Definition

Migraines can be divided into two major subtypes: migraine with aura and migraine without aura. Migraine without aura is the most common type. In both subtypes, the headaches are unilateral, often pulsatile, moderate to severe in intensity, aggravated by routine physical activity, and associated with nausea, photophobia and/or phonophobia. Furthermore, such headaches are episodic, with each attack typically lasting 4–72 hours.⁴

In a population-based study of migraine prevalence, approximately 37% of women with migraine reported symptoms of aura.¹ Migraine with aura (unlike migraine without aura) is characterized by the presence of reversible focal neurological symptoms that usually develop over 5–20 minutes and last for <60 minutes.⁴ The aura usually precedes the headache, although in many instances the aura can occur during or after the headache.

Pathogenesis

The brain's susceptibility to migraine attacks is fundamentally associated with the excitability of the neuronal cell membranes of the occipital cortex.⁵ Cortical spreading depression is the most widely accepted mechanism for the development of aura. In cases of headache following aura, cortical spreading depression is hypothesized to lead to activation of the trigeminal nucleus caudalis—part of the central pathway of migraine pain generation.⁵

The pathogenic mechanism of migraine pain in the absence of aura is still hotly debated and is an active area of investigation. The main area of dispute is whether migraine pain originates peripherally, from activation of the trigeminal nerve via release of vasoactive factors from blood vessels, or centrally, in the trigeminal nucleus in the brainstem. The exact mechanisms for the central origination of pain in the brainstem have yet to be elucidated, so the activity observed in the brainstem during migraine could be a consequence of peripheral activation and not the cause.⁶ The other possible mechanism is one in which peripheral activation is combined with a change in central pain modulation.⁷

Regardless of the source of signaling—central or peripheral—the common mechanism underlying migraine headache seems to be dilatation of the large cranial vessels beneath the dura mater. These vessels are innervated by the trigeminal nerve and form part of the trigeminovascular system. Several vasoactive peptides have been identified in the cell bodies of trigeminal neurons, including calcitonin gene-related peptide (CGRP), substance P, and neurokinin A. Of these peptides, CGRP seems to be the most important in the pathogenesis of migraine headache.⁸

Complications of migraine with aura include prolonged aura without infarction and migrainous stroke.^{9, 10} The pathophysiological mechanisms that have been linked to ischemia and the development of true migrainous infarction include neurologically mediated cortical spreading depression,¹¹ reversible vasoconstriction,¹² intracranial arterial dissection,¹³ and arterial embolism.¹⁰

Epidemiology

Pregnancy has a variable effect on the frequency and intensity of migraines. The available literature indicates that most women seem to experience either an improvement or no change in migraine frequency during pregnancy.³ The percentage of women who show improvement during pregnancy ranges from 18% to 86%,^{14, 15, 16, 17, 18} although studies performed to date have not established an objective set of criteria that can predict which patients will improve.

Studies have consistently found that migraine with aura is less likely to improve during pregnancy than migraine without aura.^{15, 18, 19} The reason for this finding is not obvious, but it could be related to increased endothelial reactivity in patients with migraine and aura, compared with those without aura. The additional, and potentially adverse, effect of pregnancy on endothelial reactivity might lead to an increase in the frequency of attacks in patients with migraine and aura. The case report in Box 1 features a patient with migraine and aura, whose attacks were frequent enough to warrant preventive therapy throughout pregnancy.

In a study of 98 women undergoing induced ovulation, those with a history of migraines (25 individuals) had a higher frequency of headaches during the phase of hypothalamus–pituitary–ovary-axis downregulation than those without migraines.²⁰ The downregulation phase of induced ovulation is characterized by very low 17-estradiol serum levels. Similarly, levels of estrogens have been observed to fluctuate in the perimenopausal state. The observation that migraine attacks increase during perimenopause supports the concept that fluctuations and abrupt changes in estrogen levels can act as triggers for migraine.^{21, 22}

Migraine headaches do not seem to directly affect the outcome of pregnancy,²³ although they could exert an indirect effect.^{24, 25} Patients with migraine have a higher risk of developing pre-eclampsia than individuals without migraine.²⁶ Of the limited number of studies that evaluated other pregnancy outcomes in patients with migraine, most did not report an increase in the incidence of other poor outcomes, such as preterm labor, low birth weights, or congenital anomalies.^{27, 28}

Additional indirect associations between migraine and outcomes in women of reproductive age were noted in an analysis of a national database of hospital discharges. This analysis showed that the prevalence of migraine in hospitalized women was only 185 per 100,000 deliveries, whereas the expected prevalence for women of childbearing age is 12–18%.²⁹ The prevalence of migraine in pregnant women might have been underestimated by this study, as only women actively affected by severe migraines were assigned a migraine discharge diagnosis. Studying only the discharge codes during pregnancy also led to an underestimation of peripartum migraine prevalence because many women do not experience migraines during pregnancy or have only very mild attacks.

