Tension Headache

Headache during pregnancy is common. Usually a patient visits the neurologist to receive reassurance that no serious medical problem is apparent. Of the headaches that occur during pregnancy, benign tension headaches are seen most often (see Chapter 69). No known association exists with hormones and, specifically, no association with the hormonal changes of pregnancy. Treatment for mild headaches often includes behavioral therapy, adequate rest, moist heat, massage, exercise, avoidance of triggering factors, and use of acetaminophen. For severe headaches, the use of a tricyclic antidepressant such as amitriptyline or nortriptyline may be helpful. No evidence of embryopathy occurs with amitriptyline, and preschool children exposed in utero to tricyclic antidepressants have normal global IQs, language, and behavioral development. Fluoxetine may cause uncommon but serious fetal risks (Chambers et al., 2006; Diav-Citrin et al., 2008; Mills, 2006).

Migraine Headache

More than 80% of women with migraine clearly show improvement during pregnancy, but 15% continue to have headaches, and in 5% headaches worsen. The prognosis for women with migraine without aura is better than that for women with migraine with aura. Headaches were more likely to persist with diagnosed menstrual migraine, hyperemesis, or a “pathological pregnancy course” in a prospective study. For women anticipating pregnancy, the physician may discontinue or reduce the dose of all migraine medications to lower the risk of possible fetal damage and offer vigorous treatment with behavioral therapy, moist heat, and the judicious use of acetaminophen or opioid preparations. Migraine usually lessens during the second and third trimesters. The diagnosis of complicated migraine or de novo migraine with aura during pregnancy requires a thorough consideration of other diagnoses. Migraine may increase the risk for preeclampsia, especially for patients with prepregnancy obesity (Adeney et al., 2005) and may increase the risk for peripartum stroke (James et al., 2005). While small increased incidences of low birth weight, preterm birth, and cesarean section were noted in an Asian population (odds ratios 1.16, 1.24, and 1.16, respectively), the single large study did not control for the use of medications in this population, where the authors note equivocal clarity in the database diagnosis of migraine (Chen et al., 2010).

Before Pregnancy

Fertility is unaffected by myasthenia gravis, and oral contraceptive agents do not weaken these patients. No single study offers certainty with regard to the cumulative risk pregnancy causes in the patient with known myasthenia gravis (Hoffman et al., 2007). Conditions may remain stable, improve, worsen, or both improve and worsen at different stages of pregnancy. Approximately two-thirds of patients report some worsening at some time during pregnancy or the puerperium. The puerperium and first trimester are times of greatest risk. The course of myasthenia gravis for a future pregnancy is not predictable by the course of previous pregnancies.

The effect of thymectomy on myasthenia gravis usually is delayed. The potential mother can be advised that the procedure may be helpful for a pregnancy beginning approximately 1 year after surgery. Generally in women with myasthenia who may become pregnant, the physician should use drugs other than azathioprine and cyclosporine. Mycophenolate is associated with an increased risk of congenital malformations and spontaneous abortion prompting the manufacturers to recommend a negative pregnancy test within one week before beginning therapy and use of two reliable forms of contraception four weeks before and six weeks after therapy. Mycophenolate may affect the effectiveness of hormonal contraception. Myasthenia gravis does not influence the contractile strength of the smooth muscle of the uterus, the incidence of postpartum hemorrhage, or the occurrence of toxemia.

During Pregnancy

The medical therapy of myasthenia gravis changes little with pregnancy. Anticholinesterase agents including edrophonium (Tensilon) and plasmapheresis are relatively safe. Rapid drug metabolism during pregnancy may require increasing the rate or dose of anticholinesterase drugs. Corticosteroids may increase the risk for gestational diabetes and preeclampsia. Abortion does not lessen the manifestations of myasthenia gravis. Although the use of intravenous human immunoglobulin (IVIG) appears safe during pregnancy, the number of myasthenic patients studied is small. In animals, azathioprine is teratogenic, and low levels cross the placenta. Physicians generally advise patients with myasthenia to discontinue azathioprine in preparation for pregnancy. A few women with myasthenia gravis receiving azathioprine during pregnancy have given birth to healthy children. Some researchers point to the uncommon reports of human teratogenicity and intimate that azathioprine might be safer than animal data suggest. Mycophenolate is contraindicated during pregnancy.

Regional anesthesia is preferred for cesarean section. When the patient is taking anticholinesterase agents, metabolism of procaine is slowed and poorly predictable; in these patients, lidocaine is favored for local anesthesia. Neuromuscular blocking agents such as curariform drugs must be avoided because they may have a greatly prolonged effect in patients with myasthenia gravis. The use of magnesium sulfate as a tocolytic agent or a treatment for preeclampsia may precipitate a myasthenic crisis and is contraindicated.

Postpartum

Breastfeeding poses no significant difficulty, despite evidence that the antiacetylcholine receptor antibodies do pass to the baby in breast milk. Cyclosporine and azathioprine are secreted in breast milk; in general they should be avoided because of their risk for immunosuppression and tumorigenic potential. When medically necessary, the small but uncertain tumorigenic potential of azathioprine must be weighed against the known benefits of breast milk.

Corticosteroids also are secreted into breast milk but in small amounts. Large doses of anticholinesterase drugs taken by the mother may lead to gastrointestinal upset in the breastfed newborn.

Pregnancy Outcome

A retrospective Norwegian study reported an increased risk for premature rupture of amniotic membranes and double the rate of cesarean section among myasthenic women (Hoff et al., 2003). A smaller retrospective Taiwanese study found statistically insignificant increased risk of cesarean section, infants small for gestational age, low birthweight, and no difference for preterm delivery (Wen et al., 2009). Premature labor may be more common in women with myasthenia gravis but varies considerably among multiple studies.

