Migraine is a complex, multifactorial disease characterized by a severe, pulsating, and often unilateral headache, lasting between 4 and 72 h, and the accompaniment of photophobia, phonophobia, and/or digestive symptoms [4]. Migraines are triggered by a stimulus, either internal or external, such as loss of sleep, weather changes, or stress. These triggers set off a cascade of cortical excitability, arterial dilation, and inflammation which trigger the pain impulses of the migraine [9]. Migraines may be present with or without aura, with diagnostic criteria displayed in Table 1 [3, 4, 10]. The most commonly supported hypothesis for the mechanism of aura is cortical spreading depression (CSD), though the origin of CSD is unclear [4].

Hormones may play a role in the pathogenesis of migraine, as evidenced by the relationship between migraine and the menstrual cycle. The luteal phase of the menstrual cycle is a common trigger for migraine [2, 11]. The characteristic steep estrogen drop in the luteal phase is thought to increase blood vessel permeability, leading to greater impact of pro-inflammatory mediators such as prostaglandins, which are elevated threefold during the same time period [11, 12]. Withdrawal from estrogen also contributes to migraine susceptibility by decreasing serotonergic tone and affecting central opioid tonus [11]. One study found that the rate of decline in estrogen levels following the luteal peak was significantly greater in patients with migraine compared to those without migraines [13]. By contrast, higher concentrations of estrogen have been associated with higher aura frequencies in patients who get migraines with aura [14]. Given the distinct interrelated nature of certain migraines with hormonal fluctuations, a subgroup of migraines are classified as menstrual migraines.

Menstrual migraine, or catamenial migraine, is defined according to criteria set forth by the International Classification of Headache Disorders (ICHD). Pure menstrual migraine (PMM) is defined as a migraine which occurs only on days 1 ± 2 (i.e., days − 2 to + 3) of menstruation in at least 2 of 3 cycles and does not occur at any other point in the cycle [15•]. PMM may occur with either the presence or absence of aura. By contrast, menstrual-related migraine (MRM) is defined as migraine which occurs on days 1 ± 2 of menstruation in at least 2 of 3 cycles as well as at other points in the cycle. Similar to PMM, MRM may occur with or without aura, though both types of menstrual migraine occur almost exclusively without aura [2].

PMM is relatively uncommon, occurring in only 10–20% of people who menstruate [2]. However, migraine at this stage in the cycle is very common. When considering all phases of the menstrual cycle, migraine without aura occurs with the highest incidence during days − 2 to + 3 of menstruation. Further, menstrual migraines are of longer duration, higher severity and disability, higher risk of relapse, and less responsive to treatment than non-menstrual attacks [2].

The risk of stroke with migraines is serious so that careful investigation of risk factors by clinicians is warranted [2, 11]. While there is extremely mixed data regarding patients with migraine without aura, there is a well-established increased risk of ischemic stroke among patients with migraine with aura, although the cause is poorly understood [6••, 11]. Interestingly, aura frequency also directly impacts a patient’s ischemic stroke risk. Compared to the individual without migraine headaches, the presence of < 1 migraine with aura per month increased ischemic stroke risk twofold, while > 1 migraine with aura per week conferred a fourfold increase in ischemic stroke risk [14].

One possibility is the migraine infarction, which is thought to be produced when the cortical spreading depression causing the aura produces such vasoconstriction that ischemia develops [16]. This can only be diagnosed if the symptoms of the stroke exactly mirror a patient’s historical aura symptoms [11].

Migraine with aura is also associated with many other ischemic stroke risk factors. Patent foramen ovale is found in approximately twice as many patients with migraine with aura as in the general population, which can pose increased ischemic stroke risk [16]. Migraine with aura is also associated with vascular risk factors such as increased rates of systemic lupus erythematosus and antiphospholipid syndrome [6••, 11, 16].

Estrogen-containing medications can increase risk of venous thromboembolic events, and the understanding of the mechanisms behind this phenomenon has advanced in recent years [17]. CHCs lead to an increase in various components of the procoagulant pathway, including fibrinogen, prothrombin, and factors VII, VIII, and X. Additionally, there is a decrease seen in hemostasis inhibitors antithrombin and tissue factor pathway inhibitor, which further increase coagulation risk [17]. These findings are seen regardless of route of administration, meaning CHCs in the form of oral contraceptives, contraceptive ring, and contraceptive patch confer the same risk [17].

