Clinical manifestations

Presentation and course

Changes in sex-hormone levels during menarche, menstruation, pregnancy, and menopause may affect the prevalence and character of migraine and other types of headaches (08). There is no gender difference in the incidence of migraine before puberty. In adults, however, the prevalence of migraine is in the United States adult has been remarkably stable over the past 19 years, affecting roughly 1 out of every 6 American and 1 in 5 women over a 3 times-month period, 9.7% of males and 20.7% of females (24). The frequency, duration, and disability of attacks tend to be higher in women (02). There is no consensus on the likelihood that female sex is a risk of factor for transition from episodic to chronic migraine. Some studies have reported that the transition is more likely in women (OR 2.9, 95% CI 1.2–6.9) (149), whereas others show a nearly indistinguishable transition rate within 1 year (5.4% in men and 4.4% in women) (104). Multiple studies across different countries have demonstrated the same findings that heritability of migraine was essentially equal between men and women, with a with a broad sense heritability of 0.45 (95% CI 0.40–0.50) (69; 56; 54). These findings suggest the prominent sex difference in migraine prevalence cannot be entirely explained by genetic factors or heritability. Furthermore, there is a significant increase in migraine risk for females with a male, co-twin as compared to females with a female co-twin (OR 1.51, 95% CI 1.26–1.81), suggesting that masculinization of the prenatal environment may actually increase migraine risk for females (54). Headache and migraine prevalence are increased in female adolescents and adult women with early menarche, which seems to increase headache susceptibility (01). These data further suggest that women’s increased migraine risk results from exposure to endogenous or exogenous factors rather than from a genetic predisposition.

The diagnostic criteria for menstrual migraine have been included in the appendix, a section reserved for proposed diagnoses pending further scientific evidence rather than formal diagnoses, of the International Classification of Headache Disorders (ICHD) since its second edition (ICHD-2). In the International Classification of Headache Disorders- 3 (ICHD-3), published in 2018, pure menstrual migraine is defined as migraine attacks that occur exclusively between days -2 and +3 of the menstrual cycle (where day 1 is defined as the first day of menstruation) in at least 2 of 3 menstrual cycle (66; 67). Menstrually-related migraine is defined as migraine attacks that occur between days -2 and +3 of the menstrual cycle in at least 2 of 3 menstrual cycles and additionally at other times of the cycle. Both pure menstrual migraine as well as menstrually-related migraine can be with or without aura (see Table 1).

Table 1. International Headache Society Definitions of Pure Menstrual Migraine and of Menstrually-Related Migraine

Prevalence

Many women report that their migraine headaches are more frequent or severe during their menses than at other times. However, increased probability of headache during menses does not always mean a patient has menstrual-migraine. A potential point of confusion is that headache in the perimenstrual period, which might occur by chance in all women affected by migraines, can be mistaken for menstrual migraine, in which attacks occur reliably around menstruation. There is a large variation in prevalence of menstrually-related migraine among studies, which is likely explained by changing diagnostic criteria among different ICHD versions, study design differences such as headache calendars versus questionnaires, retrospective versus prospective, and general population versus headache clinic populations studied.

A study of 237 women based on the ICHD3-beta appendix criteria revealed prevalence among all women was as follows: any type of menstrual migraine, 7.6%; menstrual migraine without aura, 6.1%; menstrual migraine with aura, 0.6%; probable menstrual migraine without aura, 0.6%; and probable menstrual migraine with aura, 0.3% (150). The corresponding figures among patients with migraine: any type of menstrual migraine, 22.0%; menstrual migraine without aura, 17.6%; menstrual migraine with aura, 1.7%; probable menstrual migraine without aura, 1.6%; and probable menstrual migraine with aura, 1.0% (150). A 2015 study by Pavlovic and colleagues evaluated 1697 subjects who met modified ICHD-2 criteria for migraine and found prevalence of pure menstrual migraine to be 5.5%, menstrually-associated migraine at 53.8%, and menstrually-unassociated migraine to be 40.7% (111). Women with pure menstrual migraine had an older age of migraine onset and had fewer headache days but appeared to be more impaired by attacks based on HIT-6 and MIDAS scores. When headache diaries were used, only 7% of women people with migraine were found to have menstrual migraine (85; 87). In a population-based study, the rate of menstrual migraine was significantly lower (0.85%) than that of menstrually-related, which was 8% (36). However, most studies as well as the ICHD-3 agree that the most predominant subtype of menstrual migraine is migraine without aura. In comparison, migraine that occurs for the first time during pregnancy (an uncommon occurrence) or is induced by oral contraceptive intake (high estrogen levels) is often migraine with aura.

Headaches associated with menses may occur alone or as part of premenstrual syndrome (late luteal phase dysphoric disorder). Women with premenstrual syndrome are more susceptible to severe nonmigraine headaches both during the premenstrual phase and throughout the entire menstrual cycle (74). In a Swedish study of 266 menstruating women, women who had premenstrual syndrome were found to have a significantly increased risk for migraine without aura (97).

Higher levels of estrogen are associated with increased frequency of migraine aura. Conversely, in low estrogen states, as occurs in menstrually-related migraine, the onset of menses occurs as a function of a drop in estrogen. Thus, menstrually-related migraine are predominantly migraine without aura. Another general consensus is that steady or rising levels of estrogen are not associated with migraine; however, a rapid decline in estrogen levels, for example, those that occur before menstruation, perimenopausally, etc., can precipitate individual migraine attacks, as may occur with menstrually-related migraines, or worsening intensity or frequency of attacks over a longer period. However, this is a general guideline with many exceptions. For example, some women experience new headaches or worsening of their migraine attacks during the first trimester of pregnancy, although estrogen levels are generally rising.