Perinatal mortality increases to 6% to 8% for infants of women with myasthenia gravis, which is approximately five times that of the normal population. Approximately 2% of these are stillborn. Transient neonatal myasthenia affects 10% to 20% of infants born to women with myasthenia gravis. Most infants who develop transient myasthenia gravis do so within the first day, but weakness may begin up to 4 days after delivery and usually resolves within 3 to 6 weeks. Neonates require careful observation for at least 4 days. An imperfect correlation exists between maternal levels of antiacetylcholine receptor antibodies and the likelihood that the neonate will develop transient myasthenia gravis. Intrauterine exposure to receptor antibodies rarely may result in arthrogryposis, which has a high likelihood of recurrence in future pregnancy. The role of intragestational plasmapheresis and immunosuppression to prevent this condition in subsequent pregnancies is unknown (Polizzi et al., 2000).

Myotonic Dystrophy

Pregnancy is uncommon in women who have advanced myotonic dystrophy, probably because of progressive ovarian failure. Before the development of advanced disease, there is no significant reduction in fertility. For women who are able to conceive, pregnancy can be hazardous for both mother and fetus. Myotonic weakness often worsens during the second half of pregnancy. Congestive heart failure is reported. Ineffective uterine contractions, premature labor, and breech presentation often complicate labor. Tocolysis may result in aggravation of myotonia. Oxytocin can stimulate the myotonic uterus to produce increased contractions. Myotonic dystrophy complicates obstetric anesthesia, and regional anesthesia is preferred. Patients with myotonic dystrophy are unduly sensitive to respiratory suppression with pentobarbital. After delivery, hypotonic uterine dysfunction results in an increased risk for retained placenta and postpartum hemorrhage.

Half of children born to women with myotonic dystrophy inherit the disorder. Anticipation due to an increased number of triplet repeats (see Chapter 40) is responsible for the syndrome of congenital myotonic dystrophy in the neonate (see Chapter 79). Many neonates are hypotonic, and reported rates of morbidity are high. Fetal myotonic dystrophy may affect fetal swallowing, causing polyhydramnios. Prenatal diagnostic testing with amniocentesis or chorionic villus biopsy is available.

Available data suggest that myotonic dystrophy type 2 (see Chapter 79) is more benign than type 1 and may not result in problems with general anesthesia or increase problems of delivery for the pregnant woman. Type 2 may not increase the risk for polyhydramnios or stillbirth or result in congenital myotonic dystrophy (Day et al., 2003). Another study noted that about 21% in their series first presented with myotonic weakness during pregnancy, which worsened during subsequent pregnancies (Rudnik-Schöneborn et al., 2006). Some 17% of patients miscarried, half experienced preterm labor, and 27% delivered preterm.

Facioscapulohumeral Dystrophy

Information gathered in a study with self-selected and possible recall bias provides insight that the second stage of labor may be adversely affected by skeletal muscle weakness for many, and associated with worsened irreversible weakness in 24% of patients with this dominantly inherited disorder. In order of frequency, complaints included worsening of generalized weakness, frequent falling, difficulty carrying the infant owing to worsening of shoulder weakness, worsening or new-onset pain, and difficulty carrying the infant owing to worsening of leg weakness. Despite the worsening, 90% of women reported they would choose pregnancy again. Outcomes of pregnancy were good, with a statistically higher frequency of low birth weight (16.4%), the incidence of operative delivery (23.8%), and forceps delivery was more than doubled (19%) (Ciafaloni et al., 2006).

Inflammatory Myopathy

Pregnancy worsens or activates polymyositis and dermatomyositis. Manifestations of collagen vascular disease commonly associated with myositis may complicate pregnancy. More than half of fetuses die, but surviving infants thrive. Immunosuppressive treatment is advisable for gravid women. In small studies, decreased activity of myositis correlates with a more favorable outcome for the fetus (Silva et al., 2003; Váncsa et al., 2007).

Bell Palsy

Facial nerve palsy occurs three to four times more commonly during pregnancy and the puerperium, usually occurring at about 35 weeks’ gestation (Shmorgun et al., 2002). However, reanalysis of these data suggests that this conclusion may be inaccurate, and the frequency of Bell palsy may be about the same in women of childbearing age (Vrabec et al., 2007). A retrospective chart review found the prognosis for recovery of facial nerve function to be worse when facial palsy occurs during pregnancy (Gillman et al., 2002) but not usually when the severity of the facial palsy is mild. Researchers find increased frequency of toxemia and hypertension in patients with gestational facial palsy and recommend careful and continued monitoring of the affected woman for these conditions. Herpes simplex virus type 1 is the cause of most facial palsies, and varicella-zoster far less often. Pharmacological therapy of Bell palsy during pregnancy remains controversial. When begun within 3 days of onset of facial weakness, prednisone 1 mg/kg for 5 days, tapering rapidly over a total 10-day course, may be effective in improving the prognosis in nongravid adults and is considered safest when not used during the first trimester (Vrabec et al., 2007). The routine use of antiviral drugs simultaneously has not been proven unequivocally effective in nongravid adults in the absence of a varicella-zoster syndrome (Ramsay Hunt syndrome or zoster sine herpete), including a meta-analysis (Browning, 2010; Quant et al., 2009). This combination of drugs has not been tested adequately during pregnancy. Individually, the drugs pose low risk. Patching of the eye and lubricating eye drops may help prevent corneal irritation (see Chapter 70).

Carpal Tunnel Syndrome

Approximately one in five pregnant women reports nocturnal hand paresthesias, primarily during the last trimester, often associated with peripheral edema. Excessive weight gain and fluid retention increase the occurrence of these complaints. Symptoms usually resolve spontaneously within weeks after parturition. During pregnancy, conservative therapy is indicated. Splinting of the wrist in the neutral position is helpful. Additionally, some physicians inject corticosteroids into the carpal tunnel. When hand muscles supplied by the median nerve weaken, the treatment of choice is surgical decompression using fiberoptic techniques.

