Several studies have looked at the pattern of migraine in pregnancy with consistent findings. The effect of pregnancy does not parallel that of oral contraceptive agents. In a case-control study of 100 women with migraine with aura and 200 age-matched controls, whereas oral contraceptives worsened headaches in 25% of women, pregnancy lessened migraine in 77%.⁵ In a prospective study of 49 women, migraine frequency decreased by about 50% in the first trimester, with 10.6% of patients experiencing complete remittance.⁶ In the second trimester, 87% of women reported a decrease in migraine frequency with 53% remission, and in the third trimester, 87% of women reported a decrease in migraine frequency with 79% remission. Migraine returned in 4% of women by day 2 postpartum, 34% by 1 week, and 36% by 1 month. Women who breastfed had a lower rate of early headache recurrence after delivery. Women with menstrual migraine, defined strictly⁷ as headache that occurs on the first day of menses, have the greatest improvement in headaches during pregnancy. Pregnancy appears to affect the pattern of migraine with and without aura differently. In a self-reporting study comparing 88 women with migraine with aura and 180 women with migraine without aura, women with aura were more likely to have premenstrual syndrome and headaches exacerbated by oral contraceptive agents, and they were more likely to get relief from headaches during pregnancy.⁸ Women with migraine without aura had more menstrual migraine.

Valproic acid and carbamazepine can cause neural tube and other major fetal malformations, and they should not be used for headache prevention during pregnancy.¹³^–¹⁵ Gabapentin is contraindicated given its effects on bony development and overall growth in laboratory animal studies. Experience with newer antiepileptic agents is limited, and they, too, are best avoided as migraine prophylactic agents.

Commonly used supplements that are regarded as benign should be reconsidered in pregnancy. Magnesium can be an effective preventive medication for some women, but it may delay the onset and progression of labor. Riboflavin is regarded as relatively safe in pregnancy.

A short course of corticosteroids can be used safely in mid to late pregnancy to break a persistent headache without undue exposure to the fetus. Prednisone is the preferred choice, because it crosses the placenta less than other corticosteroid preparations.

Migraines that might have been relieved by late pregnancy tend to return after delivery, and treatment during breastfeeding warrants special considerations. The American Association of Pediatrics recommends nonsteroidal anti-inflammatory agents as generally safe for breastfeeding. Highly lipophilic medications, such as sumatriptan, cross readily into breast milk. For women taken sumatriptan, a safe recommendation is to pump breast milk 6 hours after an oral dose and 4 hours after an injected dose. Caution should be used with β-blockers, because they may cause newborn bradycardia or heart block.

Cerebrovascular Issues

Epidemiology

There have been no well-designed, long-term prospective studies evaluating the incidence of stroke in pregnancy and the puerperium. Comparisons of available retrospective data are limited because authors have defined the patient groups differently, sometimes including all or only late pregnancy, all or part of the postpartum period, or including or excluding women with spontaneous or therapeutic abortions, and they have defined stroke differently, including or excluding venous sinus thrombosis, subarachnoid hemorrhage, and transient ischemic attack. Estimates of incidence of stroke in pregnancy range from a low of 4.3 to a high of 210 cases per 100,000. In recent years, several retrospective studies have examined stroke in pregnancy shedding some new light on its epidemiology. These studies agree with older ones that there is an increased risk of stroke in pregnancy and the puerperium compared with the nonpregnant state and that the distribution between ischemic and hemorrhagic stroke is roughly equal. In public hospitals in the Ile de France studied from 1989 to 1992, there were 4.3 ischemic strokes per 100,000 women and 4.6 hemorrhagic strokes per 100,000 women.¹⁶ This study included only 2 postpartum weeks, and it excluded cerebral venous sinus thrombosis, subarachnoid hemorrhage, and transient ischemic attacks from the definition of stroke. A population-based retrospective study of hospitals in the Baltimore-Washington, DC, area from 1988 through 1991 was able to compare the rate of stroke in pregnant and nonpregnant women aged 15 to 44 years to assess the risk attributable to pregnancy. This study found 11 cerebral infarctions and 9 intracerebral hemorrhages per 100,000 deliveries. There was no increased risk of ischemic stroke during pregnancy (relative risk [RR], 0.7) and only a small trend of increase in the risk of intracerebral hemorrhage (RR, 2.5; confidence interval, 1.0 to 6.4) but a significantly increased risk of stroke of both types in the 6 weeks postpartum (ischemic stroke RR, 8.7; confidence interval, 4.6 to 16.7; hemorrhagic stroke RR, 28.3; confidence interval, 13.0 to 61.4).¹⁷ The study population was almost 40% African American. It included women having abortions and stillbirths, and it considered venous sinus thrombosis in the ischemic stroke category. Another study looked at women aged 15 to 44 years from a single hospital in Taiwan during the period from 1984 to 2002 identifying first-ever cerebral infarct (including both cerebral venous thrombosis and arterial infarcts), cerebral hemorrhage, and subarachnoid hemorrhage. This study found a higher rate of strokes of 46.2 per 100,000 pregnancies (26 per 100,000 ischemic and 20.1 per 100,000 hemorrhagic), with the highest rate in the postpartum period.¹⁸ The high rate of rheumatic heart disease in the Taiwanese population likely accounts for the higher rate of ischemic strokes and emphasizes the difference in populations across these studies.