Differential diagnosis

Migraines that develop during pregnancy or after delivery can be difficult to distinguish from other potentially serious, more-common complications.³⁰ Conditions that are known to be secondary causes of headache in the nonpregnant population include vascular disease, stroke, pseudotumor cerebri, hemorrhage, tumors, and infection. Additional conditions to consider during pregnancy include pre-eclampsia and postepidural headache.³⁰

The differential diagnosis of migraine headaches postpartum requires attention to the clinical characteristics of the headache. Characteristics to consider are the association of the headache with elevated blood pressure (which could signal postpartum pre-eclampsia), the sudden onset of an atypical headache, and variations to the usual nature of the migraine, such as the onset of new neurological symptoms. Postpartum women with any of these characteristics should be evaluated in the emergency department and neuroimaging should be strongly considered. The differential diagnoses of a sudden-onset 'thunderclap' headache include reversible cerebral vasoconstriction syndrome (RCVS), aneurysmal subarachnoid hemorrhage, central venous sinus thrombosis, carotid or vertebral artery dissection, pituitary apoplexy, and intraparenchymal hemorrhage. Neuroimaging is required to rule out these differential diagnoses.

RCVS encompasses a diverse group of conditions, including hypertensive encephalopathy, and vasculopathies associated with pregnancy and postpartum (postpartum angiopathy). RCVS is characterized by sudden onset of a severe (thunderclap) headache that relapses within the first few days to weeks and resolves in most patients by approximately 12 weeks after presentation.³¹ A similar syndrome can be seen with pre-eclampsia or eclampsia occurring antepartum or postpartum. Interestingly, approximately 15% of patients with RCVS have a history of migraine headaches.¹² A diagnosis of RCVS requires the exclusion of other causes of sudden onset of headache, which is usually accomplished by brain CT or MRI, with accompanying arterial and venous imaging (CT or magnetic resonance [MR] angiography and venography) to evaluate for cerebral edema or vasoconstriction, and to exclude cerebral venous thrombosis. The imaging abnormalities associated with RCVS include vasogenic edema, which is hypodense on CT and hyperintense on T2-weighted MRI. These abnormalities reverse within 6–12 weeks.¹² Cerebrospinal fluid obtained via lumbar puncture can exclude vasculitis or infection.³¹

Neuroimaging

Neuroimaging might be considered in the evaluation of women with headaches that are of new onset or different from their usual headaches. In most patients with established migraine, however, neuroimaging is not usually necessary. Evidence-based guidelines issued by the American Academy of Neurology suggest that neuroimaging should be considered in the following patients with nonacute headache: patients with an unexplained abnormal finding on neurological examination; patients with atypical headache features or headaches that do not fulfill the strict definition of migraine or other primary headache disorders (or who have an additional risk factor, such as immune deficiency); and patients with sudden severe headache.³²

A number of features signify an increased risk of finding an abnormality on neuroimaging in patients with a nonacute headache. These indicators include a rapidly increasing headache frequency, ataxia, localized neurological signs, and the occurrence of a severe headache that either awakens a patient from sleep or can be described as the worst headache the patient has ever experienced.^{32, 33} A head CT scan (with and without contrast) is sufficient in many patients when neuroimaging is deemed necessary.³⁴ MRI is indicated when a posterior fossa lesion or cerebrospinal fluid leak is suspected. MR angiography and MR venography are indicated when an arterial or venous lesion (for example, cerebral venous thrombosis) is considered in the differential diagnosis.³⁴

Most neuroimaging diagnostic techniques pose a negligible risk to the fetus.³⁵ The average radiation to which the mother is exposed from a head CT scan is <0.01 Gy, whereas the threshold for fetal damage with ionizing radiation directed towards the maternal pelvis is >0.1–0.2 Gy.³⁵ To maintain a margin of safety, the National Council for Radiation Protection and Measurements has set the upper limit of acceptable radiation in all CT scans at 0.05 Gy.³⁶ Other forms of electromagnetic radiation, such as that produced in MRI, do not pose the same level of risk associated with ionizing radiation.

