Neurologic Disorders

Anticonvulsants and Contraception

Women with epilepsy should be advised about possible interactions between anticonvulsant drugs and oral contraceptive agents. Certain anticonvulsants, including carbamazepine, clobazam, eslicarbazepine, felbamate, oxcarbazepine, perampanel, phenobarbital, phenytoin, primidone, topiramate, and rufinamide, may interfere with the effectiveness of oral contraceptives and implanted progestins, leading to unwanted pregnancy. The possibility of contraceptive failure must be discussed with women taking these anticonvulsants and documented in the records. Regular counseling of women with epilepsy is necessary during the reproductive years because unplanned pregnancy may occur. Brivaracetam, cannabidiol, cenobamate, clonazepam, clorazepate, ethosuxamide, gabapentin, lacosamide, lamotrigine, levetiracetam, pregabalin, tiagabine, valproic acid, vigabatrin, and zonisamide have not been reported to cause contraceptive failure. When oral contraception is desired for women taking enzyme-inducing anticonvulsants, a formulation that includes at least 50 μg of ethinyl estradiol or mestranol is preferred, but the best way to ensure contraception is to use an alternative method. Conversely, combined oral contraceptives may increase the metabolism of lamotrigine, causing lower lamotrigine blood levels, and may increase seizure frequency in patients taking this medication.

Effect of Pregnancy on Seizure Disorders

Pregnancy has unpredictable and variable influences on epilepsy. When seizure frequency increases, it most commonly does so in the first trimester and usually reverts to the pregestational pattern at the conclusion of the pregnancy, although a few patients experience a permanent deterioration in seizure control. In general, control in patients with frequent seizures (i.e., more than one a month) before pregnancy is likely to deteriorate during the gestational period, whereas only a small minority of patients with infrequent attacks (i.e., less than one every 9 months) experience an exacerbation during pregnancies.

Several case series have examined changes in seizure frequency during pregnancy. Seizure frequency remained unchanged in 54% to 80% of patients, with the highest rate of stability in patients with documented medication compliance. Seizure frequency increased in 14% to 32% of patients and decreased in 3% to 24%. Seizures are more likely to be exacerbated during pregnancy in women with more frequent seizures before the pregnancy, in patients with focal seizure disorders, and in those prescribed multiple anticonvulsants. For patients who have been seizure-free for at least 9 months before conception, the likelihood of remaining seizure-free during pregnancy is 84% to 92%. Among women with catamenial epilepsy, in whom seizures cluster during certain phases of the menstrual cycle, 44% experience a >50% reduction in seizure frequency during pregnancy.

It is usually not possible to predict the outcome in individual cases, regardless of the maternal age and the outcome of previous pregnancies. The best predictor of seizures during pregnancy is the frequency of seizures prior to pregnancy; however, attacks may occur during pregnancy in patients who have been seizure-free for several years.

Epilepsy may appear for the first time during or immediately after pregnancy. It is uncommon for patients in the latter group to have seizures only in relationship to pregnancy and at no other time (i.e., gestational epilepsy). Some patients with true gestational epilepsy experience recurrent seizures during pregnancy, and the remainder have only a single convulsion. The occurrence of seizures in one pregnancy is no guide to the course of subsequent pregnancies. Seizures that occur during pregnancy do not differ clinically from seizures occurring in other circumstances. Improved compliance with an anticonvulsant drug regimen may sometimes account for the reduction in seizure frequency that occasionally occurs during pregnancy in women with epilepsy.

The increase in seizure frequency that occurs in some epileptic patients during pregnancy may be related to the metabolic, hormonal, or hematologic changes of the gestational period or to fatigue or sleep deprivation. A rapid and excessive gain in weight sometimes occurs before an increase in seizure frequency, providing some support for the belief that fluid retention may occasionally be a factor, perhaps by a dilutional effect on anticonvulsant drug concentration. It is tempting to relate any change in seizure frequency to hormonal factors because estrogens are epileptogenic in animals and progesterone has both convulsant and anticonvulsant properties. Nausea, vomiting, reduced gastric motility, or use of antacids may also lead to reduced absorption of anticonvulsant drugs.

There is sometimes difficulty in maintaining adequate treatment with anticonvulsant drugs during pregnancy. Serum levels of lamotrigine, levetiracetam, oxcarbazepine, phenobarbital, phenytoin, topiramate, and zonisamide generally decline in pregnancy and increase in the postpartum period. ^, An increase in dosage is frequently required to maintain plasma levels at prepregnancy values; monitoring of drug levels before planned conception and during pregnancy should be considered, remembering that the free level, rather than the total level, of the drug correlates best with therapeutic efficacy. The effect of pregnancy on the pharmacokinetics of other agents is less clear.