Together, these studies suggest that the overall risk of stroke during pregnancy is low and determined mainly by the risk of background risk factors and that although there is likely a very small increase in hemorrhagic risk during pregnancy, there is a large increase in risk of ischemic and more so of hemorrhagic strokes during the postpartum period.

Pathophysiology of Ischemic Stroke in Pregnancy

Strokes during pregnancy can be divided into those with mechanisms unique to pregnancy and those with more common mechanisms that are influenced by the pregnant state. The categories used in analyses differ among the available studies and missing categories and possible overlap of general and more specific categories make judgments tentative. However, it appears that the most common causes of ischemic stroke in pregnant and postpartum women are eclampsia and cerebral venous thrombosis, followed by cardioembolism and a variety of other mechanisms that are found in large studies of nonpregnant young patients with ischemic stroke. The French study mentioned earlier looked retrospectively at 348,295 deliveries from 1989 to 1992 in 63 public maternity wards in the Ile de France region for strokes that occurred through the second postpartum week.¹⁶ Cases of venous sinus thrombosis were excluded. The etiologies of the 15 ischemic strokes were found to be preeclampsia-eclampsia (7), unknown (5), dissection (1), amniotic fluid embolism (1), and protein S deficiency (1). The Baltimore-Washington, DC, study looked at all women aged 15 to 44 years who were discharged from 46 area hospitals during the period 1988 through 1991 and found 17 cerebral infarctions, of which they reported the cause in 16.¹⁷ Ischemic strokes were due to undetermined causes (6), preeclampsia-eclampsia (4), primary central nervous system vasculopathy (2), carotid dissection (1), thrombotic thrombocytopenic purpura (1), cortical vein thrombosis (1), and postherpetic vasculitis (1). If “primary central nervous system vasculopathy” represents eclampsia, which is likely (see later), then 6 of the 16 were related to eclampsia. A review of all deliveries at a Toronto hospital from January 1980 through June 1997 found 21 infarctions in pregnant and postpartum women among 50,711 deliveries.¹⁹ Eight of the 21 infarctions were venous. Of 13 arterial infarctions, 6 had unknown causes, and 4 were attributed to cardioembolism. Of the infarctions with unknown causes reported, two of six arterial infarctions and two of five venous infarctions occurred in women with preeclampsia. The study from Taiwan for first-ever cerebral infarction reported strokes among all women aged 15 to 44 years from a single hospital from 1984 through 2002.¹⁸ Among 49,796 pregnancies, 27 ischemic strokes (16 arterial, 11 venous) were related to pregnancy and the 6 weeks postpartum. The identified causes of the 16 arterial strokes were cardioembolism (9 [7 rheumatic and 2 nonrheumatic]), protein S deficiency (3), unknown (2), preeclampsia-eclampsia (1), and aneurysm (1). Of 11 patients identified with cerebral venous sinus thrombosis, identified causes were protein S deficiency (4), undetermined (4), and other coagulopathies (3) (systemic lupus erythematosus, protein C deficiency, and antiphospholipid syndrome).

Although case identification and reporting are different across studies, the data consistently indicate an increase in the risk of cerebral venous sinus thrombosis likely related to the mechanisms of hypercoagulability in effect in late pregnancy and the postpartum period, a unique risk of stroke related to preeclampsia-eclampsia also in late pregnancy and postpartum, and a small number of cases of stroke from causes seen in nonpregnant populations. The emergence of cardioembolism as the commonest cause in the Taiwanese study probably reflects the higher risk of rheumatic heart disease in that population.