Low Back Pain

Low back pain is ubiquitous in the nongravid female population and increases during pregnancy. In retrospective questionnaires, researchers find more than half the pregnant population recall pain during their pregnancy and half of those women remembered radiation of the back pain to the extremeties (Fast et al., 1987). As many as three-quarters of pregnant women report low back pain at some time in their pregnancy in prospective studies (Pennick and Young, 2008). A careful, prospective study estimated the prevalence of “true” sciatica, radiation of the pain in a dermatomal distribution to be less than 1% (Ostgaard et al., 1991). Investigators blame this torment and its propensity to increase after the fifth month of pregnancy on increasing lumbar lordosis, direct pressure from the enlarging uterus, postural stress, and hormonally-induced ligamentous laxity. In one Swedish study, nearly all women experiencing back pain during pregnancy serious enough to provoke loss of work suffered recurrence of back pain in a subsequent pregnancy and low back pain recurred commonly in the nongravid state (Brynhildsen et al., 1998). MRI and electromyography (EMG) can be helpful rarely. Risks of EMG are negligible. Authorities advise avoiding muscles that bring the EMG needle too close to the developing fetus. Risks of MRI are known incompletely (see Imaging, earlier). An extensive review of the literature concluded that interventions for the treatment of the low back pain of pregnancy are biased enough that unequivocal therapeutic direction cannot be indicated. The efficacy and risk of techniques to prevent low back pain are unknown. Studies of specifically tailored strengthening exercise, sitting pelvic tilt exercise programs and water gymnastics all reported beneficial effects. The effect of physiotherapy is small but may be of some benefit as may be acupuncture. Studies on acupuncture claim better results than for physiotherapy (Pennick et al., 2008).

When minor neurological deficits accompany a syndrome suggesting compressive disc disease, authorities recommend a conservative approach based on limited case series (Laban et al., 1995). Surgical management for compressive disc disease has been successfully employed and is suggested during pregnancy to treat severe or progressive neurological deficits and in the presence of a cauda equina syndrome (Brown and Levi, 2001; Laban et al., 1995).

We lack studies or consensus agreement on the preferred mode of delivery for patients with herniated and symptomatic lumbosacral disc disease. Laban et al., (1995) describes patients who underwent cesarean section successfully to avoid the theoretical increases of epidural venous pressure during the valsalva maneuver associated with pregnancy. This author suggests that a reflex response of skeletal muscle to pain during pregnancy may be responsible for elevated venous pressure, and that regional block anesthesia may be as effective with vaginal delivery.

Meralgia Paresthetica

The expanding abdominal wall and the increased lordosis of pregnancy stretch the lateral femoral cutaneous nerve to the thigh as it penetrates the tensor fascia lata or at the inguinal ligament. This unilateral or bilateral affliction of late pregnancy resolves within 3 months postpartum. The condition may be exacerbated during the hip flexion and increased intra-abdominal pressure of delivery and when retractors are applied for cesarean section.

Acute Polyradiculoneuropathy (Guillain-Barré Syndrome)

Pregnancy does not affect the incidence or course of acute polyradiculoneuropathy, but some investigators believe the pregnant patient may be more vulnerable to complications (Chan et al., 2004). Usually, infants of a mother without complications are born healthy. Only one case of neonatal acute polyradiculopathy resulting from maternal disease is reported. Some investigators recommend fluid loading before plasmapheresis to prevent hypotension. Others suggest avoidance of tocolytics in the presence of autonomic instability. IVIG has been used safely during pregnancy, but the number of patients who received this therapy and were studied remains small. Uterine contractions are unaffected by the disease. Cesarean section is only for obstetric indications. Severe hyperkalemia caused presumably by succinylcholine anesthesia resulting in reversible cardiac arrest has been described in a pregnant woman 3 weeks after complete recovery from acute polyradiculopathy (Feldman J.M., 1990). This case and additional related reports have prompted some authors to suggest that the combination of pregnancy and acute polyradiculopathy should lead to the cautious use or avoidance of depolarizing neuromuscular blocking agents.

Chronic Inflammatory Demyelinating Polyneuropathy

CIDP is three times more likely to relapse during the last trimester and puerperium than in the absence of pregnancy. Infants are unaffected. Corticosteroids, plasmapheresis, and IVIG treat exacerbations during pregnancy. Oral contraceptives can worsen CIDP.

Charcot-Marie-Tooth Disease Type 1

Small studies indicate that Charcot-Marie-Tooth disease type 1 worsens in approximately half of affected women during pregnancy. The magnitude of the effect of pregnancy on this disease remains unclear. Risk is less when weakness begins in adult life. After delivery, this deterioration improves in a third of patients and becomes persistently progressive in two-thirds of patients, although studies vary in their observations (Swan et al., 2007). A retrospective Norwegian study found that affected women were twice as likely as the general population to have fetal-presentation anomalies, experience postpartum hemorrhage, and undergo cesarean section—commonly on an emergency basis. Forceps delivery occurred three times as often (Hoff et al., 2005). Epidural anesthesia for labor is safe.

Gestational Polyneuropathy

Distal symmetrical neuropathy affects malnourished women. Presumably, thiamine and possibly other nutrients are deficient. The acute presentation of symmetrical neuropathy and Wernicke encephalopathy (see later discussion) in the third and fourth months may be due to the thiamine deficiency associated with hyperemesis gravidarum.

Maternal Brachial Plexus Neuropathy

Hereditary brachial plexus neuropathy is an autosomal dominant disorder characterized by periodic attacks of asymmetrical pain, weakness, atrophy, and sensory alteration of the shoulder girdle and upper limbs attributed to involvement of proximal upper-limb nerves or the brachial plexus. Symptoms are indistinguishable from those of Parsonage-Turner syndrome (see Chapter 29). Attacks may begin as early as 3 hours postpartum despite cesarean section, may recur for weeks following parturition, and may follow subsequent pregnancies. Intravenous methylprednisolone may reduce the pain but seems not to influence the course of the disease (Klein et al., 2002).

Maternal Obstetric Palsy

Peripheral nerves are occasionally compressed by the fetal head, the application of forceps, and improperly positioned leg holders. Craniopelvic disproportion, dystocia, prolonged labor, and primigravida status contribute to these injuries. Unilateral lumbosacral (L4, L5, and rarely S1) plexus injury is most common. The fetal brow strikes the nerves as they cross the posterior brim of the true pelvis. The associated sensory deficit usually involves more widespread sensory loss than that due to peroneal neuropathy. Peroneal nerve injuries often are caused when the nerve is compressed between a leg holder and the fibular head. Less common obstetric palsies include those of the femoral and obturator nerves.