The reason for the changes in drug requirements is unknown. One possibility is the dilutional effect of increasing plasma volume and extracellular fluid volume. Poor compliance with the anticonvulsant drug regimen, perhaps because of nausea and vomiting or concerns about the effect of medication on the fetus, may also be an important contributory factor, as may decreased plasma protein binding and changes in the absorption and excretion of drugs. The increased metabolic capacity of the maternal liver in pregnancy and possible fetal or placental metabolism of part of the anticonvulsant dose may influence the changes in anticonvulsant drug requirements that occur in women with epilepsy during pregnancy.

After pregnancy, clearance of anticonvulsants will return to pregestational rates. The time this takes depends on the mechanism of clearance; renal clearance and glucuronidation normalizes over 2 to 3 weeks, whereas some hepatic enzymes such as cytochrome P450 may take 1 to 2 months.

Folic acid therapy may lower the plasma phenytoin level, sometimes to below the therapeutic range, and other drugs taken concomitantly with an anticonvulsant medication also may lead to reduced plasma levels of the anticonvulsant. Antacids and antihistamines warrant particular mention because these agents are commonly taken during pregnancy.

Status epilepticus sometimes complicates pregnancy and may occur without any preceding increase in seizure frequency, occasionally because of the injudicious discontinuation of anticonvulsant drugs. This is a rare occurrence, but it may lead to a fatal outcome for the mother or fetus. The absence of hypertension, proteinuria, and edema helps in distinguishing this condition from eclamptic convulsions. Serum glucose, sodium, calcium, and magnesium levels should be checked and abnormalities corrected immediately. As in nonpregnant patients, it is essential to obtain control of the seizures as rapidly as possible, but the formerly accepted practice of terminating pregnancy is usually unnecessary.

Status epilepticus is treated with anticonvulsant drug therapy, with the pregnancy being allowed to continue to term. Intravenous lorazepam (4 mg given at a rate of 2 mg/min) or intramuscular midazolam (10 mg) is the initial treatment of choice. Either dose may be repeated once after 10 minutes if seizures continue. Fosphenytoin sodium (20 mg/kg phenytoin equivalents) or phenytoin (20 mg/kg) is then given intravenously regardless of the response to benzodiazepines to prevent seizure recurrence. An additional 10 mg/kg may be given if seizures persist. Fosphenytoin sodium is water soluble, may be infused with dextrose or saline, is better tolerated at the infusion site than phenytoin, and may be infused three times more rapidly than intravenous phenytoin up to a rate of 150 mg of phenytoin equivalents per minute. It is converted in the body to phenytoin, which may be cardiotoxic, and cardiac monitoring is required during both fosphenytoin and phenytoin infusions. Alternatively, intravenous levetiracetam (60 mg/kg, maximum dose of 4500 mg) may be given over 15 minutes. Valproate is also effective but is best avoided during pregnancy due to the risk of fetal malformations associated with its use. If seizures continue despite the fosphenytoin or levetiracetam load, next steps include 20 mg/kg intravenous phenobarbital or general anesthesia with midazolam or propofol. It is essential to maintain control of the airway throughout treatment. Diagnosis and management of eclampsia and eclamptic seizures is discussed in Chapter 45 .

Effect of Epilepsy on Pregnancy and Lactation

Only a few studies have attempted to document the effect of epilepsy on pregnancy. The results are often difficult to evaluate because of the limited number of cases reported; the lack of comparative data on women without epilepsy attending the same institutions; differences in the severity of the epilepsy and how it has been treated; differences in age, medical background, and socioeconomic status of the patients reported; and the lack of information concerning relevant social habits such as cigarette smoking and alcohol ingestion.

These methodologic challenges notwithstanding, a large meta-analysis demonstrated less than twofold increased odds of spontaneous miscarriage, antepartum hemorrhage, postpartum hemorrhage, hypertensive disorders, induction of labor, cesarean delivery, preterm birth, and fetal growth restriction among women with epilepsy. Some evidence also suggests that smoking increases the risk of preterm labor in women with epilepsy compared with women without epilepsy. Cesarean delivery is not indicated simply because of maternal epilepsy except when seizures occur frequently or during labor, when seizures are precipitated by physical activity, or when patients cannot cooperate during labor because of their neurologic disorder or mental abnormality. Fetal death can result from maternal seizures, presumably because of the accompanying hypoxia and acidosis, but evidence suggests that there are no more stillbirths among epileptic patients than in the general population. The effect of maternal seizures on placental blood flow is not established, but changes in fetal heart rate suggestive of hypoxia have been described; they may relate to reduced placental blood flow or to metabolic changes in the mother.

