).
Based on the observation that patients with migraine in a supraregional headache out-patient clinic frequently reported symptoms of RLS, we designed a case–control study of the association of migraine and RLS. In particular, we focused on the additional comorbidity with depression.
Methods
Patients were recruited from the supraregional headache out-patient clinic at the Department of Neurology, University of Münster and from the ‘Dortmunder Gesundheitsstudie’ (for methods and results see http://www.dmkg.de/grundla/epidemiolstudie.pdf). Consecutive migraine patients with and without aura were enrolled according to the criteria of the International Headache Society (26). The migraine patients enrolled via the ‘Dortmunder Gesundheitsstudie’ were representative of the German population with respect to age, sex and socioeconomic status. The study was approved by the local Ethics Committee.
After giving informed consent, patients were interviewed based on a structured questionnaire containing the criteria for RLS. This set of questions was validated against a physician judgement in a prior study of the general population also in Germany (27, 28). To date, this is the only validated instrument to assess RLS in population studies. In those patients who answered ‘yes’ to all diagnostic questions for RLS, the clinical diagnosis of RLS was confirmed by experienced neurologists who were blinded to the other questionnaire results. In addition, the migraine diagnosis and migraine features were evaluated by a neurologist experienced in headache diagnosis. The patients were also asked to fill in a self-assessment questionnaire concerning depression [Beck's Depression Inventory (BDI)] (29). The BDI contains 21 items with a maximum score of 63. A diagnosis of depression was made with a score of ≥18.
Patients with the diagnosis of RLS were asked to fill in the RLS severity scale based on the International Restless Legs Syndrome Study Group (13, 30). The RLS severity scale evaluates the severity of RLS complaints by scoring the predominant features with up to 40 points (0, no RLS; 1–10, slight symptoms of RLS; 11–20, moderate symptoms of RLS; 21–30, strong signs of RLS; 31–40, very strong signs of RLS).
Control subjects were matched by age (±0.5 years) and sex in a one-to-one fashion. Only control subjects without any history of headache suggestive of migraine were enrolled. They were recruited from waiting relatives of patients in the trauma out-patient clinic and from the ‘Dortmunder Gesundheitsstudie’. We evaluated the diagnosis of RLS, the BDI and the RLS severity scale as described for the patients.
The lifetime prevalence of RLS was calculated in our samples. The clinical and demographic data were registered, including migraine frequency per month, duration of migraine attacks, lifetime duration of migraine, number of first-degree relatives with and without migraine, number of days per month with acute migraine medication, BDI score, diagnosis of depression based on the BDI data, and the RLS severity scale score.
Data are presented as percentage or as arithmetic mean with SD. Statistical analysis was performed by non-parametric testing. For quantitative data, the Mann–Whitney U-test was used, for qualitative data the χ2 test. In addition, the odds ratio (OR) with 95% confidence interval (CI) was calculated. Significance level was set at P = 0.05.
Results
A total of 411 migraine patients (154 male, 257 female) and 411 control subjects were enrolled. Demographic and clinical data are presented in Table 1. The lifetime prevalence of RLS was 17.3% in patients with migraine and 5.6% in control subjects (OR 3.5, 95% CI 2.2, 5.8; P < 0.001). The score in the RLS severity scale was higher in migraine patients, but just failed to reach statistical significance because of a low number of responders in the control group (15.6 ± 6.6 vs. 10.0 ± 2.0; P = 0.072). The prevalence of depression as diagnosed by the BDI score was higher in the migraine patients but also failed to reach statistical significance (9.6% vs. 4.0%; P = 0.121). However, there was a significant difference in the BDI score between the patient and control groups (8.8 ± 7.6 vs. 6.1 ± 5.5; P < 0.001). Migraine patients had relatives with migraine significantly more often (41–22%; P < 0.001) and a significantly higher ratio of first-degree relatives with migraine than subjects without migraine.
Demographic and clinical data of the migraine patients and control subjects presented as arithmetic mean and SD or as percentage
BDI, Beck's Depression Inventory.
Table 2 presents clinical data of the migraine patients. Compared with other epidemiological studies, no obvious abnormalities occurred in the clinical features.
Clinical data of the migraine patients presented as arithmetic mean and SD or as percentage
Table 3 shows the differences between migraine patients with RLS and those without RLS. Migraine patients with RLS were significantly older, had a significantly longer duration of migraine and had a significantly higher BDI score. No other comparison revealed any significant differences.
Comparison of migraine patients with RLS and migraine patients without RLS with respect to migraine features, presented as arithmetic mean and SD or as percentage
BDI, Beck's Depression Inventory.
Discussion
The most important finding of this study is that migraine is significantly associated with RLS. In addition, patients with migraine showed a trend towards worse symptoms of RLS than patients who suffer only from RLS. The RLS prevalence of our control group lies in the lower range of previous population-based epidemiological studies on the prevalence of RLS; the prevalence of migraine patients is clearly higher. In a previous smaller case series (n = 50), the prevalence of RLS in migraine patients was 34% (7); however, these patients were severly affected by migraine and received antidopaminergic drugs.