Most maternal obstetric palsies are neurapraxic and resolve within 6 weeks. In future pregnancies, women with recurrent craniopelvic disproportion, dystocia, or axonal degeneration with their initial neuropathy are candidates for cesarean delivery. Otherwise, a cautious trial of labor may be prudent.

Restless Legs

Unpleasant paresthesias (described as creeping, crawling, aching, or fidgetiness) localized deep within both legs affect 10% to 27% of pregnant women. Usually they begin 30 minutes after the patient lies down and occur mainly in the last trimester. An irresistible desire to move the legs accompanies the discomfort. Symptoms resolve sharply during the first month postpartum, after which time about 5% of women remain affected. A study found lower average hemoglobin and mean corpuscular volume than in healthy subjects (Manconi et al., 2004). This same study describes gestational worsening in about 60% of patients reporting restless legs syndrome (RLS) before pregnancy, but improvement during pregnancy in some 12% of preexisting RLS. Approximately 80% of patients complaining of restless legs experience periodic movements of sleep (see Chapter 68). These stereotyped flexion movements of the legs during non–rapid eye movement sleep may awaken the patient, leading to sleep loss and excessive daytime somnolence. Caffeine ingestion, uremia, alcohol use, iron deficiency, hypothyroidism, vitamin deficiency, rheumatoid arthritis, peripheral neuropathy, and medications are important, if only occasional, associated factors. The importance of iron and folic acid supplementation remains unclear. Folic acid may be of benefit in treating restless legs during pregnancy. Anecdotal reports suggest a benefit from vitamin E, vitamin C, and magnesium supplements. Electric vibrators, stretching, walking, decreased activity, and massage also may be helpful. The use of dopaminergic agonists and levodopa to treat periodic limb movements of sleep during pregnancy has not undergone systematic study. Anecdotal reports indicate success and safety with levodopa. When a physician decides to offer either of these therapies, careful disclosure of the potential for known and also unforeseen risks to the fetus may allow the patient to make an informed decision.

Chorea Gravidarum

Chorea of any cause beginning in pregnancy is termed chorea gravidarum. Historically, rheumatic heart disease (often fatal) was associated with most cases, but rheumatic heart disease has virtually disappeared. Today, the disease linked most strongly with this condition is the antiphospholipid antibody syndrome, with or without SLE. Additional etiologies include tardive dyskinesia due to neuroleptics, hyperthyroidism, Wilson disease, Huntington disease, vascular disease, other hypercoagulable states, and a variety of intoxications.

Chorea commonly presents during the second to fifth months of pregnancy and uncommonly may begin postpartum. Subtle, sometimes severe, cognitive change may accompany the chorea. Usually this condition resolves spontaneously within weeks to months, often shortly after delivery. Choice of therapy for this often benign condition depends on the severity of the disorder and other accompanying clinical manifestations. Expectant observation, cautious use of haloperidol, and use of corticosteroids have been successful. Oral contraceptives also have been associated with the appearance of chorea. The mechanism by which pregnancy and oral contraceptives cause chorea is unknown. Chorea may recur in subsequent pregnancies.

Huntington Disease

We lack studies that demonstrate an effect of Huntington disease (HD) on the course of pregnancy or an effect of pregnancy on the course of HD. Physicians advise women known to have HD that the risk of transmitting the illness to a child is 50%. Most patients with HD die within 15 to 20 years after a symptomatic diagnosis, their later years commonly characterized by severe emotional, intellectual, and motor decline. Patients with HD have more children than the normal population. Based on this information, authorities are divided on how physicians should counsel patients. Some suggest advising fertile women with HD to consider measures to avoid pregnancy—not for the sake of the mother but to avoid the risk of propagation of HD into future generations and to avoid the predictable effect of the mother’s death and disability on the child. For those women interested in planned pregnancy, genetic testing through amniocentesis or chorionic villus sampling allows women to consider pregnancy termination in the case of an affected fetus. Preimplantation genetic diagnosis associated with in vitro fertilization can be helpful but does not eliminate the need for amniocentesis and chorionic villus sampling to confirm the accuracy of a preimplantation genetic diagnosis. Physicians debate the ethics of genetic testing when patients at risk for Huntington disease are unwilling to discover their own personal risk through presymptomatic testing yet request testing of the embryo/fetus.

Friedreich Ataxia

Little information about pregnancy and Friedreich ataxia is available. An author notes uneventful pregnancies in a retrospective study of 17 women without known adverse outcomes and recommends cardiac examination and cardiac monitoring during labor (Mackenzie, 1986). Studies point to medications such as magnesium sulfate, muscle blockers, and barbiturates as possibly problematic in isolated case reports, with cautious use of these medications recommended (Kranick et al., 2010).

Dystonia

Pregnancy carries no known additional risk to women diagnosed with dystonia. In one review, pregnancy was considered unlikely to exacerbate dystonic movement. When needed, botulinum toxin A, used with caution, seems preferable to baclofen (Kranick et al., 2010).

Rare case reports are the basis for dystonia gravidarum. This is a syndrome reported to begin during the first trimester and resolve by the end of the second trimester. Cervical dystonia and a horizontal head tremor characterized the case reported and may have been helped to some degree by clonazepam (Lim et al., 2006b).

Parkinson Disease

Rarely, patients suffering from Parkinson disease become pregnant. No reliable studies on the effect of pregnancy on the course of Parkinson disease or the effect of the disease on pregnancy are available. Information available from case reports is contradictory but indicates a possible adverse effect of pregnancy on the symptoms. No difficulty with pregnancy outcomes is known. Available limited data best supports the cautious use of l-dopa/carbidopa and the possible avoidance of amantadine during pregnancy (Kranick et al., 2010).

Tourette Syndrome

A study of 11 pregnancies in a mixed retrospective and prospective analysis found no unequivocal effect of pregnancy on the severity of tics in Tourette syndrome. The authors acknowledge the severe limitations of the study on a topic about which little information is otherwise available. The study also indicates the importance of this information for physicians treating patients whose tics worsen during pregnancy and for whom neurological consultation is requested (Stern et al., 2009).