Anticonvulsant drugs taken by the mother may be present in breast milk, but data as to whether they have any major effect on the infant are limited ( Table 66.1 ). When an infant develops sedation that is likely to be related to antiepileptic drugs in breast milk, breastfeeding should be discontinued, and the infant should be observed for signs of drug withdrawal. However, breastfeeding does not need to be discouraged for reasons related to the milk’s content of anticonvulsant medication, and current guidelines do not recommend changing the treatment regimen to a drug that does not penetrate breast milk.

Effect of Maternal Epilepsy and Anticonvulsant Drugs on the Fetus and Neonate

A woman with epilepsy who becomes pregnant is usually concerned that her unborn child may inherit a similar susceptibility to seizures. The risk of epilepsy in the child depends on the nature of the mother’s seizure disorder, and it is higher in idiopathic than acquired maternal epilepsy. Although precise quantification of the risk is not possible, it is probably quite small. The cause of this increased risk to the offspring of epileptic mothers is unknown. It may relate to genetic factors, seizures arising during pregnancy, or the metabolic and toxic consequences of seizures or anticonvulsant drugs. In general, pregnancy in epileptic women does not need to be discouraged on these grounds, but reassurance and support are necessary.

A major problem in management of epileptic patients during pregnancy is the possibility that certain anticonvulsant drugs may induce fetal abnormalities. However, epilepsy has a relatively low prevalence rate, can occur for a multitude of reasons, can vary markedly in severity, can be treated by a variety of drugs singly or in combination, and can itself be associated with an increased risk of fetal malformations. Some patients may have a common genetic predisposition to seizures and to fetal malformation. Environmental factors may be important in the genesis of congenital abnormalities, and socioeconomic backgrounds must be matched as much as possible when comparisons are made of the incidence of malformations in different patient populations.

The absolute risk of major congenital malformations based on large registries of pregnant women with epilepsy is 0.7% to 2.8% for levetiracetam, 1.9% to 2.9% for lamotrigine, 2.2% to 3.0% for oxcarbazepine, 2.6% to 5.5% for carbamazepine, 2.9% to 6.4% for phenytoin, 3.9% to 4.3% for topiramate, 5.5% to 6.5% for phenobarbital, and 6.7% to 10.3% for valproate. Based on systematic reviews, exposure to carbamazepine, phenytoin, topiramate, phenobarbital, or valproate during pregnancy carries a higher odds of major congenital malformations compared to pregnancies in women without epilepsy, whereas no increased risk was observed for lamotrigine and levetiracetam. Mothers taking more than one anticonvulsant have a higher risk of bearing children with major congenital malformations, but this appears to be based on the increased likelihood of exposure to specific teratogenic anticonvulsants in this population than combination therapy itself. Valproic acid has an especially high (1%) rate of neural tube defects including spina bifida; it may also cause cleft lip or palate, polydactyly, hypospadias, craniosynostosis, delayed development, and disorders of the cardiovascular and endocrine systems. The risk is dose dependent. Some data suggest that phenytoin and carbamazepine are associated with cleft palate, that phenobarbital may be associated with cardiac malformations, and that topiramate is associated with infants who are small for gestational age and with microcephaly, oral clefts, and hypospadias.

Animal studies lend support to the belief that some anticonvulsants are teratogenic. The mechanism involved is unclear but may include folate deficiency or antagonism. Low blood folate levels before or early in pregnancy are significantly associated with spontaneous abortion and the occurrence of developmental anomalies. It has also been suggested that certain oxidative intermediary metabolites of anticonvulsants may affect cell division and migration.

Although most children born to epileptic mothers are cognitively normal, prenatal antiepileptic drug exposure may be associated with developmental delay. In particular, exposure to valproic acid appears to carry a dose-dependent risk of impaired cognitive outcomes. In one study, children exposed to valproic acid in utero scored an average of 8 to 11 points lower on an IQ test at 6 years of age than children exposed in utero to carbamazepine, lamotrigine, or phenytoin, which do not appear to be associated with lower cognitive outcomes. Periconceptional exposure to folate was associated with higher IQs in this study regardless of anticonvulsant type.

Maternal use of phenytoin during pregnancy has been associated with the fetal hydantoin syndrome, which is characterized by prenatal and postnatal growth deficiency, microcephaly, dysmorphic facies, distal phalangeal hypoplasia, and mental deficiency. Among infants exposed to phenytoin in utero, 11% have enough clinical features to be classified as having this syndrome, and almost three times as many may show lesser degrees of impairment of performance or morphogenesis. The syndrome is similar to that ascribed to phenobarbital and carbamazepine, and it resembles fetal alcohol syndrome. A consistent facial phenotype has also been reported in children exposed to valproic acid or sodium valproate in utero.