Gadolinium-based contrast agents have been associated with the development of nephrogenic systemic sclerosis. This condition is rare and has been reported to occur in patients with compromised renal function. Gadolinium can cross the placenta into the fetal circulation and, subsequently, is excreted into the amniotic fluid, where the agent can remain for an extended period of time. No prospective studies with large numbers of patients have evaluated the risk of teratogenic or mutagenic effects. Animal studies have demonstrated that high doses of gadolinium administered over long periods of time can have adverse effects.³⁷ The American College of Radiology Guidance Document for Safe MR Practices recommends that a pregnant patient should only receive gadolinium-based contrast agents after careful consideration of the risk–benefit ratio.³⁸

Iodinated CT contrast agent has been associated with contrast-induced nephropathy in as many as 21% of patients who had a baseline glomerular filtration rate of <50 ml/min/1.73 m^{2, 39} Nephropathy induced by iodinated CT contrast agent is usually reversible, but the condition can be associated with nonrenal complications that can prolong hospital stays and increase in-hospital mortality.⁴⁰ Free iodide in the contrast medium given to the mother has the potential to depress fetal and neonatal thyroid function. Neonatal thyroid function should, therefore, be checked after delivery in such patients. The risk associated with absorption of contrast medium during lactation is small and can be considered insufficient to warrant stopping of breastfeeding.⁴¹

Migraine, pre-eclampsia and stroke

Patients with a history of migraine headaches are more likely to develop pre-eclampsia than those without such a history, and differentiating migraine from pre-eclampsia headaches during the third trimester poses a challenge for the managing physician. Pre-eclampsia occurs in 5–8% of gestations and is characterized by hypertension and proteinuria, which can be diagnosed after 20 weeks of gestation or during the puerperium.⁴² Hypertension in this setting is defined as blood pressure 140/90 mmHg on multiple measurements taken at least 4 hours apart. Proteinuria is defined as 24-hour urine protein excretion 300 mg.⁴³ Criteria for severe pre-eclampsia accepted by the American College of Obstetrics and Gynecology include two measurements of systolic blood pressure 160 mmHg or diastolic blood pressure 110 mmHg taken at least 6 hours apart while the patient is on bed rest; proteinuria 5 g in a 24-hour urine specimen or 3+ or greater on two random urine samples collected at least 4 hours apart; and oliguria <500 ml in 24 hours. Other criteria include the presence of cerebral or visual disturbances, pulmonary edema or cyanosis, epigastric or right upper-quadrant pain, impaired liver function, thrombocytopenia and fetal growth restriction.⁴³ For the sake of maternal well-being, delivery is the definitive treatment in the setting of severe pre-eclampsia.

The relationship between headache and pre-eclampsia is becoming more clearly recognized. A case–control study reported that a history of any headache before or during pregnancy is associated with an increased risk of pre-eclampsia.⁴⁴ Indeed, the risk of developing pre-eclampsia in women with migraines during pregnancy is fourfold higher than the risk in pregnant women without migraines.^{26, 44, 45}

Pre-eclampsia is characterized by endothelial dysfunction, platelet activation and inflammation. Of note, angiogenic factors and their corresponding receptors have been discovered to exhibit dysregulated levels in the maternal serum, and can be used to predict pre-eclampsia with high sensitivity and specificity.^{46, 47} Interestingly, certain pathophysiological characteristics of migraine, such as endothelial dysfunction, overlap with those of pre-eclampsia.^{48, 49} In addition, both conditions are associated with cardiovascular risk factors and a higher subsequent risk of ischemic stroke.^{50, 51}

In a case–control study of pregnancy discharges, inpatient coding for peripartum migraine was associated with ischemic stroke.²⁹ The results from this study also indicated associations between peripartum migraine and heart disease, venous thromboembolism, and cardiovascular risk factors, such as hypertension, diabetes and smoking (Box 1).²⁹ In addition, a marked overlap existed between the diagnosis of pre-eclampsia and most vascular diagnoses, such as stroke, heart disease, and venous thromboembolism. Indeed, the pre-eclampsia, eclampsia and gestational hypertension codes—because of their confounding presence—had to be excluded from the records with vascular codes (stroke, heart disease, and venous thromboembolism) to determine the independent effect of the vascular codes with migraine.²⁹ The results from this study suggest that overlaps exist between active migraine during pregnancy, pre-eclampsia, and vascular complications. Furthermore, women who experience active and severe migraine attacks during pregnancy require further study to determine whether migraine and/or vascular prevention measures should be initiated during pregnancy.