Wilson Disease

Small case series provide limited information on which to base treatment decisions during pregnancy for women with Wilson disease. Pregnancy is contraindicated only in the presence of severe liver disease (Brewer et al., 2000). Pregnancy does not appear to have an adverse effect on the course of Wilson disease, but lack of treatment of the condition can. Anecdotal but numerous reports of women who stop therapy during pregnancy documents disease progression and maternal death. A study from India and the largest case series on pregnancy outcomes reported 59 pregnancies in 16 patients with 30 successful pregnancies, 24 spontaneous abortions, 2 medical terminations, and 3 stillbirths (Sinha et al., 2004). Most of the adverse outcomes occurred in patients with Wilson disease untreated during pregnancy. Presymptomatic patients with Wilson disease had excellent outcomes for their earlier pregnancies. Debate over the most effective and safest medication continues. Penicillamine and trientine are teratogenic in animals; zinc acetate is not. A recent review suggests an advantage of zinc acetate (Kranick et al., 2010), as do advocates of zinc therapy (Brewer et al., 2000). In a series in which zinc acetate was given during 26 pregnancies, 1 case of microcephaly and 1 cardiac defect were noted. The authors mention 4 miscarriages they chose not to include in the study analysis (Brewer et al., 2000).

Primary Brain Neoplasms

Brain tumors of all types occur during pregnancy, but only at 38% of the rate expected in nonpregnant women of fertile age. Diminished fertility in women with these tumors may explain this reduction because pregnancy probably does not protect against the development of neoplasms. Studies show increased numbers of abortions before the symptoms of tumor appear. The effect of pregnancy on morbidity and mortality in women with brain tumors is not known, but certain tumors grow more rapidly during pregnancy (Pallud et al., 2009). Meningiomas have estrogen receptors, which may explain the enlargement usually seen during pregnancy. The rupture rate of spinal hemangiomas increases with the duration of gestation. Postpartum remission of symptoms of meningioma, vascular tumors, and acoustic neuromas may be due to tumor shrinkage.

Gestational brain tumors tend to present at different stages dependent on tumor type: gliomas usually present problems during the first trimester, meningiomas during and after the second trimester, and vascular tumors in the third trimester.

Malignant tumors or tumors threatening compression of vital brain structures usually require surgery during pregnancy. During pregnancy, most neurosurgical procedures appear well tolerated (Cohen-Gadol et al., 2009). Surgery for some benign tumors can wait several weeks postpartum to observe for spontaneous improvements. Babies of most women with brain tumors deliver by cesarean section. Vaginal delivery is reserved for patients whose tumor would not pose a threat of herniation with the shifts of intracranial pressure associated with labor. Pregnancy interruption is considered when increased intracranial pressure, vision loss, or uncontrolled seizures develop as a result of the tumor. Some neurosurgeons presume the safety of intracranial carmustine implants to treat high-grade glioma during pregnancy, despite the lack of empirical evidence on safety for the fetus and known teratogenicity of systemic carmustine (Stevenson and Thompson, 2005).

Administration of corticosteroids commonly lessens symptoms of brain tumors (see Chapter 52D), but fetal hypoadrenalism may result from their use. Physicians usually defer potentially teratogenic chemotherapy until after delivery. Cranial radiation therapy during pregnancy may be helpful to the mother, but no dose of radiation is completely safe for the fetus. The fetus usually is seriously affected when it receives doses greater than 0.1 Gy (10 rads), which may cause growth retardation, microcephaly, and eye malformations. The fetus may also be affected by lower amounts of radiation, particularly early in gestation. Researchers estimate that in utero exposure to 0.01 to 0.02 Gy of radiation increases the incidence of leukemia by 1 case per 6000 exposed children. Estimates of the fetal dose during radiation for brain tumors range from 0.03 to 0.06 Gy. One study suggests that alternative positioning of the patient may reduce fetal exposure to as little as 0.003 Gy when a dose of 30 Gy is delivered to the brain (Magne et al., 2001).

Pituitary Tumors

Women with untreated hyperprolactinemia often are anovulatory and infertile. Treatment with dopamine agonists restores ovulation in 90% of patients. During pregnancy, medical therapy focuses on preventing complications of tumor growth.

Bromocriptine reduces prolactinoma size, usually within 6 weeks to 6 months. Although this drug has no demonstrated teratogenic potential, some investigators recommend discontinuation of the medication unless clearly needed during pregnancy. Bromocriptine suppresses lactation, and anecdotal reports of puerperal maternal hypertension, seizures, stroke, and cerebral angiopathy are in the literature. Data on other dopamine agonists during pregnancy are limited.

The normal pituitary gland and most pituitary tumors grow during pregnancy. The woman with a pituitary microadenoma (<10 mm) may be reassured that fewer than 5% of these tumors grow enough to become symptomatic. The risk for a macroadenoma becoming symptomatic ranges from 16% to 36% but is considerably less for patients who receive radiation or surgical therapy before pregnancy. Commonly, physicians advise women with macroadenomas to have transsphenoidal surgery before attempting pregnancy or to receive bromocriptine therapy during pregnancy. Check visual fields and acuity monthly. Monitoring of prolactin levels is not helpful. MRI is indicated after delivery and should be performed if symptoms are increasing. For women with a symptomatic macroprolactinoma diagnosed during pregnancy, therapeutic options include bromocriptine therapy, pregnancy termination, or surgery.

Usually, women with pituitary tumors deliver vaginally. Studies have not demonstrated tumor growth associated with breastfeeding. Pituitary apoplexy may present within days, weeks, or occasionally years after delivery. Uncommonly, a pituitary mass presenting in late pregnancy or up to 1 year postpartum may be lymphocytic hypophysitis. Some researchers find MRI to be helpful in establishing the diagnosis and possibly reducing the need for neurosurgery (Gutenberg et al., 2009).

Choriocarcinoma

Cerebral metastases are a common manifestation of this rare tumor of trophoblastic origin. The tumor metastasizes first to the lung and then from lung to brain. This often happens months after a molar pregnancy or abortion. Approximately 15% of tumors follow normal pregnancies, but most are discovered after pregnancies characterized by spontaneous abortion or by vaginal bleeding, premature labor, and an enlarged uterus due to a molar pregnancy. Women with such cerebral metastases present with seizures, hemorrhage, infarction, or gradually progressive deficits. The tumor may invade the sacral plexus, cauda equina, or spinal canal. A ratio of serum to cerebrospinal fluid chorionic gonadotropin of less than 60 suggests the presence of choriocarcinoma brain metastasis. Chemotherapy, radiation, and surgery have yielded successful results following early diagnosis.

