Abstract
The pathophysiology of spontaneous cervical artery dissection (sCAD) is largely unknown. An association with migraine has been suggested, but not definitively proven. In the setting of a hospital-based prospective case-control study we assessed personal and family history of migraine in 72 patients with sCAD, 72 patients with cerebral infarct unrelated to a CAD (non-CAD) and 72 control subjects. Personal history of migraine was significantly associated to sCAD compared to non-CAD (59.7% vs. 30.6%; OR 3.14; 95% CI 1.41-7.01) and controls (18.1%; OR 7.41; 95% CI 3.11-17.64). As opposed to migraine with aura, migraine without aura was significantly more frequent among sCAD than among non-CAD (56.9% vs. 25.0%; OR 3.91; 95% CI 1.71-8.90) and controls (12.5%; OR 9.84; 95% CI 3.85-25.16). Similar results were observed when the frequencies of family history of migraine were compared. These findings are consistent with the hypothesis that migraine may represent a predisposing condition for sCAD.
Introduction
Migraine patients have repeatedly been reported to be at increased risk for the development of ischaemic stroke at young age (1–3). However, the possibility that the magnitude of such an association may vary according to the subtype of ischaemic stroke has been scarsly investigated. Since the pathogenic mechanisms of stroke among young adults are more heterogeneous than in the elderly, the implication of a subtype-specific relationship between migraine and ischaemic stroke would be noteworthy as it could suggest specific pathophysiological processes.
Cervical artery dissection (CAD) is one of the most frequent aetiology of stroke in young adults as it accounts for up to one fifth of cerebral ischaemic events in patients under the age of 50 (4). Despite the increasing clinical awareness and the development of noninvasive investigational tools, the pathogenesis of spontaneous cervical artery dissection (sCAD) is still largely unexplained. Although the possibility that migraine might be a causative risk factor for sCAD have been put forward by d’Anglejean-Chatillon et al. (5) about 15 years ago, it still remains to be proven. Thus far, scarse epidemiological analyses have explored this hypothesis in the setting of case-control studies or case series, with conflicting results (5–8). More recently, indirect evidence (9) and the results of a prospective study reporting a significant association between these two conditions (10) reinforced the assumption of a potential relationship between migraine and sCAD.
The present survey was conducted to test this hypothesis in the setting of a hospital-based case-control study including a group of patients with sCAD, a group of patients with cerebral infarct unrelated to a CAD (non-CAD) and a group of controls.
Subjects and methods
Patients with sCAD consecutively admitted to the Department of Neurology of the University Hospital of Brescia and to the Department of Neurosciences of the University Hospital of Parma between January 1999 and January 2002 were prospectively enrolled. All the subjects in whom clinical and/or duplex ultrasound findings were consistent with the hypothesis of acute CAD, underwent neuroradiological investigations in order to confirm the diagnosis. Four-vessel conventional angiography and/or magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) of the brain and neck were included in the diagnostic work-up. The presence of the double lumen sign (a false lumen or an intimal flap), luminal narrowing with the ‘string sign’, and gradual tapering ending in total occlusion of the lumen (flame-like occlusion) were considered reliable angiographic findings of CAD, while a narrowed lumen surrounded by a semilunar-shaped intramural haematoma on axial T1-weighted images was considered the pathognomonic MR sign. Dissections were classified as spontaneous when occurring spontaneously or in association with a minor trauma (8).
Age and sex-matched patients with ischaemic stroke not related to a CAD (non-CAD) consecutively admitted during the same time period were also included in the study. A detailed description of the standard diagnostic work-up and the adopted criteria for aetiological stroke subtype classification have been presented previously (11). Controls were age and sex-matched individuals with no known history of vascular disease, recruited from the staff members of our hospitals.
Demographic data and conventional vascular risk factors such as history of arterial hypertension, diabetes mellitus, cigarette smoking, and hypercholesterolemia were assessed in all subjects (11). Body mass index (BMI) was calculated as weight divided by height squared (kg/m2) as a measure of relative weight. Obesity was considered to be present when BMI was > 30. Oral contraceptive use was also considered for women. The study was designed and carried out in observance of the ethical principles established by the local Institutional Guidelines on Clinical Investigation. Written, informed consent was provided by all the study participants.
Diagnosis of migraine
Personal and family history of headache was assessed by three investigators (CB, PI, EB) unaware of the study hypothesis, during a face-to-face interview based on a semistructured questionnaire (12). The diagnosis of migraine without aura, migraine with aura and migrainous disorder (group 1.7 of the IHS classification) was made according to the diagnostic criteria of the International Headache Society (13). Since all the patients with migrainous disorder in our series (n = 5) fulfilled all the criteria but one for migraine without aura, they were included in the diagnostic category of migraine without aura. Family history of migraine was considered positive when at least one first-degree relative (i.e biological parent or sibling) suffering from migraine was present.