An important distinction to be made is whether menstrually-related migraine occurs in setting of exogenous hormone use versus as a result of an unimpeded hypothalamic-pituitary-ovarian (HPO) axis. Sixty four point nine percent of the 72.2 million women aged 15 to 49 in the United States use a form of contraception (37). The most common contraceptive methods currently used were female sterilization (18.6%), oral contraceptive pill (12.6%), long-acting reversible contraceptives (LARCs) (10.3%), and male condom (8.7%) (37). The reported incidence of headache with hormonal contraception use is variable and related to the agent’s composition and the estrogen dose (96). Migraine is more likely to worsen or remain the same, rather than improve, following oral contraception initiation; however, the effects that an individual woman may experience are largely unpredictable (11). In a study of women who were taking high-dose oral contraceptives (50 or greater µg of ethinyl estradiol), 10% of the women who had not had headaches before oral contraceptive use reported headaches when they started taking oral contraceptives (80). In a double-blind study of 704 women, the use of low-dose combined oral contraceptives (20 µg ethinyl estradiol/100 µg levonorgestrel) was not associated with a higher headache incidence than placebo (35). The use of progestin implantable contraceptives is associated with a higher headache rate (10% to 30%) compared with combined oral contraceptives (20). In a literature review, Loder and colleagues concluded that in most contraceptive trials there was no statistically significant difference in overall headache complaints between treatment and control subjects (82). There is also evidence that migraine with aura is more likely to be exacerbated with oral contraceptive pill intake compared with migraine without aura (93; 94).

Pregnancy can change the pattern of established migraine. Migraine improves during pregnancy in 50% to 90% of women, mostly in the second and third trimesters (88; Silberstein 1997a; 122; 16). The course of migraine during pregnancy was assessed in a prospective Italian study of 49 pregnant patients with migraine (122). Of the 47 women who had migraine without aura, migraine improved in 46.8%, 83%, and 87.2% in the first, second and third trimester, respectively. Improvement during pregnancy was more likely when the migraine was menstrually related (98). Migraine attacks may even cease during pregnancy, especially in the second and third trimesters. The improvement in migraine that most women experience during pregnancy has been attributed to the rising or sustained high levels of estrogen that occur at that time. Only 4% to 8% of women had migraine worsen during pregnancy. Worsening during pregnancy is more common in women who have migraine with aura. Breast-feeding seems to exert a prophylactic effect in the postpartum period (152), whereas bottle-feeding is associated with an accelerated return of migraines (122). It has been posited that oxytocin may play a role in the protective effect of breast-feeding against migraine attacks.

Migraine may increase in frequency and severity in the premenopausal and perimenopausal periods. In an observational study of 100 postmenopausal women, headache (38%) was the most frequent pain complaint (99). This is probably due to erratic estrogen secretion at these times (52). After menopause, when sex hormone levels stabilize, most women experience improvement (107). The prevalence of new onset migraine during menopause may vary between 8% and 13% (60; 77). Some evidence suggests that women with premenstrual syndrome and migraine without aura might be more prone to headache exacerbation in the perimenopausal period (154). The use of oral hormonal replacement therapy may exacerbate migraine (102), likely due to unpredictable absorption and plasma levels, but there is evidence to the contrary as well (62). However, in a multivariate analysis of 18,221 postmenopausal women treated with hormone therapy, an intermediate dose of estrogen of 0.625 mg/d was associated with reduced headache frequency compared with higher or lower doses (100). In contrast to oral therapy, transdermal hormonal replacement therapy does not seem to exacerbate migraine and can even have preventive effects (93; 94), again possibly due to more stable absorption and less fluctuation in hormonal plasma levels.

In contrast to migraine, cluster headache is rarely related to menses, but this relationship has been described. A report described a woman with episodic cluster headache whose attacks were temporally linked to her menstrual cycle and occurred regularly on days 4 to 9 of her menstrual cycle. Suppressing menstruation with oral contraceptives resulted in cessation of her cluster headache attacks. However, a study of 224 women with cluster headache showed that menstruation, pregnancy, menopause, and the use of oral contraceptives had little effect on their headache attacks (148).

Tension-type headache is less likely to be exacerbated during menstruation as well (72). In a large, prospective, questionnaire-based study, patients with migraine were 3.12 times more likely to experience a headache attack during menstruation compared to patients with tension-type headache. However, pure menstrual headaches were experienced by 4.4% of menstruating migraine patients and 8.2% of people experiencing menstruating tension-type headache (p = 0.032). In the same study, the odds ratio for disappearance of headache in migraineurs during pregnancy compared to tension-type headache patients was 2.56 (72). Oral contraceptives worsened migraine significantly more than tension-type headache, whereas menopause had a slight improving effect on migraine compared to tension-type headache (p = 0.046).

Chronic paroxysmal hemicrania (a cluster-like headache characterized by daily, short duration, high-frequency headaches responsive to indomethacin) is more common in women than men. Pregnancy seems to have an ameliorating effect on attack frequency and severity. Headache onset sometimes occurs immediately after delivery (137).

Prognosis and complications

The relationship between estrogen containing contraceptives, stroke, and migraine has been a widely investigated and debated subject, dating almost to the initial launch of continuous hormonal contraceptives in the 1960s. However, it is important to note that the first generations of continual hormonal contraceptives used in trials dating back to the 1970s and 1980s generally used products that were almost uniformly high-dose estrogen containing pills, defined as 50 µg or more of ethinyl estradiol. In subsequent decades, the dose of estrogen was successively reduced, so much so that some pills on the market today contain only 10 µg of ethinyl estradiol.

The relationships among oral contraceptives, migraine, and stroke are complex. American College of Obstetrics and Gynecology in 2010 noted that extended-cycle or continuous hormonal contraceptives, including oral and parenteral products, might provide relief of migraines by eliminating the drops in estrogen levels that precipitate them (07). Conversely, a 2016 update from the U.S. Centers for Disease Control and Prevention on hormonal contraceptive use in various conditions, deemed combined hormonal contraceptives safe in those with migraine without aura (28). However, in the case of migraine with aura, the consensus was that the risk of stroke outweighs benefit of continuous hormonal contraceptives in patients with migraine with aura. Current guidelines restrict the use of combined hormonal contraceptives in the setting of migraine with aura; migraine without aura is not a contraindication to oral contraceptive use.

Several studies suggest migraine alone is a risk factor for ischemic stroke in young women, with odds ratio ranging from 1.9 to 4.3 (146; 11; 31; 42; 126). In a case control study, ischemic stroke was strongly associated with migraine, both migraine without aura (OR 3.0, 95% CI: 1.5 to 5.8), and migraine with aura (OR 6.2, 95% CI: 2.1 to 18.0). The risk of ischemic stroke was substantially increased for migrainous women that were using oral contraceptives (OR 13.9) or who were heavy smokers (≥ 20 cigarettes/day) (OR 10.2) (147). However, a metaanalysis identified a relative risk of 8.72 for an ischemic stroke in patients with migraine taking oral contraceptive pills. This risk rises disproportionately in women with migraine who smoke.