Maternal use of barbiturates (60 to 120 mg daily) in late pregnancy may be associated with neonatal withdrawal symptoms beginning a week after birth. Withdrawal symptoms include restlessness, constant crying, irritability, tremulousness, difficulty in sleeping, and vasomotor instability, but not seizures.

Clinical or subclinical coagulopathy may occur in a neonate whose mother received anticonvulsants during pregnancy. In affected infants, levels of factors II, VII, IX, and X are decreased, and levels of factors V and VIII and fibrinogen are normal. The abnormalities are similar to those produced by vitamin K deficiency. As a result, some experts advocate maternal ingestion of vitamin K ₁ (100 μg daily) during the last month of pregnancy. However, it is unclear whether routine prophylaxis in this manner is justifiable, as some studies suggest that such hemorrhagic complications are rare and even call into question whether maternal antiepileptic use increases the risk of hemorrhage in neonates at all. The routine practice of administering 1 mg of vitamin K to all neonates is likely to be sufficient to mitigate any risk of hemorrhagic complications that could result from maternal antiepileptic use.

General Therapeutic Approaches

It is difficult to make more than general therapeutic recommendations about pregnancy in a woman with epilepsy. Epilepsy should be treated with the smallest effective dosage of an anticonvulsant drug, and monotherapy is preferable to polytherapy. Drug selection is based on seizure type, clinical status, and maternal and fetal risks. Switching from valproate to an alternative monotherapy should occur before conception, and a similar approach could be considered for women taking topiramate or phenobarbital. Folate supplementation (4 mg daily) is usually provided. Because many pregnancies are unintended and congenital malformations may have already occurred by the time a woman realizes she is pregnant, providing folate supplementation for any woman of childbearing age taking antiepileptic medications can be considered. Similar reasoning suggests that it may be advisable to avoid valproate in epileptic women of childbearing age.

Prenatal counseling is important. If a nonpregnant epileptic woman asks about pregnancy, it is appropriate to inform her that there is a small risk of having a malformed child because of the seizure disorder or the drugs used in its treatment. This risk is probably about double that for a patient without epilepsy, but there is still a more than 90% chance that she will have a normal child.

Data concerning the relative safety and therapeutic effectiveness of different anticonvulsant drugs in the management of pregnant patients with epilepsy are insufficient to guide the physician responsible for the care of these patients. It seems clear, however, that trimethadione should not be used and that valproic acid should be avoided. If valproic acid must be used, prenatal testing for maternal serum alpha fetoprotein levels or with ultrasound is advisable to detect neural tube defects so that therapeutic abortion can be considered if necessary. Substitution of one anticonvulsant drug for another in epileptic women whose first medical visit is after the first trimester should be avoided because any major malformation of the fetus has probably occurred already.

The principles of drug management of a seizure disorder in a pregnant woman are the same as those for a nonpregnant woman. Anticonvulsant drugs are as necessary to epileptic patients during pregnancy as at other times. A detailed account of the drugs used in the treatment of epilepsy is unnecessary here, but several points are worthy of comment.

A solitary seizure, unrelated to toxemia, should not lead to a diagnosis of epilepsy because there may be no further attacks. Only time will tell whether an individual who has a single seizure is going to have further attacks, thereby justifying a diagnosis of epilepsy and necessitating prophylactic anticonvulsant drug treatment.

Although some physicians start a patient on anticonvulsant medication after one convulsion, others prefer to withhold medication until the patient has had at least two seizures, at least in the nonpregnant state. During pregnancy, many physicians initiate anticonvulsant therapy after a single seizure and arrange for neurologic reevaluation after delivery. This approach warrants emphasis because many patients with so-called gestational epilepsy have only a single convulsion, and continued treatment in such circumstances may be unnecessary. Simple medical and neurologic investigations are indicated in an adult who has an isolated seizure and is otherwise well with no neurologic signs: hematologic and biochemical screening tests, electroencephalogram, and magnetic resonance imaging (MRI) of the head. If the findings of such investigations are unremarkable, discuss the controversial issue of anticonvulsant drug treatment with the patient but generally recommend that treatment be withheld unless a future attack occurs.

In pregnant women experiencing two or more seizures, prophylactic anticonvulsant drug treatment is indicated. In patients with a progressive history, abnormal neurologic signs, or a focal electroencephalographic abnormality, it is necessary to exclude an underlying structural lesion with MRI of the head.