The adverse effects of early CHCs became apparent shortly after the release of the first oral contraceptive. In response, prescribing guidelines were updated to advise against their use for patients with risk factors including histories of hypertension, stroke, deep vein thrombosis, and smoking [14]. A critical study published in 1975 demonstrated an increased stroke risk posed by COCs, with an incidence of stroke 4 to 5 times higher in users than nonusers. However, the study did not correlate the risk with the concentration of estrogen. Twenty-three of the 25 patients who suffered from a stroke while taking a mestranol-containing formulation were on a 100-µg dose, and all 20 of the 20 women who experienced a stroke on ethinyl estradiol were taking 50-µg pills. All formulations tested in the study were high doses compared to today’s standard of ethinyl estradiol dosing between 10 and 35 µg [10].

Other studies that showed the increased risk of stroke in CHC users rarely stratified risk by hormone dosing, and as the hormone levels in CHCs were reduced with new formulations over the years, decreases in adverse events have been consistently observed across numerous studies [13]. One pooled US study reviewed 3.6 million woman-years and found no increased stroke risk in users of low-dose COCs. While some European studies have continued to find a minimal increase in risk, this difference may be explained by an increased incidence of smoking and higher doses of estrogen used in those studies [14].

In a 2017 national-scale case–control study, Champaloux et al. identified first-ever strokes among women aged 15–49 between 2006 and 2012 based on healthcare claims data of a private insurance, ultimately reporting 25,887 ischemic strokes among 33,218,977 females. They sought to investigate the individual and combined associations between migraines, with and without aura, and contraceptive use, with ischemic stroke [6••]. In their study, migraine with aura had an increased OR of ischemic stroke of 2.9 (95% CI, 2.2–3.9). Migraine without aura had a smaller increased OR of ischemic stroke of 2.1 (95% CI, 1.8–2.5). Compared to never- or former-users of CHCs, current CHC-users had an increased OR of 1.3 (95% CI, 1.1–1.6). This data supports previous studies which have shown that migraines and CHC use are independently associated with increased ischemic stroke risk.

They also examined combined effects, creating a reference group who had no diagnosis of migraine and were not using CHCs. The highest increased risk of ischemic stroke was found as a six-fold increased risk in patients with migraine with aura using CHCs (OR 6.1, 95% CI, 3.1–12.1), followed by patients with migraine with aura not using CHCs (OR 2.7, 95% CI, 1.9–3.7). Migraine with and without aura was found to elevate risk when used with CHCs; however, the risk was not significantly different than that found in patients with migraine without aura not using CHCs. This study was notable for its large sample size and statistically significant findings which align well with many prior studies in their identification of the additive effects of migraine with aura and CHCs on ischemic stroke risk [14, 18]. However, their finding of increased risk associated with migraine without aura is less consistently reported and therefore needs further research.

Although studies of this magnitude are helpful in gaining confidence in clinical recommendations, there are still many limitations. Given the timeline of this study, the study population using CHCs were exposed to modern dosing of estrogens, though in previous studies and older reviews, the same assumption cannot be made as easily [10, 13, 14]. Even among lower dosed prescriptions, there are many different formulations of CHCs with a range of estrogen levels and different types of progestin which were not individually compared. Additionally, a limitation of this study and others like it is that in basing data points off diagnostic billing codes, patients are reduced to a binary of having or not having a condition, which does not represent the spectrum of a disease experience. For instance, given the evidence that aura frequency significantly impacts the magnitude of individual ischemic stroke risk, it would be beneficial to understand the frequency of migraines among the individuals who experienced a stroke in the study.

Another limitation is that migraine, both with and without aura, is a highly prevalent disease and is underdiagnosed [19]. This suggests that individuals may be excluded or incorrectly assigned as patients without migraines if they have not sought or have not had access to medical care for a formal migraine diagnosis. Also, patients coded as having migraines may also represent a more severe end of the migraine disease spectrum [6••]. Another consideration is that a prescription for a CHC noted in a patient chart does not equate to consistent use of that CHC. Finally, many studies are not able to control for other risk factors for stroke, such as smoking and hypertension.