Treatment options

Migraine therapy in pregnancy is complicated. The available treatment options (Table 1) cannot always be applied to the pregnant or postpartum patient, because of the potential effects of these compounds on the fetus and (via lactation) the newborn. For example, the use of narcotic medications either as first-line therapies for migraine headaches or at progressively higher doses in patients experiencing recurrent migraines leads to frequent complications. In both cases, rebound headaches can develop as a result of narcotic withdrawal. Such narcotic effects might have contributed to the headache pattern observed in our patient (Box 1).

Table 1 | Medications used for migraines during pregnancy

• Full table
• Figures and tables index
• Download Power Point slide (66 KB)

Preconceptional counseling is important for patients with any medical condition that could affect pregnancy. Many patients with migraine use medications to prevent recurrent migraines. These medications include -blockers, antiepileptic medications, calcium channel blockers and antidepressants. When a patient decides to attempt conception, such medications should be discontinued and alternative forms of migraine prevention should be recommended. If a patient cannot achieve acceptable control of her symptoms without pharmacological therapy, the lowest effective doses of the lowest number of agents considered safe during pregnancy should be used. The specific risks of each medication during pregnancy should be discussed with the patient.⁵²

The US Recommended Dietary Allowance for folate in pregnancy is 600 g per day.⁵³ Patients using antiepileptic drugs, such as valproic acid, for migraine or seizure control are recommended to take high-dose folic acid supplements (4 mg) daily before conception to reduce the risk of neural tube defects and other potential fetal anomalies.⁵⁴ This amount of folic acid is well above the dose of 400 g a day that is generally recommended for pregnant patients. The safety of receiving the 4 mg dose has not been well studied and a lower dose might be preferable.⁵⁵ Topiramate, a drug often used for the prevention of migraine, has been associated with an increased rate of fetal anomalies. The UK Epilepsy and Pregnancy Register, however, reported a malformation rate with topiramate monotherapy of 4.8% (95% CI 1.7–13.3%), which is no higher than the background rate.⁵⁶ The patient should also be counseled about the background risk of any major and/or minor congenital anomaly that reaches 3–6% and about the options for prenatal screening and ultrasound evaluations.

Most women are able to manage headaches during pregnancy with nonpharmacological treatments, including biofeedback-assisted relaxation, hydration, bed rest, and a reduction in work or home responsibilities. The second line of treatment is abortive medication used early in the onset of headache. In patients with milder headaches, biofeedback methods combined with the occasional use of over-the-counter analgesics (for example, acetaminophen) usually suffice and are safe to use during pregnancy. Approximately 25% of women will require additional pharmacological methods for headache pain management during pregnancy.³⁰ The decision to use additional forms of preventive or abortive medications rests on the patient and her physician, and depends on gestational age, as well as on the type and severity of the associated symptoms. FDA guidelines are commonly consulted to determine which medications to use during pregnancy. These guidelines are, however, based on overall risk and benefit, which are different from teratogenic risk, the main concern. Additional concerns include the risk of fetal intrauterine growth restriction, alterations in functional organ maturation, and neonatal withdrawal symptoms that can be associated with medication exposure during pregnancy. Limited correlations exist between risks assigned by the FDA and the risk of teratogenesis, growth restriction or narcotic withdrawal as assessed by other online automated databases. This lack of consensus leads to confusion regarding drug safety during pregnancy.^{57, 58}

Medications used to treat acute migraines include triptans, ergot alkaloids, and NSAIDS. Ergot alkaloids have been widely used to treat migraines in nonpregnant women. Derivatives of ergot, such as ergotamine and dihydroergotamine, are highly nonselective and interact with virtually all monoamine receptors,⁵⁹ leading to multiple effects including increased sympathetic activity and vasoconstriction.⁶⁰ Adverse effects related to ergot derivatives, including generalized arterial vasoconstriction with decreased uterine blood flow and increased uterine muscle contractility, preclude their use during pregnancy. In the past, ergot derivatives have been used as abortifacients. The theoretical risk of congenital anomalies caused by the vasoconstrictive effects of these agents has not been substantiated in large epidemiological studies.⁶¹