Idiopathic Intracranial Hypertension (Pseudotumor Cerebri)

Idiopathic intracranial hypertension (IIH) (see Chapter 59) usually worsens with pregnancy. Some researchers advise a delay in pregnancy until all signs and symptoms of preexisting IIH abate. Termination of pregnancy is of unknown value and is not indicated. Healthy babies usually result regardless of whether IIH begins before or during pregnancy.

Commonly, IIH develops during the 14th gestational week and disappears after 1 to 3 months, but it sometimes persists until the early puerperium. Typically these women are obese and gain weight rapidly with pregnancy. Brain imaging is normal. Protein concentration may be slightly low in otherwise normal spinal fluid.

Frequent checks of optic fundi, visual acuity, and visual fields are recommended to monitor the condition and the results of treatment. Initial cerebrospinal fluid pressures that exceed 350 mm H₂O usually indicate more severe disease. Careful studies of the effectiveness of treatment are unavailable. Most physicians advise moderation in diet to reduce weight gain. Two-week courses of corticosteroids, most commonly dexamethasone or prednisone, are added for vision loss. Four to six serial lumbar punctures should be performed before considering optic nerve sheath fenestration or lumboperitoneal shunting.

The use of acetazolamide remains controversial; human studies are inadequate to determine its efficacy or teratogenic potential. Nevertheless, acetazolamide has been used to treat IIH during many pregnancies productive of healthy infants. Some physicians recommend restricting its use until after 20 weeks’ gestation. A retrospective study of 12 patients taking acetazolamide 500 mg twice a day resulted in normal children (Lee et al., 2005).

Acetaminophen with or without codeine may improve headache. More aggressive therapy usually is reserved for vision loss. Adequate pain control during labor may decrease expected rises in intracranial pressure. Usually these patients can undergo vaginal delivery with epidural analgesia. Recurrence of IIH in a subsequent pregnancy is unusual.

Maternal Considerations

Women with epilepsy have approximately 15% fewer children than expected. Reasons offered for this decrease in fertility include social effects of epilepsy, menstrual irregularity, the effect of some antiepileptic medications on the ovaries, and an effect of seizures on reproductive hormones. In an Indian registry-based study, 38.4% of women with epilepsy were infertile. Researchers identified age, lower education, and polytherapy with antiepileptic medications as risk factors (Sukumaran et al., 2010).

Convulsive seizures during pregnancy can result in blunt trauma to the mother. Trauma is the leading nonobstetric cause of maternal death in women with epilepsy, but the incidence is very low. The effect of pregnancy on seizure occurrence can be predicted from the control of epilepsy during the 9 months preceding gestation. The fewer seizures there are in the 9 months before conception, the lower the risk for worsening during the pregnancy. Women who have at least one seizure in a month can be expected to have more seizures during pregnancy. Women who have less than one seizure in 9 months usually do not experience an increase in seizure rate during pregnancy.

Many studies suggest that approximately one-fourth of women experience an increase in seizure rate during gestation and that somewhat less than that experience fewer seizures. Researchers postulate that for some women, increased seizure frequency may result from lowered levels of circulating unbound antiepileptic drugs (AEDs). Pregnancy alters protein binding of many AEDs and increases the volume of distribution and metabolism of many drugs. However, even when blood levels of drugs are maintained adequately, approximately 10% of women experience worsened seizure control during pregnancy. During labor, approximately 1% to 2% of epileptic women have a convulsive seizure, and another 1% to 2% have a seizure within 24 hours of delivery. Other factors that may theoretically contribute to a possible increase in seizure rate include hormonal changes, sleep deprivation, mild chronic respiratory alkalosis, the use of folic acid supplements, and emotional factors. Seizure type did not play a role in some studies, but in others, complex partial epilepsy worsened more often during gestation.

A prospective Finnish study is more optimistic, suggesting that the frequency of seizures during gestation does not change or may even decrease (Viinikainen et al., 2006). The International Registry of Antiepileptic Drugs and Pregnancy (EURAP) reports that nearly 60% of women with epilepsy and careful monitoring do not have seizures during pregnancy (EURAP Study Group, 2006). A consensus statement and review of the evidence available concluded that there is insufficient evidence to determine whether gestation is associated with any predictable change in the frequency of seizures or status epilepticus reported in populations of women with epilepsy (Harden et al., 2009b).

Epilepsy does not affect the course of pregnancy to a clinically significant degree. Epilepsy either does not increase the rate of cesarean section or increases it only modestly (Kelly et al., 2009), has no effect on late-pregnancy bleeding, but may increase premature labor, premature contractions, or premature delivery to a small degree, primarily in women who smoke tobacco. While the effect of epilepsy on the rate of gestational hypertension, preeclampsia, or spontaneous abortion is uncertain, in one carefully conducted study, researchers found no increased risk for preeclampsia (Harden et al., 2009b; Viinikainen et al., 2006).

When women with epilepsy take antiepileptic drugs researchers in Norway find the risk of severe pre-eclampsia, early pregnancy bleeding, pregnancy induction, and cesarean section to be increased, approximately five fold, 6.4 fold, 2.3 fold, and 2.5 fold respectively. In this same study, forceps delivery and preterm birth were increased for those women not taking antiepileptic drugs (Borthern et al., 2011).