If prophylactic anticonvulsant drug treatment is necessary, it is generally continued until the patient has been seizure-free for at least 2 or 3 years. Treatment is started with a small dosage of one of the anticonvulsants, depending on the type of seizure experienced by the patient and the considerations outlined earlier. The dosage is increased until seizures are controlled, blood concentrations reach the upper end of the optimal therapeutic range, or side effects limit further increments. If seizures continue despite optimal blood levels of the anticonvulsant drug selected, a second drug should be substituted for the first. Patients often respond better to one or another of the various drugs that are available.

Patients must take medication as prescribed, and treatment should be controlled by frequent monitoring of the plasma concentration of the anticonvulsant drug. Monthly follow-up visits during pregnancy usually permit satisfactory supervision of the patient. At the initial visit, trough values of total and free concentrations of drugs known to have altered pharmacokinetics during pregnancy (see above) should be measured. Total levels should then be measured each month in patients whose seizures are well controlled; free levels should be monitored monthly in patients with poor seizure control, seizures during pregnancy, or a marked (>50%) decline in total level. Poor compliance with an anticonvulsant drug regimen can often be improved by encouragement and by explaining the importance of taking medication regularly. Simplifying the dosage schedule so that medication is taken just once or twice daily may be helpful.

As the pregnancy continues, the dosage of the anticonvulsant drug may need to be increased if seizures become more common or the free level of the anticonvulsant drug declines by more than about 30%. In some instances, the required dosage may reach a level that would probably cause toxic side effects in a nonpregnant patient. If the anticonvulsant dosage is increased during the pregnancy, reductions will probably be necessary in the puerperium to prevent toxicity, but this change must be based on clinical evaluation and measurement of the plasma concentration of the drug because the period over which drug requirements decline varies considerably. For anticonvulsants cleared renally or by glucuronidation, such as lamotrigine, a taper over 2 to 3 weeks after delivery to pregestational levels will usually maintain seizure control and avoid toxicity. Drug levels should be checked when steady state has been reached. Because of the poorly defined risks of increased obstetric complications among pregnant epileptic women, close supervision of these patients by the obstetrician is mandatory, and delivery in a hospital is advised.

After delivery, the infant must be inspected for congenital malformations and given an injection of vitamin K ₁ (1 mg intramuscularly). Clotting factors may be studied after about 4 hours, and further injections of vitamin K ₁ can be given if necessary. If hemorrhage occurs, infusions of fresh-frozen plasma or of factors II, VII, IX, and X may also be necessary. Breastfeeding of a healthy infant by an epileptic mother should not be discouraged. The effect of enzyme-inducing anticonvulsants on oral contraceptives and implantable progestins should be discussed.

Summary and Recommendations

• The risk of untreated epilepsy to the fetus is greater than the risk of maternal anticonvulsant use.

• Patients should be counseled on interactions between anticonvulsants and oral contraceptives.

• Valproate should be avoided in women with childbearing potential.

• Women taking anticonvulsants should be given folic acid supplementation before and during pregnancy.

• Serum anticonvulsant levels should be monitored before and regularly during pregnancy.

Migraine

Migraine is an important cause of headache among women of childbearing age. Women with migraine outnumber men by a factor of two or three. Migraine with aura, in which episodic headache is preceded by visual, sensory, or motor symptoms, affects 20% of individuals who experience migraine. Typical auras are visual, with photopsias (flashes of light), scintillations (flickering lights), scintillating scotoma (areas of visual loss surrounded by flickering light), or fortification spectra (jagged, bright zigzags of light) most commonly described. Other focal neurologic symptoms preceding or accompanying the headache may include tingling, numbness, weakness, or impairment of consciousness. The remaining 80% of migraineurs do not experience premonitory focal symptoms. Headaches may be lateralized or generalized, usually have a gradual onset, and usually last for less than 1 day, although they may persist for a longer time. They may be dull or throbbing; are commonly accompanied by nausea, vomiting, and photophobia; and are often associated with blurring of vision, lightheadedness, and scalp tenderness. Positive visual and sensory phenomena are important diagnostic clues for migraine, but in patients with purely negative symptoms (e.g., numbness, weakness, visual field deficit, aphasia), structural or vascular brain disease should be considered.

Many women with migraine link the periodicity of some of their attacks to the menstrual cycle, with headaches occurring usually just before or during menstruation. Migrainous attacks without aura are most likely to be related to the menstrual cycle. Some patients may have headaches that occur only in relation to the menstrual cycle, although this pattern is much less common. The manner in which hormonal factors provoke migraine remains unclear, although some evidence suggests that the decrease in estrogen that precedes menstruation is a trigger.