Stroke risk is higher for migraine with aura compared to migraine without aura. In one study, migraine was associated with an increased risk of ischemic stroke (OR, 3.54 for all migraines; 3.81 for migraine with aura, 2.97 for migraine without aura) but not of hemorrhagic stroke (OR, 1.1). A landmark Women’s Health Study, including data from 27,798 women with a follow-up period of 9 years, found patients who had migraine with aura between ages of 45 and 55 years had the highest relative risk for cardio and cerebrovascular disease, including stroke, which correlated directly with aura frequency (79). Aura frequency less than once a month conferred 2 times higher risk than in women without migraine, and more than 4 times higher risk when aura frequency exceeded once a week. By contrast, migraine without aura was not a risk factor for stroke (73).

Because the prevalence of stroke in young women is low, the absolute stroke risk in young female patient who experiences migraine is still low (17/100,000 to 19/100,000 women per year) in the absence of other risk factors. However, a World Health Organization study of stroke in young women found that the adjusted risk of ischemic stroke was significantly and directly associated with aura frequency (42). A large-scale epidemiologic study from the Netherlands showed that migraine patients, especially those with migraine with aura, had a less favorable cardiovascular risk profile compared with nonmigraine controls (125). There is some evidence that levels of platelet activating factor as well as von Willebrand factor increase during migraine attacks (29; 123). The increase in the platelet activating factor may be mediated through release of the CGRP from the trigeminovascular system.

The results of these clinical studies have been supported by neuroimaging findings. In a population-based MRI study from the Netherlands, patients with migraine had a higher prevalence of silent cerebellar infarcts compared with age- and sex-matched controls (78). Migraine was also associated with an increase in deep white matter lesions in women but not in men.

A 1996 World Health Organization study reported an increased risk of stroke among Europeans taking high-dose combined oral contraceptives (OR 5.30, 95% CI 2.56–11.0). However, the risk of stroke was not statistically significant in studying preparations containing less than 50 μg of ethinyl estradiol (OR 1.53, 95% CI 0.71–3.31) (06). By comparison, the average relative risk, compared with nonusers, is 4.1. With the use of low-dose oral contraceptives, the risk for stroke is lower and, in some studies, not significantly different from that of nonusers (136). Oral contraceptive use is a strong risk factor for cerebral venous thrombosis (17). An increased risk for stroke among patients who use triptans has not been found thus far, although triptans should be used in setting of an appropriate cardiovascular risk evaluation (65).

The decision to prescribe oral contraceptives to a patient with migraine should be individualized, and the risk/benefit ratio considered. Most young women with migraine without aura can take low-dose (35 µg or less of ethinyl estradiol daily) combined oral contraceptives safely unless they have other vascular risk factors. Patients should be strongly advised to refrain from or stop smoking. Any modifiable risk factor (eg, hypertension) should be controlled. For women with migraine with aura, the decision should be made on a case-to-case basis and should consider the woman’s age and other risk factors.

It is also important to note, current guidelines note that progestin-only pills can be safely offered to women with migraine with aura. However, progestin-only pills must be taken consistently and exactly on time to ensure contraceptive efficacy and are often difficult for patient compliance and, therefore, may increase risk of unintended pregnancy. Additionally, a progestin-only arm implant is considered safe to use in women with migraine with aura. Women with a family history of thromboembolic events should undergo screening for prothrombotic states and, if confirmed, should not use oral contraceptives.

A new-onset aura or a dramatic change in an established aura requires discontinuing estrogen-containing oral contraceptives (11). Older women and women with additional risk factors should be offered nonestrogen methods of contraception.

Interestingly, estrogen is likely to play a role in the pathogenesis of migraines and breast cancer, and studies have shown an association between them. However, a 2022 metaanalysis revealed that there was statistically significant inverse relationship between migraine and breast cancer in case-control studies 0.68 (95% CI: 0.56, 0.82), but no significant relationship was found in cohort studies 0.98 (95% CI: 0.91, 1.06) (68).

Migraine alone does not increase the risk of birth defects or pregnancy and labor complications (151; 09). Patients with migraine do not differ from the general population in the prevalence of miscarriages, toxemia, congenital anomalies, or stillbirths (131). However, in an Italian case-control study, pregnant women who had preeclampsia were more likely to report a history of migraine without aura than control women with uneventful pregnancies (46). Women who have migraine may also bear infants of lower birth weight.

Several studies have shown that approximately 70% of people who experience migraine have improvement in migraine symptoms during the second and third trimester of pregnancy, but 20% have no change in headache pattern, and 10% have their first attack of migraine during pregnancy (84; 88). Improvement is more common in women with menstrual migraine than in women with nonmenstrual migraine. Women whose migraine began at menarche have a higher remission rate than women whose headaches began at other times.

After menopause, when sex hormone levels stabilize, most women have improvement in their migraine. Women who have a physiologic menopause experience a more favorable course of migraine than women who have surgical menopause (84). In a study of 112 women taking hormone replacement therapy, 50 women reported improvement, 52 felt that their migraine worsened, and 10 noted no change. When the types of hormone replacement therapy were analyzed, significantly more women improved using transdermal estrogen than oral estrogen (83). For women whose migraine worsens with conventional hormone replacement therapy, the selective estrogen-receptor modulators may be appropriate. The new selective estrogen-receptor modulator, raloxifene, has been shown to increase mineral bone density and to lower total and low-density lipoprotein cholesterol without stimulating the endometrium (40). It can be used if a woman requires, but cannot tolerate, estrogen.

Clinical vignette

Vignette 1. A 30-year-old woman presented to the clinic complaining of headache attacks associated with her menses. The attacks consisted of unilateral, severe, throbbing headache associated with photophobia, phonophobia, and nausea. The headaches would start about 2 days prior to the onset of the menstrual flow, occur daily for 5 days and then disappear. The attacks were rather disabling. In mid-cycle, around the expected ovulation day, she would also develop an attack of lesser severity, typically lasting for 1 day and resolving. The duration of the cycles was quite steady. She had a past history of similar attacks, which for several years were not so tightly associated with menses, although the attacks were more frequent perimenstrually.