Amniotic fluid embolism is a very rare cause of stroke specific to pregnancy. It occurs during labor or immediately after childbirth. Affected women are usually multiparous and older than 30 years. Amniotic fluid enters the circulation to embolize simultaneously the lungs and, via a patent foramen ovale or other right-to-left shunt, the cerebral circulation. Thus, it presents with cyanosis, dyspnea, hypotension, rarely seizures, and stroke. It may rapidly progress to cardiopulmonary collapse.

Metastatic choriocarcinoma is a rare cause of either hemorrhagic or ischemic stroke that may be seen after pregnancy, abortion, or molar pregnancy. Angioinvasive cells of the choriocarcinoma can cause vascular obstruction directly, or they may invade the arterial wall to cause thrombosis and aneurysm with subsequent hemorrhage.

Peripartum cardiomyopathy may lead to ischemic stroke. This condition is most commonly seen in multiparous African American women over the age of 30. It is usually diagnosed between the last 2 months of pregnancy and the sixth postpartum month when a woman presents with dyspnea due to congestive heart failure. Through mechanisms that remain unclear, all four chambers of the heart become dilated, and stroke occurs from cardioembolism of mural thrombi. The incidence of stroke in women attributable to peripartum cardiomyopathy is not known. There is a risk of recurrence in future pregnancies.

Peripartum cerebral angiopathy or “angiitis” has been reported as an entity in the clinical literature. The diagnosis is typically based on the presence of headache, visual symptoms, confusion, seizures, or focal neurological signs in a peripartum woman with a cerebral angiogram showing multiple segmental narrowings. Cases have not been reported with biopsy-proved angiitis, and the description just given is consistent with eclampsia, suggesting that most reported cases are, in fact, cases of eclampsia. Some cases have been associated temporally with the use of sympathomimetic agents or bromocriptine. Unless there is clear evidence of angiitis, the best treatment is removal of any potentially offending agents, the administration of magnesium sulfate, and control of blood pressure and seizures as for other cases of eclampsia.

Pathophysiology of Hemorrhagic Stroke in Pregnancy

Subarachnoid hemorrhage has been cited as the third most common cause of nonobstetrical death in pregnancy.²⁰ Different classification of intracerebral hemorrhage including or excluding subarachnoid hemorrhage as an etiology of stroke makes comparison of studies of intracranial hemorrhage tentative.¹⁶ Sharshar and colleagues’ 1995 study of women in the Ile de France found 16 hemorrhagic strokes among 348,295 deliveries. The identified causes were eclampsia (7), unknown (3), arteriovenous malformations (2), aneurysm (2), and cavernous angioma (2).¹⁶ Kittner and colleagues’ 1996 study of women in the Baltimore-Washington, DC, area found 14 intracerebral hemorrhages among 234,023 pregnancies (including 140,167 live births, 1076 stillbirths, and 92,780 spontaneous or induced abortions). They reported the cause in 13. The identified causes were indeterminate (4), arteriovenous malformations (3), preeclampsia-eclampsia (2), cocaine use (2), primary central nervous system vasculopathy (1), and sarcoid vasculopathy (1).¹⁷ The Baltimore-Washington, DC, population-based study of risk supports the suggestion that risk of hemorrhage is increased in the postpartum period, but no hemorrhages in this study were attributed to aneurysm. There remains some controversy about whether or not there is an increased risk of hemorrhage from aneurysms and arteriovenous malformations in the peripartum period. Although studies of total rates do not find such an increase, an analysis of the risk of rupture per day suggests a several-fold increase on the day of delivery.^21,²² The increased blood volume during the second trimester and the increase in blood pressure and mobilization of fluids at the time of delivery likely lead to a slightly increased likelihood of rupture of arteriovenous malformation and aneurysms at these times.²³ For both arteriovenous malformations and aneurysms, the increasing levels of estrogen as pregnancy progresses and their vasodilating effects may also contribute to an increased susceptibility to blood vessel rupture.

Clinical Presentation

The presentation of strokes is similar in pregnant and nonpregnant patients, and the specific symptoms may provide insight into the mechanism. Headaches and seizures are common symptoms of the hypertensive encephalopathy of eclampsia, venous sinus thrombosis, and intracranial hemorrhage, and focal neurological deficits are common to all types of strokes. Although not specific to it, visual changes should raise suspicion of the hypertensive encephalopathy of preeclampsia-eclampsia.

Management of Stroke During Pregnancy

Treatment of stroke in pregnancy should proceed as it does in the nonpregnant state with some consideration to the trimester of pregnancy. If it is thought that aspirin is needed in women with ischemic strokes, then low-dose aspirin can be given. Aspirin (pregnancy category D) inhibits placentation and may promote early closure of the ductus arteriosus, so it is best to avoid it during the first and third trimesters, but it has been used safely in women with pregnancy-induced hypertension and in those with a history of recurrent loss of pregnancy. Because it may delay onset of labor, prolong labor, and increase bleeding during delivery, it is best to stop it approximately 2 weeks before expected delivery.