Fetal Considerations

Nearly 90% of epileptic women deliver healthy, normal babies. However, risks for miscarriage, stillbirth, prematurity, developmental delay, smallness for gestational age, and major malformations are increased in the offspring of epileptic mothers. Available literature suggests that there is no increase in perinatal mortality. Maternal seizures, AEDs, and socioeconomic, genetic, and psychological aspects of epilepsy affect outcome. Although AEDs may cause significant problems for the fetus, the consensus among neurologists has been that maternal seizures probably are more dangerous. Convulsive seizures cause fetal hypoxia and acidosis and are associated with the potential for blunt trauma to the fetus and placenta (Holmes et al., 2001). Fetal heart rate slows during and for up to 20 minutes after a maternal convulsion, which suggests the presence of fetal asphyxia. The child of an epileptic mother experiencing convulsions during gestation is twice as likely to develop epilepsy as the child of a woman with epilepsy who does not have a convulsive seizure during gestation. A retrospective study identified five or more generalized convulsive seizures during gestation as an independent risk factor for lowered verbal IQ scores in children (Adab et al., 2004). Pregnant women may be reassured that current data do not indicate an increased risk for malformation to their fetus from an uncomplicated single seizure during the first trimester.

A prospective study in which clinicians physically examined children of epileptic mothers and controls (Holmes et al., 2001) reported the frequency of major malformations (structural abnormalities of surgical, medical, or cosmetic importance not including microcephaly, growth retardation, or hypoplasia) to be 1.8% in their normal control population. With one AED (phenytoin, carbamazepine, or phenobarbital) this rate rose to 3.4% to 5.2%, and with two or more to 8.6%. Some women with a history of seizures did not take anticonvulsants during pregnancy and had a rate of major malformation statistically the same as the control population at 0%. Women who suffered seizures during the first trimester and were taking their anticonvulsant drugs had a rate of major malformation of 7.4% to 7.8%. This carefully gathered information lays blame for teratogenesis primarily on the use of AEDs.

Studies of various design suggest that valproic acid causes a higher rate of teratogenicity than other commonly used AEDs. Increased risk may be related to increasing dose of valproic acid and was 9.1% for higher valproic acid doses in one study (Morrow et al., 2006). Higher-dose lamotrigine (>200 mg/day) also possibly increased teratogenicity up to 5.4% (Brodie, 2006). Based on these trends, some researchers have suggested caution in the use of AEDs, particularly valproic acid, in treating fertile women with epilepsy and avoidance of this medication during the first trimester if possible.

AEDs with known teratogenic effect exert their most serious effects during the first 2.5 months of gestation. Change medication before or during the first trimester to be maximally useful. The neural tube closes between 3 and 4 weeks. Cleft lip and palate occur with exposure before 5 and 10 weeks, respectively, whereas congenital heart disease due to anticonvulsant exposure occurs before 6 weeks’ gestation.

A syndrome described initially as fetal hydantoin syndrome—including midfacial hypoplasia, long upper lip, low birth weight, cleft lip and palate, digital hypoplasia, and nail dysplasia—occurs with carbamazepine, primidone, and valproic acid and is more accurately called fetal anticonvulsant syndrome. Some minor anomalies usually disappear during the first years of life. Investigators speculate that midfacial hypoplasia associated with hypoplasia of the facial bones could be a marker for cognitive dysfunction. Trimethadione has such a high teratogenic potential that its use during pregnancy is contraindicated and should be avoided in women who might become pregnant.

Adequate human studies of newer anticonvulsant drugs during pregnancy are lacking. The teratogenic potential of gabapentin, vigabatrin, tiagabine, zonisamide, topiramate, clobazam, levetiracetam, lacosamide, and oxcarbazepine is incompletely understood. A consensus statement on outcomes (Meador et al., 2008) concluded that the risk of fetal malformations with newer AEDs is “largely unknown, and consequences to individual offspring could be severe and lifelong.” At this time, the physician may consider reevaluating the need for these anticonvulsants in a patient planning pregnancy or substituting an agent with known potential risks.

Several studies focused on cognitive effects of AEDs employed during pregnancy. Holmes and colleagues (2000) described normal behavior in children of women with epilepsy on no AEDs during gestation, compared to matched controls. Vinten et al. (2005) found that children exposed to valproic acid in utero have significantly lower verbal IQ scores and memory function than children exposed to carbamazepine or phenytoin. While contradictions exist in published literature, the effects of carbamazepine and phenytoin may have little effect on cognition, but methodological problems hinder these analyses. A small United Kingdom pregnancy registry study compared use of valproic acid (mean dose 800 mg) and levetiracetam (mean dose 1700 mg) taken throughout pregnancy to a control group. Assessing development in children less than 2 years of age, 8% who were exposed to levetiracetam fell within the below average range, while the corresponding statistic was 40% for valproic acid, and 12% born to control mothers (Shallcross et al., 2011). Larger-scale prospective studies are needed to adequately quantify the cognitive effect of in utero AED exposure.

Common Advice and Management Strategy

The need for AED therapy should be reevaluated before conception. Once pregnancy has begun, discontinuing medications becomes more problematic. A supplement to a practice parameter from the American Academy of Neurology (AAN) states: “Although many of the recommendations in this parameter suggest minimizing AED exposure during pregnancy, for most women with epilepsy, discontinuing AEDs is not a reasonable or safe option” (Harden et al., 2009c). Whether this approach may be helpful for selected patients is not addressed.

When an AED is needed, monotherapy at the lowest effective dose is preferred. Warning the patient about the effects of sleep deprivation and noncompliance with the drug regimen may be helpful when paired with a thorough description of the potential consequences of seizures and benefits of AEDs.

Physicians often monitor and adjust serum AED concentrations with increased frequency during gestation and the postpartum period. However, we lack data demonstrating the effectiveness of such a plan. Some reviewers observe that lamotrigine levels drop more commonly during pregnancy and suggest that if there is a benefit to increased attention to AED levels, lamotrigine therapy might be most deserving of scrutiny. A review of available literature associated with a practice parameter (Harden et al., 2009a) provides a recommendation that monitoring of lamotrigine, carbamazepine, and phenytoin levels are a consideration, while monitoring levetiracetam and oxcarbazepine are discretionary.

Carbamazepine should be weaned in women with a family history of neural tube defects, particularly if there is a suitable substitute. A consensus statement published by the AAN recommended that for a fertile woman taking valproic acid as monotherapy for epilepsy, avoidance of valproic acid may be considered during the first trimester, but when valproic acid is taken as part of epilepsy polytherapy, avoidance of valproic acid should be considered (Harden et al., 2009c).

Women with epilepsy taking AEDs during gestation should be managed on an individual basis. During pregnancy and particularly after the period of organogenesis has passed, changes in medications are likely to cause more harm than good.