Statistical analysis
Differences among groups were examined with the χ2 test and with Anova F-test, when appropriate. P-values were computed by Montecarlo (MC) procedure. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by Multinomial Logistic Regression Models with sCAD, nonCAD, and control groups as outcome variables, and with migraine variables (personal history of migraine, migraine with aura, migraine without aura, family history of migraine, age of migraine onset, and frequency of migraine attacks) as repressor, controlling for centre (Brescia, Parma), age (in 6 age strata), gender, arterial hypertension, smoking status, and obesity. Analyses were conducted with the SPSS (version 11.1) software package.
Results
The study group was composed of 72 patients with sCAD, 72 patients with non-CAD ischaemic stroke and 72 controls. Among patients with sCAD, 62 had a single vessel-dissection, involving the carotid artery in 41 (56.9%) cases and the vertebral artery in 21 (29.2%) cases. Multiple vessel dissections were observed in 10 (13.9%) patients. Fifty-eight (80.6%) patients had a transient ischaemic attack or a complete stroke, whereas 24 (33.3%) patients suffered from an oculosympathetic paresis. Sixty (83.3%) patients experienced headache at the time of arterial dissection.
Demographic features and history of vascular risk factors in each group are summarized in Table 1. History of hypertension was significantly more prevalent in the group of patients with sCAD than in the group of controls (23.6% vs. 9.7%; P < 0.05). Though patients with non-CAD ischaemic stroke showed a higher prevalence of hypertension and diabetes in comparison with sCAD patients, only smoking tended to be significantly more prevalent (48.6% vs. 30.6%).
Baseline characteristics of the study group
In women ≤ 45 years of age.
A history of migraine was diagnosed more frequently in the group of patients with sCAD (43 migraine sufferers; 59.7%) compared to the group of patients with non-CAD ischaemic stroke (22 migraine sufferers; 30.6%) and the group of controls (13 migraine sufferers; 18.1%). These differences turned out to be significant when patients with sCAD were compared to patients with non-CAD ischaemic stroke (OR 3.14; 95% CI 1.41–7.01) and to controls (OR 7.41; 95% CI 3.11–17.64) in the multinomial logistic regression analysis. In contrast, borderline difference was found between the group of patients with non-CAD ischaemic stroke and the group of controls (OR 2.35; 95% CI 0.98–5.65) (Table 2).
Prevalence distributions and Odds Ratios of the pairwise groups comparison of the migraine variables adjusted for strata (centre, age, gender) and covariates (smoking status, hypertension and obesity) by Multinomial Logistic Regression modelling
Thirty-three patients (76.7%) had migraine without aura, 2 patients (4.7%) had migraine with aura (MA), and 4 patients (9.3%) had a migrainous disorder. Four patients (9.3%) had attacks of migraine both with and without aura.
Including subjects with migraine without aura and subjects with both migraine with and without aura.
Including subjects with migraine with aura and subjects with both migraine with and without aura.
When specific subtypes of migraine were considered, a significant association was observed between migraine without aura and the group of patients with sCAD in comparison with both the group of patients with non-CAD ischaemic stroke (56.9% vs. 25%; OR 3.91; 95% CI 1.71–8.90) and the group of controls (12.5%; OR 9.84; 95% CI 3.85–25.16). Conversely, borderline difference was detected between patients with non-CAD ischaemic stroke and controls (OR 2.51; 95% CI 0.95–6.62). As opposed to migraine without aura, migraine with aura did not show a significantly higher prevalence in the group of patients with sCAD compared to the other groups (Table 2).
The group of patients with sCAD was also significantly associated to a positive family history of migraine with respect to the group of patients with non-CAD ischaemic stroke (43.6% vs. 25.4%; OR 3.23; 95% CI 1.36–7.64) and controls (12.5%; OR 8.69; 95% CI 3.22–23.46), while non-CAD-control comparison proved a borderline difference (OR 2.69; 95% CI 0.99–7.28). Age of migraine onset and frequency of migraine attacks did not differ significantly between groups (data not shown) (Table 2).
Finally, patients with multiple vessel sCAD showed a stronger association with migraine than patients with single vessel sCAD (Table 3).
Prevalence distributions and Odds Ratios of the migraine variables within the group of patients with sCAD (multiple-vessel dissection vs. single-vessel dissection), adjusted for strata (centre, age, gender) and covariates (smoking status, hypertension, and obesity) by Binary Logistic Regression modelling
Values are number and (%). mvCAD, multiple-vessel cervical artery dissection; svCAD, single-vessel cervical artery dissection.
Discussion
The main result of the present study is that a personal history of migraine is associated with the occurrence of sCAD as compared to ischaemic stroke of other aetiology. This association seems to be stronger and more significant in patients with dissections involving multiple vessels. In contrast, we did not detect an association between sCAD and migraine with aura.