Warfarin (category X) crosses the placenta and can cause fatal fetal hemorrhage and developmental malformations, so it is contraindicated during pregnancy. In women who require full anticoagulation for mechanical heart valves, cerebral venous sinus thrombosis, a hypercoagulable syndrome, or other indications, heparins are the preferred treatment, because these large molecules do not cross the placenta. Both unfractionated heparin (category C) and low-molecular-weight heparin (category B) can be administered subcutaneously throughout pregnancy, and unfractionated heparin can be given intravenously as well. Intravenous heparin can be stopped around the time of delivery and restarted immediately afterward to minimized bleeding risks. Fondaparinux, an antithrombin III–mediated factor Xa inhibitor, is classified as category B. Because there is limited experience with its use in pregnancy, its use should be reserved for women who cannot tolerate heparins. Ginsberg and colleagues reviewed the use of antithrombotic agents during pregnancy in detail.²⁴

Based on theoretical considerations, tissue plasminogen activator (tPA) is contraindicated in pregnancy, unless it is thought that the benefits largely outweigh the risks in the setting of a major stroke. In fact, a few pregnant women have received tPA with no apparent adverse outcome.^25,²⁶

Hemorrhagic strokes should be treated according to accepted neurological and neurosurgical principles based on the clinical circumstances. For large (≥7 mm) asymptomatic or symptomatic aneurysms, neurosurgical guidelines recommend clipping or coiling before delivery, whenever possible. If the aneurysm cannot be clipped, cesarean delivery is generally advised even though this has not been shown to improve outcome. Treatment of aneurysmal rupture during pregnancy depends on the stage of pregnancy. The preferred option is cesarean section followed immediately by aneurysmal clipping if the woman is close to term. If rupture occurs earlier, then treatment proceeds as in nonpregnant patients. Where possible, mannitol use should be avoided or minimized to avoid fetal dehydration. Nimodipine (category C) can be used to prevent vasospasm after a significant subarachnoid hemmorhage with acceptable risk.

Treatment of arteriovenous malformations without hemorrhage during pregnancy is less clear. Based on a series reported in 1974, Robinson and colleagues ²⁷ argue for a high likelihood of bleeding during pregnancy.²⁷ Later series dispute this increase risk. Horton and colleagues²¹ found a hemorrhage rate of 3.5% per year during pregnancy and the puerperium, a rate similar to that of nonpregnant women with arteriovenous malformation and no history of prior hemorrhage. Robinson reported on arteriovenous malformations discovered after hemorrhage, whereas Horton and colleagues population included all arteriovenous malformations. This may account for some of the differences in their findings. Weir and Macdonald²² have argued that pregnancy increases the bleeding rate of arteriovenous malformations more than of aneurysms based on the relatively low ratio of aneurysms to arteriovenous malformations found as the cause of intracranial hemorrhage during pregnancy (1.3 in pregnancy versus 8.4 in all patients with intracranial hemorrhage). Both Horton and colleagues ²¹ and Dias and Sekhar²⁸ found that vaginal delivery added no risk of hemorrhage in patients with arteriovenous malformations. Based on these findings, vaginal delivery with epidural anesthesia is a reasonable choice for women with arteriovenous malformations.

Prognosis

The risk of maternal death from ischemic stroke is low and less than that from hemorrhagic stroke. Combining the French and Toronto studies mentioned, there were no maternal deaths among 36 women with ischemic stroke.¹⁶ Hemorrhagic stroke is the third most common cause of nonobstetrical death in pregnant women, and 7 of the 29 women with hemorrhage in these two studies died.²⁰ Among 20 women with ischemic stroke before delivery and known outcomes, there were 4 premature deliveries, 2 stillbirths, 1 miscarriage, and 1 termination. Of 26 women with intracerebral hemorrhage before delivery, there were 7 premature deliveries, 3 fetal deaths due to maternal death, 2 stillbirths, and 1 termination.