She was given treatment with frovatriptan, a long-acting triptan, at 2.5 mg daily, starting with the onset of the first headache and continuing daily for 5 days. This resulted in excellent and consistent control of the migraine attacks. She later became pregnant. During the first trimester, she experienced erratic migraine attacks, but early in the second trimester, the attacks disappeared completely for the duration of the pregnancy.

Vignette 2. A 33-year-old woman came to the clinic complaining of a headache that began at work. She was sitting at her desk and noticed zig-zag lines in her left visual field. This lasted for 30 minutes and was followed by a severe, throbbing headache in her right temporal area. The headache was associated with nausea and vomiting. She had no past history of headaches. She missed her last menstrual period but did not believe she was pregnant. Her neurologic examination was normal.

A urine beta-HCG test was positive. An MRI of the head without gadolinium was normal. A diagnosis of migraine with aura was made. She was educated on the benign nature of her disease and told that her migraine would probably improve during the later stages of pregnancy. A nonpharmacologic treatment plan (including advice on avoiding triggers, getting regular sleep, maintaining a regular meal schedule, and doing relaxation exercises) was initiated. With this treatment plan, she continued to have infrequent, mild to moderate migraine attacks (1 to 2 per month) that responded to oral ibuprofen. She became headache-free during the last trimester of her pregnancy. She delivered a healthy baby at 38 weeks of gestation. Four weeks after delivery, she had a typical migraine with aura attack that responded again to nonsteroidal anti-inflammatory drugs (NSAIDs). Over the following months, despite optimal nonpharmacologic treatment, her attack frequency increased to 4 headaches per month. A typical attack lasted 12 to 24 hours, and she had significant disability during the attack. She was, therefore, put on amitriptyline 25 mg/day as a preventive medication. When the dose was increased to 50 mg/day, she had a good response, with decreased migraine attack severity and frequency. Acute attacks were managed with almotriptan 12.5 mg with good results.

Biological basis

Etiology and pathogenesis

The etiology of migraine in general and of migraine during the different periods and events in a woman’s life that are associated with hormonal changes is not known. Genetic factors play a role in making a woman susceptible to migraine (57). Hormonal changes act as triggers for migraine attacks in susceptible women.

Menstrual migraine appears to be an abnormal response to normal, rather than abnormal, hormonal fluctuations (105; 84). Estrogen can affect neuronal excitability by both genomic and nongenomic mechanisms (44). Plasma concentrations of follicle-stimulating hormone, luteinizing hormone, prolactin, and testosterone obtained throughout the menstrual cycle are normal (45). Somerville studied 14 women who had migraine exclusively related to menstruation and showed that migraine appeared to be triggered by falling levels of estrogen during the late luteal phase of the normal menstrual cycle.

When migraine attacks were delayed by maintaining high plasma estradiol concentrations, endogenous progesterone levels fell as usual, resulting in menstruation that was not associated with a migraine attack. Only when plasma estradiol levels fell did an attack ensue (139; 140; 141; 142). Once the migraine mechanism is activated by falling estrogen levels, subsequent estrogen administration does not influence the occurrence of an attack. A study of 38 women with migraine without aura recorded migraine attack frequency and urinary levels of estrone-3-glucuronide and pregnanediol-3-glucuronide, metabolites of estradiol and progesterone, over 3 menstrual cycles. Migraine attacks were most likely to occur during the late luteal and early follicular phases, when estrogen levels were falling or low, and were least likely to occur when estrogen levels were rising (86).

Additional evidence that menstrual migraine is due to estrogen withdrawal include the following: (1) some women who take combined oral contraceptives experience migraine during the pill-free week, when estrogen falls after 21 days of high levels; (2) migraine may improve during pregnancy, when estrogen levels rise gradually, but can reoccur immediately postpartum, when estrogen levels plummet; (3) some women have migraine during the week without estrogen replacement therapy in the old regimen of 21 days on and 7 days off treatment; (4) headache frequency increased when an estrogen preparation was followed by a nonestrogen preparation in a study of women with bilateral oophorectomies; and (5) migraine often improves after menopause, when the ovarian cycle ceases, estrogen levels remain low, and there is little fluctuation (84). A study showed that gene expression of pain-associated neuropeptides in the trigeminal ganglia of female mice is modulated by estrogen through estrogen receptor alpha (116). Estrogen also increases the production of calcitonin gene-related peptide (CGRP), a peptide well known to be implicated in pain perception, by the trigeminovascular system (55). A plethora of evidence also supports an estrogen-mediated increase in nitric oxide production, through activation of estrogen receptor a (ER-a). Despite its established vasodilatory and cardioprotective properties, the role of nitric oxide in migraine may be far more complicated because it has been shown to promote the production of proinflammatory mediators and activate trigeminal afferents (75). Polymorphisms in the estrogen receptor and in the follicle stimulating hormone receptor (FSH-R) can also account for the variability in migraine propensity and clinical phenotype (110). Moreover, estrogen alters the expression of genes coding for proteins that may be involved in inflammatory pain, eg, extracellular signal-regulated protein kinase (117). Modulation of neurotransmission by estrogen within the trigeminal nucleus caudalis is another potential mechanism (91). Polymorphisms of the progesterone receptor have been implicated as well (34; 21).

Neuroimaging. Brain MRI studies support both structural and functional sex differences in patients with migraine. Maleki and colleagues used high-field MRI to compare male and female patients with migraine to healthy controls (90). Women with migraine were found to have thicker posterior insula and precuneus cortices as compared to male patients and healthy controls. Using functional MRI, noxious thermal stimuli produced stronger responses in the amygdala and parahippocampus in female as compared with male participants with migraine. This study was of particular interest because the anatomical differences found in the females who experience migraine are attributable neither to the female sex nor the migraine disease state alone. One can, thus, hypothesize that a combination of being female and having migraine may lead to changes in brain anatomy and sensory responsiveness and result in differences in migraine characteristics (30).

Neurophysiology. Cortical spreading depression has been implicated in the pathogenesis of migrainous aura. Effects of sex hormones on cortical spreading depression might be another explanation of the variability of headache syndrome during a woman’s lifetime. Both estrogens and, more surprisingly, progesterone increase cortical spreading depression susceptibility. Furthermore, L-kynurenine is more efficient in suppressing cortical spreading depression when progesterone levels are high. This might be relevant for the aggravation/appearance of migraine with aura during pregnancy or intake of a combined contraceptive pill (33).