Migraine headaches are commonly exacerbated in women using oral contraceptives, but improvement can occur in some patients. Such exacerbation usually becomes apparent within the first few months of oral contraceptive use. Preparations with a relatively high estrogen content are most likely to influence the headache pattern and are generally not as well tolerated as low-estrogen preparations. Migraines are more likely to occur during the hormone-free week, and in such cases eliminating the hormone-free interval through the use of continuous combined hormonal contraceptives is an effective strategy. Because migraine with aura and combined oral contraceptives are independent risk factors for stroke, with the combination of both risk factors resulting in a sixfold increased odds of stroke, the American College of Obstetricians and Gynecologists (ACOG) advocates avoiding combined oral contraceptives in patients with migraine with aura; progestin-only pills are considered a safe option. Combined oral contraceptives are probably safe in patients with migraine without aura, especially if a preparation containing less than 25 μg of ethinyl estradiol is used.

Migraine often improves considerably after the first trimester of pregnancy, regardless of whether the attacks are related to the menstrual cycle. It occasionally worsens or occurs for the first time during pregnancy, most commonly during the first 3 months of gestation. The response of migraine to pregnancy does not correlate with sex of the fetus or with differences in plasma levels of hormones.

Management of migraine consists of the avoidance of precipitating factors coupled with prophylactic or abortive drug treatment, if necessary. In general nonpregnant populations, when simple analgesics do not abort the headache, treatment with extracranial vasoconstrictors (e.g., ergotamine, dihydroergotamine), serotonin agonists (e.g., sumatriptan), phenothiazines (e.g., prochlorperazine), or other drugs may be necessary. For migraineurs experiencing more than three headaches per month, prophylactic treatment with β-blockers (e.g., propranolol), calcium channel blockers (e.g., verapamil), or tricyclic antidepressants (e.g., amitriptyline) is usually offered. Although menstrual migraine should be treated with the same approach as nonmenstrual migraine, the predictable timing of the headaches allows other strategies, such as a short course of twice-daily triptans or 1.5 mg transdermal estradiol gel daily during the menstrual week.

During pregnancy, medication is best avoided if possible. Dietary and other precipitants of headache should be avoided. Biofeedback and relaxation should be attempted, and the patient should be reassured that most women experience a decrease in headache frequency as pregnancy proceeds. When drugs are required, simple analgesics should be used. Acetaminophen is preferred over aspirin and other nonsteroidal antiinflammatory drugs because the latter are weakly associated with miscarriage in the first trimester and premature ductus closure in the third trimester. Similarly, caffeine is an effective migraine abortive that is probably safe in pregnancy, although it has also been associated with increased rates of miscarriage in the first trimester. Metoclopramide is an effective migraine abortive that is probably safe in pregnancy. Triptans can be considered if these approaches are not effective, given data from meta-analysis demonstrating no increase in premature labor, spontaneous abortions, or teratogenicity. As a last resort, opiates can be considered, although these are associated with rebound headaches and medication-overuse headache. Ergotamine-containing preparations should be avoided because of the potent contractile effect this drug may have on the gravid uterus and its potential teratogenicity. If a preventive agent is necessary due to headaches occurring more than three times per month, propranolol, verapamil, or tricyclic antidepressants can be considered. Several devices are effective in aborting migraine, including single-pulse transcranial magnetic stimulation (contraindicated in patients with epilepsy), noninvasive vagus nerve stimulation, transcutaneous trigeminal nerve stimulation, and remote electrical stimulation applied to the upper arm, but their safety in pregnancy is not known. Transcutaneous supraorbital neurostimulation is also an effective preventive therapy, and both acupuncture and botulinum toxin A are effective in patients with more than 15 headache days per month. The data regarding the safety on the use of botulinum toxin in pregnancy and during lactation are limited. A prospective study of 32 pregnant women with refractory chronic migraines treated with botulinum toxin A in pregnancy did not observe any adverse impact of the drug on pregnancy outcomes. Although there were no congenital malformations and all of the women delivered healthy babies with normal birth weights at term, the authors cautioned that the numbers are too small to draw conclusions regarding safety in pregnancy. Because botulinum toxin’s safety during pregnancy and lactation is not known, its use for nondisabling symptoms is not recommended in these settings. In patients with frequent headaches, comorbidities such as depression require special consideration.

Women with migraine during pregnancy may have a higher prevalence of preterm delivery. This did not appear to influence delivery outcome.

Postnatal Headache

About one-third of women experience headaches in the week after delivery, and most of them have a personal or family history of migraine. The headaches, which are usually mild and bifrontal, respond well to simple analgesics and are self-limited.

Summary and Recommendations

• Migraine frequency and character often change during pregnancy.