The risk of recurrent stroke has been investigated in two recent studies.^29,³⁰ In these studies of young women with a history of stroke, the risk of recurrent stroke in subsequent pregnancies was very low (0% to 1.8%), with the greatest risk during the postpartum period in one study. Yet, in one study, more than 75% of young women who had had a stroke did not have subsequent pregnancy due to fear of having another stroke in pregnancy, to medical advice against another pregnancy, or to physical limitations resulting from the stroke.²⁹

Preeclampsia-Eclampsia

Preeclampsia-eclampsia is a systemic disorder of mid to late pregnancy characterized by hypertension and proteinuria. The criteria for diagnosis recommended by the National High Blood Pressure Program Working Group on High Blood Pressure in Pregnancy are given in Table 112-2. Endothelial cell dysfunction plays a major role in the pathogenesis of preeclampsia and likely accounts for many of the additional clinical features that may be a part of the full syndrome, including thrombocytopenia, microangiopathic hemolytic anemia, and hepatic and renal dysfunction. The acronym HELLP has been applied to the combination of hemolytic anemia, elevated liver function tests, and low platelets in the setting of preeclampsia-eclampsia.³¹ Eclampsia is defined as preeclampsia complicated by seizures. More recent definitions accept preeclampsia with encephalopathy but without seizures as defining eclampsia. Which terminological convention is chosen does not affect our understanding of the pathophysiology.

TABLE 112-2 Criteria for the Diagnosis of Preeclampsia and Eclampsia

Preeclampsia

Gestational hypertension: systolic BP > 140 mmHg or diastolic BP > 90 mmHg in a woman who was normotensive before 20 weeks’ gestation. Onset of hypertension is typically after 20 weeks’ gestation (or earlier in trophoblastic diseases such as hydatidiform mole or hydrops), and

Proteinuria: ≥300 g protein/24 hr (usually ≥30 mg/dL; ≥1+ on dipstick)^*

Eclampsia

Preeclampsia accompanied by seizures that are not attributable to other causes.^†

* In the absence of proteinuria, preeclampsia should be strongly considered when gestational hypertension is accompanied by headache, blurred vision, abdominal pain low platelet counts, or elevated liver enzyme values.

† In the absence of seizures, coma or other features suggesting severe encephalopathy are sometimes termed eclampsia, although this extension of the definition remains non-standard.

From Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol 2000; 183: S1-S22.

Epidemiology

Preeclampsia affects 3% to 5% of pregnant women.³² It is a major cause of maternal and infant mortality and morbidity. Approximately 1 in 2000 pregnancies in the United States is complicated by eclamptic seizures, and the incidence is much higher in developing countries.^33,³⁴ Risk factors include obesity and poor nutrition, nulliparity, multiple gestations, age greater than 35 years, extrauterine and molar pregnancies, preexisting hypertension, diabetes mellitus, and thrombophilia.

Pathophysiology

Although the pathophysiology of preeclampsia-eclampsia is still not clear, abnormal placentation leading to impaired placental perfusion is likely to be the event that triggers downstream events in the disease process. Remodeling of the spiral arteries creates the low-resistance system that perfuses the intervillous space. This process is abnormal in preeclampsia leading to placental hypoperfusion. Women with preeclampsia have increased systemic vascular tone and heightened sensitivity to mediators of vasoconstriction. This increased vascular tone underlies the systemic hypertension, vasospasm, and decreased organ perfusion. There is also much evidence for a disorder of endothelial dysfunction. This endothelial cell dysfunction promotes the instability of vasomotor tone and hypertension as well as the increased vascular permeability, edema, and proteinuria characteristic of preeclampsia. The mechanism by which the placental insufficiency leads to vasomotor hyperreactivity and endothelial dysfunction is not well established. Proposed mechanisms include autoantibodies and oxidative stress, prompting activation of neutrophils and monocytes and release of inflammatory mediators. There is also clinical evidence for genetic predisposition.³² Population data argue for maternal susceptibility, yet some findings, such as a lack of concordance between monozygotic twins, an increased risk with changed paternity, and an increased risk in partners of fathers whose mothers had preeclampsia, point to a fetal contribution to susceptibility as well.³⁵^–³⁷