In a pilot study of 11 women receiving combined hormonal contraceptives who were developing migraine attacks occurring exclusively during the estrogen-withdrawal period, the investigators evaluated the nociceptive reflex, a neurophysiological assessment of pain control systems, during the third week of active treatment and during the hormone-free interval (38). During the hormone-free interval, the nociceptive withdrawal reflex threshold was significantly lower than that of the third week of hormonal treatment. These results indicate that estrogen withdrawal may mediate increased sensitivity to somatosensory stimuli in women with migraine attacks occurring during the hormone-free interval.

Cell biology. A role for intracranial mast cell activation in migraine pathophysiology has been proposed (145). Fluctuations in dural mast cell density during the estrous cycle were demonstrated. This effect was abolished by splenectomy, which suggests estrogen-related recruitment of mast cells from the spleen. This might, at least in part, explain the higher frequency of migraine in women as well as the fluctuations in headache during the menstrual cycle (14).

In 1 study, nitroglycerin, a well-known nitric oxide donor that provokes spontaneous-like migraine attacks, administration in male and female rats induced a significant Fos expression in the paraventricular and supraoptic nucleus of the hypothalamus, central nucleus of the amygdala, nucleus of tractus solitaries, area postrema and nucleus trigeminalis caudalis. Sex, however, influenced the Fos expression, with higher expression in the specific nuclei in the female rats. Ovariectomy significantly reduced nitroglycerin-induced Fos immunoreactivity in all the nuclei that were activated in the intact females, suggesting that at least partially the vascular effect of estrogens may be mediated through nitric oxide and that these effects are associated with an increased production of nitric oxide (61). An alternative mechanism of action of estrogens could be linked with calcitonin-gene related peptide (CGRP), a vasodilator agent that plays a role in determining migraine attacks.

Neurochemistry and endocrinology. The high estrogen levels sustained during pregnancy are thought to protective against migraine attacks. It has also been suggested that the improvement might be related to changes in serotonin metabolism during pregnancy and increased endorphin concentrations during the last 2 trimesters (130). The rapidly falling concentrations of estrogen that follow childbirth are thought to be responsible for postpartum headaches.

Serotonergic pathways are certainly involved. A PET study showed that migraine patients had lower brain serotonin levels between attacks compared with controls and that those levels increased to the normal range during an attack (121). The study suggested that migraine may be associated with a low-serotonin state between attacks. Furthermore, there is evidence that serotonin synthesis is influenced by ovarian hormones (12). Data from primate experiments support the notion that the serotonergic tone is enhanced by estrogen (93; 94).

In general, ovarian steroid hormones regulate nociceptive pathways at multiple levels. The trigeminal vascular system is significantly modulated by estrogens and is probably the major determinant of sexual dimorphism observed in migraine (64). This relationship is well exemplified by the interaction between CGRP-ergic and serotoninergic pathways with estrogens (64).

In a case-control study of patients with menstrually-related migraine, the investigators confirmed that the premenstrual withdrawal of estradiol influences the function of the trigeminovascular system (70). In the same study, the investigators also found that there is disturbed systemic as well as trigeminovascular vasodilator system cyclicity in patients with menstrually related migraine, which may augment their susceptibility to migraine around the time of menstruation.

Prostaglandins intensify nociception. This might provide a partial explanation for the commonly held view that menstrual migraine is generally more severe and less responsive to treatment compared with migraine not related to the menstrual cycle. However, there is little support for prostaglandins as primary mediators in migraine. Nattero and colleagues found that controls and migraine sufferers had similar prostaglandin levels during the cycle. However, migraine patients had significantly higher levels during a migraine attack than on any other day of their menstrual cycle (106). These results do not lend support to the idea that elevated prostaglandin levels per se trigger menstrual migraine. Because dysmenorrhea is clearly related to elevated prostaglandin levels (which intensify uterine contractions), one would expect that migraine and dysmenorrhea would correlate if prostaglandin levels played a primary role in menstrual migraine pathophysiology.

In a study in rats, deficiency of female sex hormones was found to influence prostaglandin E2 (PGE2) and CGRP levels in the midbrain periaqueductal gray. More specifically, the authors found that circulating 17β-estradiol and/or progesterone influence the levels of PGE2 and CGRP in the periaqueductal gray, a lower level of 17β-estradiol and/or progesterone augments PGE2 and its EP3 receptor, and that PGE2 plays a role in regulating expression of CGRP in the periaqueductal gray (153).

Mechanisms relating to dopamine and opioids have also been proposed for menstrual migraine, including reduction in serum dopamine beta-hydroxylase concentrations (89), dysfunction of opioid control of the hypothalamic-pituitary-adrenal axis (49), reduced levels of magnesium (51), and platelet dysfunction.

Estradiol and progesterone seem to have opposite effects on neuronal excitability, with estradiol being excitatory and progesterone inhibitory. Estradiol augments N-methyl-D-aspartate-mediated glutamate receptor activity, whereas progesterone enhances gamma-aminobutyric acid-mediated chloride conductance. Sex steroid regulation of the balance of neuroexcitatory and neuroinhibitory activities may also have a role in modulating clinical susceptibility to migraine (53).

Other hormonal systems also have been implicated in the pathophysiology of migraine. Insulin resistance has been found in migraine patients compared with control subjects (27), raising the question of whether diabetes could be linked to migraine. Bosco and colleagues detected an increase in plasma prolactin levels during migraine attacks in migraine and tension-type headache patients with microprolactinoma--although the sample was small and did not include control subjects (15). Chronic migraine and chronic tension-type headache could be associated with abnormal sleep melatonin secretion (23). Whether this is the cause or the effect of chronic headache remains to be clarified. The orexin system has also been linked to migraine and medication overuse headache. In a study by Sarchielli and colleagues, CSF levels of corticotrophin releasing factor (CRF) and orexin-A were elevated in patients with chronic migraine and medication overuse headache (124).

Epidemiology

Fourteen percent of women report that their attacks are exclusively related to their menstrual periods, but headache diaries reveal that only 0.85% to 7% of women with migraine have pure menstrual migraine. Eleven percent of women have migraine onset at menarche (58). Women who had migraine onset at menarche are more likely to have menstrually-related migraines (19).