• Intracranial venous sinus thrombosis should be considered in patients with new, severe, unremitting headaches during pregnancy.

• Combined oral contraceptives should be avoided in women with migraine with aura.

Arterial Occlusive Disease

Arterial disease is not unusual, even in the absence of diabetes or severe hypertension, in women of childbearing age. Major arterial occlusion accounts for most cases of nonhemorrhagic hemiplegia that develop during pregnancy or the puerperium. Numerous cases of occlusion of the middle cerebral artery or one of the other major intracranial arteries have been described during pregnancy, with occlusion usually occurring in the third trimester or the postpartum period. Stroke is usually caused by the development of a thrombus on a preexisting atheromatous plaque. Predisposing factors may include age, anemia, hormonal influences, hypertension, diabetes, smoking, migraine headaches, increased platelet aggregation, reduced tissue plasminogen activity, changes in blood coagulation factors (especially factors V, VII, VIII, IX, X, and XII and fibrinogen) during late pregnancy, preeclamptic toxemia with hypertension, and puerperal septicemia. Other causes of stroke in young women include carotid or vertebral artery dissection; protein C, protein S, and antithrombin III deficiencies; hyperhomocysteinemia; arteritis; meningovascular syphilis; sickle cell disease; antiphospholipid antibodies; polycythemia and other hematologic disorders; prosthetic cardiac valvular disease; and cardiomyopathy.

An embolus resulting from rheumatic or ischemic heart disease, subacute bacterial endocarditis, or a cardiac myxoma may occur. Rare instances of arterial occlusion by paradoxical embolization from a pelvic vein through a patent foramen ovale have also been described. Rarely, a fat, air, or amniotic fluid embolism may occur in relation to childbirth and may have a fatal outcome. Hypotension as a consequence of hemorrhage or related to anesthesia during labor may lead to watershed cerebral infarction.

Transient cerebral ischemic attacks may precede occlusion of one of the major intracranial arteries. The neurologic disorder and the underlying arterial disease must be investigated and treated as in nonpregnant patients. Investigations should include complete blood cell count, serum glucose and electrolytes, hemoglobin A _1c , fasting serum cholesterol and triglyceride levels, prothrombin and partial thromboplastin times, electrocardiography, echocardiography, and radiologic procedures. CT is an important means of excluding intracranial hemorrhage, although MRI may be preferable if it can be obtained in a timely manner to avoid radiation exposure to the fetus. Carotid ultrasound is necessary to rule out carotid stenosis in anterior circulation strokes. CT and magnetic resonance angiography and venography are helpful to diagnose arterial occlusion or venous sinus thrombosis.

Because treatment of acute ischemic stroke is time sensitive, patients must be managed with utmost urgency, and protocols are usually in place in most hospitals to triage stroke patients efficiently. Pregnancy is not an absolute contraindication to thrombolytic agents—tissue plasminogen activator has been used safely and effectively in pregnant patients—although they may increase the risk of miscarriage and may be more dangerous in some pregnancies than others (e.g., placental abruption at 36 weeks versus normal pregnancy at 10 weeks). Intravenous tissue plasminogen activator is approved for use within 3 hours of stroke onset in the United States and within 4.5 hours in Europe, although off-label use between 3 and 4.5 hours in the United States is common. Mechanical embolectomy greatly improves outcomes in patients with stroke resulting from an accessible arterial occlusion when performed within 6 hours of onset and is a viable option for pregnant patients. CT angiography is necessary to determine eligibility for mechanical embolectomy and must be performed urgently in patients with suspected large-vessel occlusion. Select patients with a large ischemic penumbra identified by CT perfusion who present between 6 and 24 hours of onset also benefit from thrombectomy.

Aspirin is indicated within 48 hours of ischemic stroke or transient ischemic attack. Anticoagulation is reserved for patients with atrial fibrillation, mechanical heart valves, certain hypercoagulable states, or venous occlusive disease. Warfarin is best avoided, if possible, because it crosses the placenta and increases hemorrhagic complications and because of the risks for teratogenicity and fetal wastage, especially during the first trimester. However, in select patients with mechanical heart valves, warfarin may be used throughout pregnancy or in the second and early third trimesters due to increased rates of thromboembolism with the use of low-molecular-weight heparin throughout pregnancy in this population. Direct oral anticoagulants should also be avoided because they, too, cross the placenta and may cause fetal complications. Patients requiring anticoagulation during pregnancy are maintained instead on subcutaneously administered low-molecular-weight heparin, which is usually discontinued when labor begins and resumed about 12 hours after vaginal delivery or 24 hours after cesarean delivery. Careful monitoring of anti-Xa activity is required. Carotid endarterectomy should be performed for severe carotid stenosis (70% to 99%) and can be considered for moderate stenosis (50% to 69%), although the benefit in women with moderate stenosis is unclear. With regard to subsequent obstetric management, vaginal delivery, unless specifically contraindicated, is preferable to cesarean delivery. Other diseases that may be associated with arterial occlusive disease in pregnancy (e.g., eclampsia, thrombotic thrombocytopenic purpura) are discussed in Chapter 53 .