Migraine patients with RLS are significantly older than patients who have only migraine. One reason for this could be that the prevalence of RLS increases with age, whereas migraine mainly appears at 20–50 years old. Age and the duration of migraine, which are also significantly greater in migraine patients with RLS, are of special importance because patients with migraine and RLS have probably taken headache medication for a longer time. These patients may more often exhibit a symptomatic RLS based on renal insufficiency due to medication overuse. Older migraine patients could also have taken ergotamine derivatives as a headache medication, because ergotamine was the standard medication in the 1970s and 1980s, whereas triptans have been available only since the early 1990s. Many younger migraine patients might never have taken ergotamine. Ergotamine can cause myalgia and paraesthesia as a side-effect of clinical or subclinical ergotism (31). In these cases, the symptoms of the vascular side-effects could mimick RLS or ergotamine use, with subsequent affects on the spinal tract, and polyneuropathy (32) might lead to symptomatic RLS.
Two smaller observational studies have also found an increased rate of RLS in migraine patients (6, 7). Other extrapyramidal movement disorders with involvement of the dopaminergic system have also been linked to migraine, such as Tourette's syndrome (33), Parkinson's disease (34), dystonia (35) and essential tremor (36). The underlying mechanisms of all these comorbidities are still unexplained.
From a pathophysiological point of view, we also have no explanation for the correlation between migraine and RLS, although this was not the aim of the study. On the one hand, RLS seems to be based on a dysfunction of the dopaminergic system (14, 15), and the standard therapy for primary RLS is medication with dopamine or dopamine agonists (37). On the other hand, there is involvement of the dopaminergic system especially as a hypersensitivity of peripheral and central dopaminergic receptors in migraine patients (38). Based on the well-known interactions between central amines in migraine, reduced release of serotonin between attacks could lower dopamine release, which would lead to receptor hypersensitivity. Hypersensitivity to extrastriatal dopamine receptors has also been shown in patients with both Parkinson's disease and migraine (38). In conclusion, the connection between RLS and migraine according to the dopaminergic system might explain the correlation. Conversely, dopamine receptor antagonists are effective therapeutic agents in migraine prophylaxis. Genetic data suggest that molecular variations within dopamine receptor genes play a modifying role in the pathophysiology of migraine (39). This might also be the case in RLS. Finally, modulation of dopaminergic neurotransmission has been suggested for the therapeutic management of migraine (40).
Our results on the prevalence of depression are in contrast to those previously reported (41). In our study, depression was not significantly more frequent in migraine patients than in the control group, whereas several other authors have stated that migraine patients suffer from depression more often than controls. This is difficult to explain. The most plausible explanation is that we made the diagnosis of depression based exclusively on the BDI score (only a self-assessment test), which is clearly not comparable to a diagnosis according to the International Classification of Diseases or Diagnostic and Statistical Manual of Mental Disorders criteria. In future studies, an external assessment test might give more reliable results for the impact of depression on the comorbidity of migraine and RLS. Furthermore, relatives of trauma patients might show a higher depression score than other parts of the healthy population. Nevertheless, we found a significant difference in BDI scores between the migraine and control groups, which is in concordance with the several studies reporting depressive symptoms in migraine patients (42).
The type of migraine (occurrence of aura or not) seems to have no influence on the association with RLS, but the group of migraine patients with aura was too small to draw any further conclusions. Furthermore, the migraine features as recorded in our study did not differ between migraine patients with and without RLS.
This study has some limitations. First, we did not screen for symptomatic RLS (diabetes, renal insufficiency, specific medication), nor did we subclassify the headache medication (ergotamine, triptans, non-steroidal anti-rheumatic), which could be a major factor explaining the occurrence of RLS in migraine patients. Second, we did not evaluate specific symptoms of the migraine attack, the intensity of the pain, or other symptoms accompanying the migraine attack. It may be a correlation between migraine intensity and semiology and the occurrence of RLS. Third, we did not enquire about familial RLS and familial depressive disorders. Detection of a familial pattern of comorbidity should be included in future studies. This could help to identify a common genetic background of both diseases. Even the familiy history of migraine in our patients was not reliable, because relatives of our patients were not interviewed directly (43). Finally, it would be interesting to have data on the appearance of RLS symptoms and of migraine attacks at the same time. Several dopaminergic-associated symptoms such as yawning, drowsiness, and food craving can occur as premonitory or concomitant migraine symptoms.
In conclusion, our case–control data suggest a comorbidity of migraine and RLS independent of the migraine features and RLS features. This comfirms previous observational studies (6–8) and, thus, provides a further step to a population-based study. A longer duration of migraine might be a risk factor for RLS, but this needs confirmation in future studies; the underlying mechanism could be a secondary RLS due to longer medication intake. A co-association of depressive symptoms for both disorders was also found, which could also explain part of the comorbidity.