Migraine improves during pregnancy in 50% to 90% of women, mostly in the second and third trimesters, but remains unchanged in 5% to 30% of cases (88; Silberstein 1997a; 122). After menopause, when sex hormone levels stabilize, most women have improvement (52). Neri and colleagues, in a sample of 556 postmenopausal women, found a 13.7% prevalence of primary headaches (107). Migraine improved in almost two thirds of postmenopausal patients with migraine, but tension-type headache worsened or did not change in 70% of women. None of the study patients reported migraine with aura or cluster headache. Two thirds of patients with migraine who underwent surgical menopause reported a worsening of their headache. Ten percent of women may continue to experience migraine in the seventh and eighth decades of life.

Prevention

There is no primary or secondary prevention for migraine.

Differential diagnosis

Headaches that are exclusively linked to the menstrual cycle are almost always benign, and diagnostic studies are rarely indicated.

Headaches that begin during pregnancy can raise concerns about serious underlying illness, such as eclampsia or preeclampsia, idiopathic intracranial hypertension, subarachnoid hemorrhage from aneurysm or arteriovenous malformation, cortical vein thrombosis, pituitary tumor, and choriocarcinoma (131). Pregnancy can also unmask an intracranial neoplasm. The diagnosis of a serious disorder can easily be missed, as the symptoms of headache, vomiting, and visual disturbance can be encountered in pregnant women with or without preeclampsia (32). The history of the headache onset and a careful neurologic examination, including blood pressure measurement, ophthalmoscopy, and a thorough check for signs of meningeal irritation, are usually sufficient to rule out these problems. Migraine usually can be diagnosed on clinical grounds alone, but uncertainty may result when the patient is seen during the initial attack, particularly if the migraine is associated with aura. In such cases, neuroimaging should be considered.

Headache in the postpartum period is common and usually caused by migraine (143; 122). Sances and colleagues found that migraine recurred during the first week after delivery in 34% of women and after the first month in 55% (122). Other causes of headache, such as venous sinus thrombosis, should be considered if the headache is not typically migrainous in nature, even if it is not accompanied by fever, convulsions, altered consciousness, hemiplegia, or papilledema (10). Postpartum headache may also result from low intracranial pressure due to a CSF leak caused by an inadvertent dural puncture during epidural anesthesia (113).

In rare instances, hormone replacement therapy triggers benign intracranial hypertension (129). Transient visual obscurations and tinnitus commonly accompany benign intracranial hypertension. These features should prompt lumbar puncture (after CT) to exclude this important cause of headache.

Diagnostic workup

Neuroimaging studies are rarely indicated for typical migraine linked exclusively to the menstrual period. An unexplained increase in the frequency or severity of attacks requires clinical evaluation but is rarely a result of serious organic disease if high-risk features are absent.

Headaches that are of recent onset or are associated with high-risk features require neuroimaging studies. When considering imaging studies for a pregnant woman with headache, the potential risk to the mother and fetus should be weighed against the potential benefit of the procedure. Pregnancy is not a contraindication to neuroimaging studies. In most cases, MRI is the neuroimaging study of choice for pregnant women (128).

Head CT is relatively safe during pregnancy (exposure of 0.001 cGy to the fetus) and is the study of choice for head trauma and for suspected intracranial hemorrhage. The potential risk, if any, of MRI during pregnancy is still unclear. Gadolinium crosses the placental barrier and is excreted through the fetal kidneys; even though no ill effects have been demonstrated, gadolinium injection should be avoided, as should CT with a contrast agent (127; 131).

Management

The treatment of acute migraine attacks caused by hormonal fluctuations is similar to that of migraine attacks unrelated to hormonal fluctuations. Patients with typical episodic migraine without aura are less likely to be referred to neurologists than are patients with unusual features, comorbid disorders (medical and psychiatric), severe clinical disability, or refractoriness to standard therapy. Menstrual migraine is said to be more difficult to treat than migraine not associated with the menses. Clinical evidence supports this notion. In a study of 64 women with migraine, perimenstrual attacks were found to be significantly longer, less responsive to acute treatment, and cause greater work-related disability compared with attacks not associated with the menses (58). In general, treatment options that can be considered include acute attack treatment, preemptive treatment or short-term prophylaxis, standard migraine prophylaxis, hormonal treatment, and others.

Treatment of acute attack. The drugs of choice for acute attacks include NSAIDs, the triptans, and dihydroergotamine (47). In a randomized placebo controlled trial, a single dose of naratriptan 2.5 mg was significantly superior to placebo in alleviating pain and associated symptoms of menstrually-related migraine attacks (95). Another placebo-controlled study of 93 women with menstrually-related migraine showed that sumatriptan 100 mg was effective in the acute treatment of migraine for both perimenstrual attacks and nonmenstrually-related attacks (43). Antiemetics are important adjuncts for those who experience severe nausea or vomiting. Second-choice abortive agents include narcotics, corticosteroids, neuroleptics, and a course of intravenous dihydroergotamine. Rapoport and colleagues reported that migraine attacks occurring during the menstrual period responded better to zolmitriptan (a 2-hour response rate of 72%) than attacks that occur outside the menstrual period (a 2-hour response rate of 67%) (118). However, there were no comparisons of patients with migraine attacks confined to the menses versus patients with attacks unrelated to the menstrual cycle. However, Bhambri and associates reported that when comparing treatment of migraine attacks during menstrual and nonmenstrual time periods with eletriptan, there were higher recurrence rates and lower sustained response rates at 2 to 24 hours in the group treated during the menstrual time period (67; 13). In a pooled analysis of 3 double-blind, randomized, crossover, multicenter clinical trials, frovatriptan was found to be more suitable for the management of menstrual migraine occurring during the hormone withdrawal phase, producing pain free and pain relief state in 25% and 51% of patients respectively (nonsignificant difference against rizatriptan, almotriptan, or zolmitriptan), but produced more sustained relief, with 71% and 83% of patients being pain free or having pain relief at 24 hours (04).