A noninflammatory cerebral angiopathy may complicate an otherwise normal pregnancy or the postpartum period ( Fig. 66.3 ). The presentation typically involves a thunderclap headache, and the vasculopathy can be complicated by infarction and intracranial or subarachnoid hemorrhage. It is sometimes associated with hypertension or the use of vasoactive drugs, and although it may simulate vasculitis angiographically, it is treated by removing vasoactive agents as opposed to by immunosuppression.

Intracranial Venous Occlusive Disease

Intracranial venous occlusive disease is an uncommon complication of pregnancy and childbirth. When the thrombosis occurs in the first trimester, it usually follows a complication such as spontaneous abortion, therapeutic abortion, or missed abortion, but it may occur in an otherwise normal pregnancy. Intracranial venous thrombosis is more likely in the third trimester or in the puerperium and is sometimes related to preeclampsia.

Intracranial venous thrombosis is characterized clinically by headache, weakness, focal or generalized convulsions, drowsiness, and confusion. Disturbances of speech, sensation, or vision may also occur, and patients may have mild pyrexia. There may be signs of meningeal irritation resulting from subarachnoid bleeding caused by cortical infarction, and fluctuating hypertension is sometimes found. Papilledema may be present, particularly if the superior sagittal sinus is involved. Cerebrospinal fluid pressure may be increased, and the protein or cell content is often elevated; occasionally, the fluid is stained with frank blood. The diagnosis may be confirmed by CT and magnetic resonance angiography, which are also necessary to exclude arterial pathology and vascular malformation ( Figs. 66.4 and 66.5 ). The symptoms and signs of intracranial venous thrombosis are sometimes mistakenly ascribed to eclampsia, but the absence of previous signs of preeclampsia should help prevent diagnostic confusion.

The prognosis is grave, with a 15% rate of death and dependency based on the pooled results of multiple cohort studies. However, in some series, 80% of patients have an excellent recovery, although lasting neuropsychiatric deficits may have been underreported. Intracranial venous thrombosis is not a contraindication to future pregnancy; the risk of recurrence is probably about 1%. Women with an intracranial venous thrombosis have a higher risk for early spontaneous abortion with subsequent pregnancies.

The etiologic basis of aseptic intracranial venous thrombosis is uncertain; coagulation abnormalities, changes in the constituents of the peripheral blood, and intimal damage to the dural sinuses have been suggested as causes. Inherited prothrombotic states that have been associated with intracranial venous thrombosis include protein C, protein S, and antithrombin III deficiencies; factor V Leiden deficiency; and hyperhomocysteinemia. It has also been linked to COVID-19 adenoviral vaccine-induced thrombotic thrombocytopenia. In addition to pregnancy and oral contraceptive use, other acquired conditions associated with intracranial venous thrombosis are antiphospholipid antibody syndrome and systemic malignancy. Testing for these conditions in young women with either arterial or venous occlusive disease is reasonable. On the basis of several small randomized trials and additional observational studies, anticoagulation with unfractionated or low-molecular-weight heparin is recommended, even in the presence of associated intraparenchymal hemorrhage. Anticonvulsant drugs may be necessary if seizures have occurred; treatment of increased intracranial pressure with osmotic agents (e.g., mannitol) or neurosurgical intervention may also be required. If the thrombosis occurred early in pregnancy, labor can usually be allowed to commence spontaneously with forceps assistance of delivery. However, if thrombosis occurs shortly before or during labor, cesarean delivery may be necessary. Anticoagulation should be continued after delivery for at least 3 to 6 months.

Pituitary Infarction

Sheehan syndrome is a well-recognized complication of the peripartum period (see Chapter 62 ).

Summary and Recommendations

• CT distinguishes between cerebral infarction and hemorrhage.

• Ischemic stroke during pregnancy can be treated with intravenous thrombolysis within 3 to 4.5 hours of onset and mechanical embolectomy within 24 hours of onset.

• CT angiography is the test of choice to determine eligibility for embolectomy within 6 hours and the combination of CT angiography and CT perfusion between 6 and 24 hours.

• Magnetic resonance venography is indicated if intracranial venous sinus thrombosis is suspected.