Almost all women with preeclampsia have absolute hypertension. Although the degree of hypertension may be mild, there is usually a significant elevation of blood pressure compared with the premorbid blood pressure during pregnancy.³⁸ If systemic hypertension and endothelial cell dysfunction reach a threshold of adequate severity, then there is a breakdown of cerebral autoregulatory vasoconstriction resulting in cerebral hyperperfusion and vasogenic edema of the brain. The endothelial dysfunction probably allows for edema to form at lower absolute perfusion pressures than in other forms of hypertensive encephalopathy. The nature and distribution of the brain lesions, as defined by computed tomography, magnetic resonance imaging, and single-photon emission computed tomography, are the same as in patients with hypertensive encephalopathy.³⁸ Hyperperfusion is most pronounced posteriorly, and edema is most prominent in the subcortical white matter of the occipital lobes and other structures fed by the posterior circulation (Fig. 112-3).^39,⁴⁰ This pattern of hyperperfusion and edema is believed to reflect the relative low density of vasoconstricting sympathetic receptors in the posterior circulation.⁴¹ Yet, gray matter and anterior circulation structures are often involved, especially in severe cases. If limited to vasogenic edema, the lesions are reversible, but intraparenchymal hemorrhage and, less commonly, infarctions due to vasospasm and thrombosis may occur. Seizures may occur as a result of the subcortical and cortical edema. Hence, the defining seizures of eclampsia appear to be the severe end of the spectrum of dysfunction caused by hyperperfusion and brain edema.

Figure 112-3 Magnetic resonance image in a woman with eclampsia. A 23-year-old primigravida woman at 30.5 weeks’ gestation awoke with a headache, nausea and vomiting, followed by a generalized tonic-clonic seizure. On presentation to the hospital, her blood pressure was 160/100. She was obtunded with mild right arm and leg weakness, hyperreflexia, and a right Babinski sign. She was treated with magnesium sulfate, labetalol, and hydralazine and emergency cesarean section. One month later, her blood pressure and neurological examination were normal. A–C, Axial fluid-attenuated inversion recovery (FLAIR) images done on presentation show the high intensity lesions involving the white and gray matter of both occipital lobes. There are no associated hyperintensities on diffusion-weighted images, consistent with cerebral edema. These lesions are typical for eclampsia in quality and location. D and E, Axial FLAIR images done on presentation show similar lesions in both basal ganglia and most prominently in the external and Internal capsules bilaterally and in both frontal lobes. Most of these lesions are not seen on diffusion-weighted imaging, consistent with cerebral edema, but there is a small focus of bright signal on diffusion-weighted imaging (with corresponding low apparent diffusion coefficient [not shown]) consistent with a small focus of infarction. F, Axial FLAIR images at 5 days after onset show resolution of the occipital lesions and partial resolution of the lesions of the left external capsule and basal ganglia, consistent with resolving edema.

Clinical Presentation

Headache and visual phenomena consistent with occipital lobe edema are the most common symptoms of eclamptic encephalopathy. Seizures may be focal or secondarily generalized. Some women have encephalopathic features without seizures, such as confusion, aphasia, cognitive deficits, and depressed level of consciousness. Although hypertensive encephalopathy and seizure are most commonly preceded by typical features of preeclampsia, it is not uncommon for women to present with seizures in late pregnancy and in the early postpartum period with only mild hypertension.⁴²

Diagnostic Evaluation and Differential Diagnosis

The diagnosis of preeclampsia is based on the fulfillment of the clinical criteria of hypertension and proteinuria. The diagnosis of eclampsia is based on the finding of encephalopathy or seizures in late pregnancy or the postpartum period with computed tomography scanning or magnetic resonance imaging that shows evidence of characteristic edema and that rules out alternative brain lesions. Magnetic resonance imaging is more sensitive for detection of edema than computed tomography, and diffusion-weighted imaging is useful to distinguish reversible edema from ischemic stroke. Technetium single-photon emission computed tomography imaging may demonstrate hyperperfusion characteristic of hypertensive encephalopathy, but this is not usually necessary for diagnosis. Patients with intracerebral or cerebellar hemorrhage should have vascular imaging to rule out an underlying vascular anomaly. Almost all women with eclampsia have hypertension, although the elevation may be mild or only relatively raised from a low baseline blood pressure of pregnancy.³⁸

Management

The mainstays of management are the rapid control of hypertension and the administration of magnesium sulfate. Mean arterial blood pressure should usually be lowered by 15% to 20%. This is best achieved rapidly with intravenous medications such as labetalol, hydralazine, or nicardipine. Nitroprusside and nitroglycerin can cause fetal cyanide toxicity, so they should be avoided during pregnancy. Angiotensin-converting enzyme inhibitors are also contraindicated during pregnancy (category D).⁴³ Most eclamptic women have intravascular volume contraction and will benefit from volume replacement, but invasive hemodynamic monitoring may be useful to guide fluid therapy in severely ill women with refractory hypertension, oliguria, or pulmonary edema.^44,⁴⁵