Triptans are 5-hydroxytryptamine (5-HT) receptor agonists that are effective in treating acute headache attacks. Sumatriptan, which was shown to activate 5-HT_1B, 5-HT_1D and 5-HT_1F receptors, was the first and has been the most widely used of the triptans.⁶² These drugs have vasoconstrictive effects,⁶³ but also act as presynaptic neuroinhibitors within the CNS, where they suppress the release of pain-generating neuropeptides within the neurovascular unit.⁶⁴

The FDA classifies all triptans as pregnancy category C drugs, which implies that a risk of teratogenicity in humans has not been ruled out. Available data, however, support the use of such drugs during pregnancy, with little evidence to suggest that triptans are teratogenic or alter uteroplacental blood flow.^{65, 66, 67} Nevertheless, the possibility remains that triptan use is associated with a small increase in the risk of very rare birth defects.³⁰ Population studies and registries have revealed no differences in the frequency of birth defects between control populations and patients using sumatriptan.⁶⁸ One study, however, found a small increase in the frequency of esophageal atresia, although the overall incidence was no higher than in the general population.⁶⁹

NSAIDs are frequently used for the management of headache. Indomethacin has been shown to be particularly useful for patients with migraine who experience nausea, as the drug can be given rectally. NSAIDs can also be combined with acetaminophen and caffeine for effective pain control. Frequent use of NSAIDs, as with most medications used for acute headache management, can lead to rebound headaches. During pregnancy, NSAIDs can only be used up to 32 weeks' gestation and in short courses. Extended use can cause closure of the ductus arteriosus and persistent pulmonary hypertension in the fetus and the neonate. Evidence of ductal constriction was found on fetal echocardiography in 4 out of 54 (7%) fetuses of women who were treated with ibuprofen for premature labor. Oligohydramnios and neonatal renal failure have been reported in pregnant women who have received prolonged NSAID therapy for the purpose of inhibiting premature labor during the second or third trimester.⁷⁰

Intravenously delivered ondansetron, metoclopramide, chlorpromazine or prochlorperazine can each be used alone or in combination with NSAIDs or opioids in the treatment of acute migraine headache. The efficacies of all of these antiemetics for migraine have been demonstrated in randomized placebo-controlled trials.⁷¹ Antiemetics are not, in general, associated with an increased teratogenic risk in humans, although the phenothiazines have been associated with an increased risk of intrahepatic cholestasis during pregnancy.⁷²

Benzodiazepines, opioids and barbiturates are also options for the treatment of migraine. Agents from all three classes should not be used chronically, however, because they are associated with dependence, can contribute to the development of rebound and chronic daily headaches in the mother, and are associated with a withdrawal syndrome in the neonate. The severity of the neonatal withdrawal is directly attributable to the frequency and dose of the drug used. Chronic headache related to medication overuse is most often observed with the use of narcotics, although any medication used for acute migraine relief can cause an analgesic rebound headache. To circumvent the problems of opioid overuse and rebound headaches, a prophylactic strategy for migraine headache management should be adopted. Furthermore, available practice guidelines recommend non-opioid medications as the first-line therapies for severe migraine.³²

Daily prevention of migraines should be considered in women who have at least three or four severe migraine attacks per month throughout pregnancy. Therapeutic options include -blockers, calcium channel blockers, and selective serotonin reuptake inhibitors (SSRIs; Table 1). Use of the -blocker propranolol has been associated with increased risk of intrauterine growth restriction during pregnancy. Metoprolol and labetalol have been used as alternatives to propranolol during pregnancy; however, with all of these agents, close ultrasound surveillance of fetal growth is necessary, particularly throughout the latter half of gestation.

The use of calcium channel blockers can be an effective strategy for the management of migraines during pregnancy, as well as to treat hypertensive disorders during pregnancy (Box 1). Short-acting forms of these drugs can, however, lead to a sudden drop in maternal blood pressure and decreased uteroplacental perfusion, especially among hypertensive patients.

The SSRIs, especially paroxetine, have been associated with an increased risk of congenital heart defects and neonatal pulmonary hypertension. The SSRIs have all been reported to be associated with some degree of neonatal withdrawal syndrome after birth.⁷³ Most of the literature concerning the association between SSRIs and congenital anomalies is, however, retrospective and inconsistent.