Preemptive treatment or short-term prophylaxis. Several drugs have been used as short-term prophylaxis for menstrual migraine, including NSAIDs, ergotamine, dihydroergotamine, magnesium, and the triptans (19; 132). Three triptans have been shown to be effective in menstrual migraine prevention. These include sumatriptan 25 mg 3 times a day, beginning 2 to 3 days before the expected day of headache onset and continued for a total of 5 days; naratriptan 1 mg twice daily for 5 days, beginning 2 days before the expected onset of menses; and frovatriptan 2.5 mg once or twice daily, starting 2 days before the anticipated start of the headache and continuing for 6 days (Newman 1998; Newman 2000; 131; 101). The combination of a triptan with an NSAID (such as sumatriptan and naproxen sodium) can be an effective way of managing menstrual migraine while also addressing other prominent menstrual symptoms (03; 92).

Menstrual migraine can also be treated by the premenstrual use of standard prophylactic medication (antidepressants, beta-blockers, valproate, or calcium channel blockers). The following medications may be used: amitriptyline, 25 mg to 50 mg at bedtime; nadolol, 40 mg to 160 mg every day; verapamil, 80 mg to 160 mg 3 times a day; or valproate, 250 mg 3 times a day. Women who already use prophylactic medication for nonmenstrual migraine can increase the dose prior to their menses.

Hormonal treatment. Hormonal manipulation was also shown to be effective as short-term prophylaxis for menstrual migraine. Percutaneous estradiol gel (1.5 mg in 2.5 gr gel), started 2 days before the menses and continued for 7 days, was associated with reduced frequency, severity, and duration of menstrual migraine (39). The transdermal formulation of estradiol (100 µg) is the most convenient. In their study, Pradalier and colleagues found that a perimenstrually applied 100 mcg estradiol patch was effective in preventing menstrual migraines, whereas a 50 mcg patch was not (114). It is, thus, reasonable to believe that estrogen levels need to be kept above a certain cutoff to prevent an attack. In a placebo-controlled study, MacGregor and colleagues examined the efficacy of estradiol gel (1.5 mg) in menstrual migraine prevention (86). Women treated 6 cycles in a cross-over design. Gel was applied daily from the 10th day after ovulation to the second day of menses. During the time of gel application, estradiol treatment was associated with a 22% reduction in migraine days compared with placebo. However, migraine increased in the estradiol group during the 5 days following cessation of gel application compared with the placebo group. Patch administration of estrogen was evaluated in 2 small studies and seems to be much less effective, although they may have been underpowered (114; 138). In a retrospective review of 229 patients with menstrually-related migraine, Calhoun and Ford found that hormonal manipulation reduced chronification of headache and analgesic use (25). NSAIDs (eg, naproxen sodium 250 mg 3 times daily or indomethacin 25 mg 3 times daily) may also be used perimenstrually for migraine prophylaxis. The drug is started 2 days before menses and continued for 5 days. Dihydroergotamine mesylate given intranasally to 40 women with menstrual-related migraine for 6 days, starting 2 days before the expected onset of the migraine attack, was shown to reduce the severity of headache (133). Frovatriptan 2.5 mg a day was more efficacious than transdermal estrogen and naproxen sodium for the management of menstrual migraine in a small, open-label, nonrandomized trial (63). Based on a systematic review of 19 prospective, double-blind, randomized trials, Pringsheim and colleagues concluded that grade B recommendations could be made for sumatriptan 50 and 100 mg daily, mefenamic acid 500 mg and rizatriptan 10 mg for the acute treatment of menstrual related migraine (115). For prevention, the same authors made grade B recommendations for transcutaneous estrogen 1.5 mg daily, frovatriptan 2.5 mg twice daily, and naratriptan 1 mg twice daily. The choice of agent should be based on patient profile, drug side effects, and physician experience.

Other. A short course of a neuroleptic or corticosteroids may be used. These include: prednisone, 80 mg/day for 7 days to 10 days with a rapid taper; dexamethasone, 20 mg/day for 4 days to 7 days with a rapid taper; or chlorpromazine 10-s 50 mg twice a day for 4 days to 7 days.

A placebo-controlled, double-blind study of 24 women with premenstrual syndrome and migraine has shown that oral magnesium (360 mg of magnesium pyrrolidone carboxylic acid) decreases the severity of the premenstrual symptoms and the duration and intensity of menstrual migraine occurring prior to the onset of menstruation (50; 51; 48).

For refractory cases, more radical approaches can be undertaken (in conjunction with the patient’s gynecologist). Inducing a temporary state of anovulation, either with the synthetic androgen danazol (26), the antiestrogen tamoxifen (108), or GnRH analogs (103), may be efficacious in prophylaxis.

Table 2. Current Acute Headache Treatments and Prescription Directions for Adults

Standard migraine prophylaxis. Preventive treatment for migraine should be considered when 3 or more attacks that are prolonged and unresponsive to abortive measures occur each month (59). However, many women remark that standard prophylactic medications such as propranolol effectively eliminate their headaches, except for the 1 headache a month they experience just before or during their menses.

Treatment of migraine during pregnancy. Treatment for young, healthy women with menstrual migraine includes NSAIDs administered 2 days to 3 days before menses and continuing through menstruation.

Popular, but ineffective, therapies include diuretics and vitamins.

There is no place for hysterectomy in the management of menstrual migraine (84).

Because migraine usually improves after the first trimester of pregnancy, many women can manage their headaches with reassurance and nonpharmacologic means of coping, such as ice, massage, and biofeedback. Riboflavin is safe at physiological doses, but safety has not been proven at the high doses used for migraine prevention. Despite the fact that 67% of women take medications during pregnancy and 50% take them during the first trimester, insufficient knowledge exists about the risks of birth defects from drug exposure. Acetaminophen (alone or with codeine), codeine, or other narcotics can be used during pregnancy. Narcotics should be avoided in the late third trimester because of possible neonatal withdrawal symptoms. Indiscriminate codeine use may present a risk to the fetus during the first or second trimester. Cleft lip, cleft palate, inguinal hernia, hip dislocation, and cardiac and respiratory system defects have been reported (131). However, migraine per se does not constitute a risk factor for congenital malformations (22).

NSAIDs should be avoided in the late third trimester because they can lead to premature closure of the ductus arteriosus. Aspirin can also increase intrapartum blood loss and impair neonatal hemostasis. Aspirin and NSAIDs in the third trimester also increase the risk of preeclampsia, prolonged labor, and neonatal pulmonary hypertension (09). For prolonged migraine attacks (status migrainosus), intravenous hydration and 10 mg of intravenous prochlorperazine can be given for head pain and nausea, supplemented by intravenous narcotics or corticosteroids (131).