Taken together, these findings support the hypothesis that migraine and arterial dissection may share common pathogenic mechanisms and underlying susceptibility factors. Given its character of association study, our analysis cannot provide direct informations on the precise mechanisms by which migraine may affect the risk of sCAD. However, recent reports provide indirect biological plausibility to this assumption. Current theories suggest the concept that migraine is a neurovascular disorder (14). Since vascular disturbances may act as a trigger for neurological manifestations, it is possible that molecular mechanisms that affect vascular function may be involved in migraine susceptibility. In this regard, it is generally assumed that an underlying arteriopathy related to extracellular matrix defects represents a predisposing condition for sCAD (15). Because of an increased activity of serum elastase, a metallopeptidase which degrades specific elastin-type aminoacid sequences, patients with migraine might be at increased risk of extracellular matrix degradation (9) and sCAD occurrence. Sparse observations indicating that serum levels of alpha1-antitrypsin, a serine proteinase inhibiting the proteolytic effect of elastase and collagenase, is reduced in patients with sCAD are in line with this hypothesis (16–18). It is also tempting to speculate that these extracellular matrix abnormalities might be involved in the pathogenic processes linking both diseases to patent foramen ovale, a developmental defect predisposing to migraine (19, 20) and for which a relationship with sCAD has been suspected (21).
In line with previous observations of altered common carotid artery distensibility in patients with sCAD (22), Lucas et al. (23) recently reported that the endothelium-dependent vasodilation assessed in the brachial artery is significantly impaired in these subjects. Interestingly, similar vascular changes have been observed in migraine patients during interictal periods (24). This supports the hypothesis that a common generalized vascular disorder, which extends beyond the cervical-cranial circulation, might be a predisposing condition for both diseases.
Finally, the analysis of small families has shown that the structural abnormalities related to sCAD might be familial and follow an autosomal dominant pattern of inheritance (25, 26). This implicates that genetically determined alterations of the extracellular matrix may play a crucial pathogenic role and that candidate genes involved in the regulation of the endothelial and the vessel wall functions, might increase susceptibility to both conditions (27–29). As neither the age of migraine onset nor the frequency of migraine attacks turned out to be significant predictors of sCAD occurrence in our series, we can speculate that the link between migraine and arterial dissection may be unrelated to migraine activity and likely reflect shared susceptibility factors.
The literature on the relationship between migraine and specific aetiological subtypes of ischaemic stroke is sparse. In particular, several retrospective studies provided contradictory conclusions regarding the association between migraine and sCAD (5–8). Although such variability might theoretically be ascribed to differences in the studied populations, methodological shortcomings inherent in the retrospective design of these analyses cannot be ruled out a priori. Despite this may represent a likely explanation for such inconsistencies, the only two prospective studies reported to date also gave conflicting results. In particular, in a systematic analysis of the Lausanne Stroke Registry, Milhaud et al. found a lower frequency of arterial dissection among young stroke patients with active migraine than among those without (30). More recently, Tzourio et al. (10) observed a specific association of migraine and sCAD in a case-control analysis including consecutive patients with acute cerebrovascular disorders. In agreement with the results of the latter study, in our analysis we found that only migraine without aura was significantly more prevalent among patients with sCAD and that patients with multiple-vessel dissections had the highest migraine prevalence. In contrast to Tzourio et al. we found neither a higher frequency of migraine attacks among sCAD patients nor differences in the age of migraine onset.
Several limitations of our study should be noted. First, the interviewers were unaware of the study hypothesis but not unblinded to the status (case or control) of the participants. Therefore, they might have involuntarily diagnosed migraine more frequently in cases (sCAD and non-CAD) than in controls. However, since the frequency of migraine is at the upper end of the range found in the literature both in the group of patients with non-CAD ischaemic stroke and in the group of controls (30.6% and 18.1%, respectively), the possibility of an interviewer bias seems unlikely.
Second, cases could be more prone to remember their headaches than controls. Although such a recall bias cannot be completely ruled out in case-control comparisons, this is unlikely to operate when patients with sCAD and non-CAD ischaemic stroke are compared, as both groups were recruited for a cerebral acute vascular event. The prospective and consecutive recruitment of cases also excludes the possibility of a selection bias. Third, obtaining migraine status of first-degree relatives through proband cannot exclude a possible family information bias (31) and an underestimation of migraine prevalence among family members. However, given the fact that the interviewers were unaware of the study hypothesis, we believe that the effect of such a potential underestimation is equally operant in the three subgroups and, thus, it is unlikely to have a substantial influence on the comparison. Finally, given the relatively small size of our sample, the results of subgroup analyses should be interpreted with caution, and verified in larger studies.
Despite these limitations, our findings provide compelling arguments that a history of migraine may represent a predisposing condition to sCAD. Although further studies are needed to substantiate the hypothesis that a common vascular disorder might increase the propensity to both conditions, the finding of such coaggregation has potential implications in understanding the complex interaction between migraine and ischaemic stroke.
Footnotes
Acknowledgements
We gratefully acknowledge Dr Caspar Grond-Ginsbach, Department of Neurology, University of Heidelberg, and Dr Giorgio Dalla Volta, Headache Centre, Istituto Clinico Città di Brescia, for comments and suggestions on the manuscript.