Magnesium sulfate therapy now has strong support from clinical trials both to prevent further seizures in women with eclampsia and to prevent seizures in women with preeclampsia.^46,⁴⁷ Therapy can start with a loading dose of 4 to 6 g of magnesium sulfate followed by a 2 g/hr infusion. A supplemental loading dose of 2 g may be given if seizures recur shortly after the initial load. Tendon reflexes, respiratory function, and urine output should be followed closely. It is common practice to follow serum levels of magnesium, although it has not been established that this helps to guide therapy. It is most important to follow women for signs of neurological depression and to adjust magnesium sulfate doses accordingly. At levels of 8 to 10 mEq/L, tendon reflexes are typically depressed. At levels above 10 to 12 mEq/L, there is a high risk of respiratory depression. When patellar reflexes are lost, magnesium sulfate should be discontinued. If respirations are depressed, calcium gluconate should be given. Doses of magnesium sulfate should be adjusted for renal insufficiency, for example, by reducing the loading dose to 4 g and halving the maintenance dose. Although recent data argue that magnesium sulfate is better than phenytoin or diazepam for eclamptic seizures, refractory seizures should be treated aggressively with traditional antiepileptic drugs in addition to magnesium sulfate.

Epilepsy

In past years, women with epilepsy were discouraged from having children, and in some states laws required sterilization or prohibited marriage of women with epilepsy. A more scientific understanding of the true medical issues has corrected this discriminatory practice. Although the vast majority of epileptic women have normal pregnancies and deliveries, there is still a need for careful neurological care and monitoring during preconception, pregnancy, and the postpartum period.

Effect of Pregnancy on Seizure Frequency

Epilepsy affects approximately 1% of the population, and about 1 million women in their reproductive years have epilepsy. It is estimated that 0.3% to 0.5% of all births occur in epileptic women. Seizure frequency seems to increase by approximately 25% to 33% during pregnancy.⁴⁸ Older data suggest that the greatest increase occurs in the first trimester, but more recent data suggest a similar effect throughout pregnancy.^49,⁵⁰ Approximately 1% to 2% of women have a seizure during labor and delivery, and another 1% to 2% have a seizure within 24 hours after delivery. There are many possible explanations for this high rate of seizures. Many of these women do not have epilepsy but rather eclampsia or other symptomatic seizures. Among epileptics, the proconvulsant effect of the increased estrogen levels, third trimester and postpartum sleep deprivation, and decreasing levels of antiepileptic drugs due to more rapid metabolic inactivation and increased total body water may account for some of the increase in seizure frequency. Noncompliance with medication due to fear of side effects or an inability to tolerate oral medications during early pregnancy and the delivery period may also contribute.

Consequences of Seizures During Pregnancy

Prolonged generalized tonic-clonic activity during pregnancy has been shown to lead to maternal acidosis, hypoxia, and decreased fetal heart rate.⁵¹ Trauma from a seizure can lead to poor fetal and maternal outcomes, including premature rupture of membranes, infection, placenta abruption, placenta previa, and miscarriage. Most reports of poor outcome are from literature dating from the mid-twentieth century. More recent studies have found that complications of seizures are rare.⁵² Abruptio placenta was seen in 1% to 5% women after minor injury and in 20% to 50% after major seizure-induced injury. The effects of nonconvulsive status epilepticus during pregnancy are not well known.

Management

Pregnant patients should be warned about the effects of trauma during pregnancy and should be counseled about safety measures to prevent harm. If antiepileptic drugs are indicated, then it should be emphasized that compliance will minimize risk of undue harm from a seizure.

Treatment of women with epilepsy was outlined in the American Academy of Neurology Consensus Guidelines in 1998.⁵³ Counseling and treatment should be undertaken long before planned conception. It is recommended that women be treated with an antiepileptic drug most appropriate to their seizure type and that they be maintained on a single agent whenever possible. If antiepileptic drug withdrawal before conception is planned, then it is best accomplished six months before the planned date. Folic acid, 0.36 to 5 mg daily, is recommended, because it is safe, and, although not conclusive, studies suggest that folate supplementation lowers the risk of teratogenicity attributable to antiepileptic drugs.

The effects of pregnancy on drug distribution, pharmacokinetics, and metabolism are complex due to the simultaneous and interdependent changes in total body water, estrogen and progesterone levels, albumin levels, and transaminase activity. Estrogens increase hepatic enzyme activity, which leads to an increased rate of metabolic inactivation of most antiepileptic drugs throughout pregnancy and the early postpartum period. Total body water begins to increase in early pregnancy and returns to baseline by about 2 to 3 months postpartum. This increase in total body water results in a decrease in serum albumin concentration despite the increase in total protein synthesis that is stimulated by estrogens. Absolute drug levels decline as pregnancy progresses, but much of this decline is compensated by an increase in the ratio of free (active): total drug. It is usually necessary to increase antiepileptic drug doses to maintain therapeutic levels in late pregnancy. After delivery, doses should be lowered gradually to coincide with the return of total body water volume and albumin levels to prepregnancy values over the 2 to 3 months postpartum.