The effect of sumatriptan on pregnancy outcome has been evaluated (81). In a review of pregnancy registries, Reiff-Eldridge and colleagues found a 3.8% rate of birth defects in women who were exposed to sumatriptan during the first trimester. This is comparable to the risk in the general population (3% to 5%). No defect pattern was observed (119). Olesen and colleagues identified 34 women who were exposed to sumatriptan during pregnancy and compared their pregnancy outcomes with those of healthy women and migraine controls (109). The risk of preterm delivery was elevated among women exposed to sumatriptan during pregnancy compared with both migraine controls (OR, 6.3; CI, 1.2-32.0) and healthy women (OR, 3.3; CI,1.3-8.5). The risk of having a low birthweight newborn was elevated for all patients with migraine compared with healthy controls (OR, 3.0; CI, 1.3-7.0). Kallen and colleagues identified 912 infants who had been exposed to antimigraine drugs prenatally, mostly sumatriptan (71). There was an insignificant increase in the rate of preterm deliveries and low birthweight infants. No increase in the rate of congenital malformations was seen. Current data show no evidence for teratogenicity from sumatriptan use in pregnancy; however, the drug should be used with caution, weighing potential risks against benefit.

Pharmacologic migraine prophylaxis is rarely indicated in pregnancy and should be used only as a last resort and after the risks have been explained to the patient. Beta-blockers have been used in these circumstances. Propranolol has been used safely during pregnancy to treat hypertension, and there is no evidence that it is teratogenic (18). Metoprolol is also considered safe during pregnancy (112). Dey and colleagues reported on a woman with severe migraine and preeclampsia who was successfully treated with labetalol and delivered a healthy baby with no complications (41).

Valproate may cause neural tube defects and is contraindicated during pregnancy. Amitriptyline must be stopped at least 2 weeks before the due date because respiratory distress and feeding difficulties have been described in infants born of women taking tricyclics up to the time of birth. None of the preventive agents are classified as category A (controlled studies show no risk).

CGRP antagonists, erenumab, galcanezumab, fremenumab, eptinezumab, as well as “gepants,” such as ubrogepant, atogepant, and rimegepant should be avoided in pregnancy as there are no long-term safety data. However, if a patient chooses to continue on these medications, their outcomes may be reported and prospectively followed. Wash out period of at least 5 half-lives should be completed during pregnancy planning in order to clear the antibodies prior to conception as they have long, multi-month half-lives.

Although there are limited high powered safety studies on pregnant patients receiving onabotulinumtoxinA, it is generally considered safe in pregnancy by experts. In a study of 32 patients who remained on treatment with onabotulinumtoxinA through pregnancy, many showed a good response, whereas 11 out of 13 who withdrew from treatment experienced a relapse (155). In a retrospective review of 9 subjects who received onabotulinumtoxinA injections for migraine during pregnancy, none of the 10 pregnancies reported any organ malformations at delivery (144).

Although the medical risks of pregnancy have declined significantly over the past 50 years, the medicolegal risk of caring for the pregnant patient has increased dramatically. Many experts fear that any malformation (which occurs spontaneously in 2% to 3% of births) will be blamed on medication (112). The current medicolegal environment makes it difficult to recommend any preventive therapy as safe for all headache patients. Full disclosure of potential risks and documentation of informed consent is advisable. For those patients who require specific antimigraine acute or preventive therapy, consultation with a medical geneticist or perinatologist can be helpful.

When migraine is first diagnosed or worsens during pregnancy, a secondary cause of headache (eg, subarachnoid hemorrhage, cerebral sinus venous thrombosis, benign intracranial hypertension) must be carefully excluded.

Migraine treatment and menopause. For the perimenopausal woman with migraine, continuous combined estrogen and progesterone (or estrogen alone, if the uterus has been removed) replacement is the preferred therapy. This can be achieved with a 50 µg/day estrogen skin patch or an oral estrogen, with half the daily dose given every 12 hours to maintain optimal stability. When the uterus is present, progesterone should be added, either as low-dose medroxyprogesterone acetate 2.5 mg every evening or micronized progesterone 100 mg every evening (52).

Women who develop headaches as a result of hormone replacement therapy are difficult to manage. Reducing the dose of estrogen or changing the estrogen type may reduce headache (134). When headaches are associated with estrogen withdrawal, changing from interrupted to continuous administration may be effective. Changing the route of administration may also reduce headache occurrence. Raloxifene, a selective estrogen receptor modulator, can be used if a woman requires, but cannot tolerate, estrogen. Another approach is to use targeted drug delivery, such as a progesterone-containing vaginal gel. This maximizes progesterone’s effect on the uterus while minimizing its potential adverse effects, including headaches.

Estrogen and selective estrogen receptor modulators increase the risk of venous thromboembolism 3-fold. The risk for breast cancer increases after 5 years of hormone replacement therapy. The excess risk of estrogen use is 2/1000 users over 20 years in women aged 50 years to 70 years if hormone replacement therapy has been used for 5 years.

The Women’s Health Initiative trial included 27,500 postmenopausal women. The active treatment in this trial was conjugated estrogen for women without a uterus and estrogen with medroxyprogesterone acetate for women with an intact uterus. The study was terminated early because of increased risk of breast cancer, and evidence for greater overall risk than benefit in the active treatment group. More surprisingly, the risk of myocardial infarction (and of stroke) was higher in the hormone replacement therapy group; however, the absolute risk was small. There was no difference in mortality between the active treatment and the control groups (120; 05). Migraine has not been shown to be a risk factor for stroke in postmenopausal women (76). Therefore, in considering symptomatic hormone replacement therapy for postmenopausal patients with migraine, the usual indications and contraindications should be applied.

Special considerations

Pregnancy

Migraine that develops for the first time during pregnancy usually appears during the first trimester. Approximately 10% of patients with migraine report that symptoms began during pregnancy (09). Migraine with aura is more common than migraine without aura when onset occurs during pregnancy. As noted above, approximately 70% of migrainous women improve during pregnancy, usually during the second and third trimesters.

Forty percent of women will have migraine attacks during the week following childbirth (143). As mentioned above, breast-feeding may have a protective effect towards migraine recurrence.