Antiepileptic drug levels are best checked at least every month throughout pregnancy and then every 2 weeks in the third trimester. During the postpartum period, levels should be monitored every 1 to 2 weeks to keep pace with the rapid fluid shifts. Because the complex relationships of changes in volume of distribution, induction of metabolism, decline in albumin, and redistribution of the free to total ratio are not predictable, it is important to measure both total and free levels when highly protein-bound agents such as phenytoin and valproate are used. If a drug is minimally protein bound, then the relationship is usually not clinically significant, and total levels provide adequate guidance.

Selection of the proper antiepileptic drug is determined largely by standard considerations based on seizure type and individualized experience with effectiveness and tolerance. Development of congenital anomalies, fetal loss, and stillbirth all occur at an increased rate in women with epilepsy. A huge volume of data addresses the risk of fetal malformations induced by early exposure to antiepileptic drugs. Epileptic women taking antiepileptic drugs have a rate of major congenital malformations of about 4% to 6%. This is higher than the rate of 2% to 3% in the normal population. The baseline rate of major congenital malformations in epileptic women not taking antiepileptic drugs also may be elevated, but studies have not consistently found this increased background rate.⁵⁴ Lack of controlled data greatly limits conclusions about the risk of particular antiepileptic drugs. The risk of an adverse outcome increases with increasing number of antiepileptic drugs; therefore, the recommendation to minimize polypharmacy is strong.

Although there may be little difference in the risk conferred by the commonly used agents, certain risks are notable. The developing fetus is most sensitive to induction of malformations by exogenous agents between days 21 and 56 of gestation. A 1982 French study documented an increase in neural tube defects in women on antiepileptic drugs, especially valproate and carbamazepine.⁵⁵ Women taking carbamazepine and valproate were noted to have an increased rate of spina bifida aperta.¹³^–¹⁵ These congenital malformations are thought to be associated with the low serum and red blood cell folate levels that lead to hyperhomocystinemia. A family history of neural tube defects further increases the risk, so these drugs should be especially avoided in women with a positive family history. From the small population evaluated so far, the Gabapentin Pregnancy Registry has found a rate of malformations fetal similar to the rate in the general population.⁵⁶ A review of the International Lamotrigine Pregnancy Registry, again with a small sample, found a 2.9% risk of major congenital malformations, also comparable to that of the general population and to registries of other drugs.^57,⁵⁸ Valproic acid alone may confer a greater risk of malformations than other agents, and the combination of valproic acid, carbamazepine, and phenobarbital may confer a particularly high risk.⁵⁴

Older enzyme-inducing antiepileptic drugs cause vitamin K deficiency in the newborn by decreasing the transport of vitamin K across the placenta. This is most common after exposure to phenobarbital, primidone, and phenytoin, but it can also occur with other antiepileptic drugs, including carbamazepine, benzodiazepines, and ethosuximide. A woman who has been maintained on one of these antiepileptic drugs throughout pregnancy should receive 10 mg/d of oral vitamin K during the last month of pregnancy. At birth, the baby should be given 1 mg of intramuscular vitamin K.

During labor and delivery, antiepileptic drugs should be given in oral or intravenous form. Many newer antiepileptic drugs that are not available in intravenous form are long acting and may sustain the mother throughout delivery. If a woman taking an exclusively oral agent has labor extending beyond the duration of action of the oral agent, then she may be given intravenous lorazepam in doses of 1 mg every 8 hours during delivery. In such a case, the pediatrician should be notified to prepare for the possible delivery of a sluggish baby.

Breastfeeding is generally safe for mothers taking antiepileptic drugs. The secretion of antiepileptic drugs into breast milk depends on the molecular size and charge and the degree of protein binding. Drugs that are highly protein bound have much lower levels in the breast milk than in plasma. The ultimate level of the antiepileptic drug achieved in the blood of the infant also depends on the half-life of the drug. Most drugs achieve a low enough level in the infant to allow their safe use during breastfeeding. Due to their long half-lives, phenobarbital and many benzodiazepines may accumulate in the infant causing sedation and withdrawal. This is especially true of phenobarbital, which has a longer half-life in infants than in adults, up to 300 hours. There is so far little information about the safety of breastfeeding using newer agents.