Women who take folic acid supplements before and during pregnancy lower their risk for delivering a child with major malformations. The use of folic acid has become routine, but recommendations vary. The Department of Health in the United Kingdom and the Centers for Disease Control and Prevention in the United States have recommended, respectively, 5 and 4 mg of folic acid daily for women who have had a child with a neural tube defect and 0.4 mg for all other women planning pregnancy. Anticonvulsants inhibit the absorption of folic acid. Occasionally, folic acid lowers anticonvulsant concentrations. Some investigators suggest that 5 mg of folic acid be given daily to women treated with valproic acid or carbamazepine. Others recommend 2 to 4 mg daily for all women with epilepsy who are taking anticonvulsants, beginning as long as 3 months before conception until 12 weeks’ gestation.

In one study, women taking AEDs and a multiple vitamin supplement containing folic acid had no reduction in the risk to their infants for developing cardiovascular defects, oral clefts, or urinary tract defects compared with women who took no supplements (Hernandez-Diaz et al., 2000). One detailed review found “insufficient published information to address the dosing of folic acid and whether higher doses offer greater protective benefit” (Harden et al., 2009a). These reviewers concluded that offering women with epilepsy supplements of at least 0.4 mg of folic acid before pregnancy “may be considered.” A review of similar data recommended folic acid 0.4 to 0.8 mg for all women capable of or planning pregnancy (U.S. Preventive Services Task Force, 2009).

For patients taking AEDs during pregnancy, a second-trimester high-resolution ultrasound evaluation helps exclude spina bifida aperta, cardiac anomalies, and limb defects. When results of the ultrasound scan are inconclusive, consider amniocentesis and obtain α-fetoprotein and acetylcholinesterase concentrations.

A deficiency of vitamin K–dependent clotting factors occurs in some neonates born to women who take phenobarbital, primidone, carbamazepine, ethosuximide, or phenytoin. Although rarely reported, neonatal intracerebral hemorrhage may be attributable to a vitamin K deficiency. In an attempt to lower this risk, some physicians prescribe oral vitamin K₁ 10 to 20 mg daily beginning 2 to 4 weeks before expected delivery and until birth. The American Academy of Pediatrics has recommended that physicians inject every newborn with 1 mg of vitamin K₁ intramuscularly. When hemorrhage occurs, fresh frozen plasma corrects the hemorrhagic state acutely. A study by Kaaja and colleagues (2002) could not support the hypothesis that maternal enzyme–inducing AEDs increase the risk for bleeding in offspring. A literature review finds inadequate evidence to determine whether the newborns of women with epilepsy taking AEDs have an increased risk of hemorrhagic complications (Harden et al., 2009a). An accompanying practice parameter noted insufficient evidence to support or refute a benefit of prenatal vitamin K supplementation (Harden et al., 2009b).

Occasionally, seizures occur for the first time during pregnancy. Pregnancy has little effect on the use of diagnostic examinations and treatment considerations. The most common causes of seizures during childbearing years include idiopathic epilepsy, trauma, congenital defects, neoplasms, meningitis, intracerebral hemorrhage, and drug or alcohol toxicity. In addition, pregnancy predisposes women to certain epileptogenic conditions such as eclampsia, water intoxication, thrombotic thrombocytopenic purpura, sinus or cortical venous thrombosis, and amniotic fluid embolus. Common iatrogenic causes of epilepsy include hyponatremia due to intravenous fluid infusion during the intrapartum period, and the use of epidural or parenteral anesthetics.

A single first-onset seizure resolving within minutes usually can be managed acutely without anticonvulsants. Once the physician determines the cause for the seizure and whether further seizures are likely, the need for anticonvulsant medication can be reviewed.

No special considerations exist during pregnancy when potentially fatal generalized convulsive status epilepticus is treated. Physicians agree that prompt application of a specific treatment regimen with which the physician is familiar generally assures the best chance of success. Monotherapy with phenobarbital or lorazepam and combined therapies with phenytoin are effective.

The newborn ingests anticonvulsant drugs secreted in breast milk and may become sedated and hyperirritable. Infants may show withdrawal reactions from phenobarbital after lactational exposure. Known health benefits of breast milk probably outweigh potential subtle and theoretical effects of older AEDs on the nervous system. The World Health Organization Working Group on Human Lactation and the American Academy of Pediatrics disagree on the safety of breast milk containing ethosuximide, which may cause hyperexcitability and poor suckling. Little information is available on newer anticonvulsants concerning their secretion into breast milk and effects on the infant. The use of these drugs while nursing must be individualized based on the known benefits over alternatives.

Arteriovenous Malformations

The risk for repeat hemorrhage from a previously ruptured arteriovenous malformation (AVM) generally outweighs the risk from surgical excision or an obliterative procedure. Surgery can usually be performed shortly after the diagnosis is made and before pregnancy is considered. When performing proton beam irradiation, some authorities advise women to wait 2 years before conception. The decision concerning management (excision, embolization, or irradiation) remains controversial. No specific therapeutic recommendation is available for patients planning pregnancy.

The risk for hemorrhage from an AVM (whether unruptured or previously ruptured) rises from a low point during childhood and teenage years to a higher risk during childbearing years. Whether pregnancy poses an additional risk remains uncertain. The best retrospective review suggests that risk for hemorrhage during pregnancy resulting from an unruptured AVM may be as low as 3.5%. This is probably no different from the risk to nonpregnant women with unruptured AVMs. Multiple pregnancies do not increase the rate of hemorrhage. In the past, physicians routinely advised women with an AVM, previously ruptured or not, to avoid pregnancy. This conclusion from early retrospective studies indicated that 87% of AVMs rupture during pregnancy and that 25% to 30% of initial ruptures are fatal. Subsequent analysis contradicted these dismal estimates. Still, we have no prospective studies. The clinician must exercise caution in interpreting this information.

Women whose AVM is repaired surgically can undergo vaginal delivery. Physicians usually perform cesarean section for incompletely repaired or partially treated previously ruptured AVMs. Epidural anesthesia is preferred.

Intracranial Hemorrhage

Women presenting with pregnancy-associated stroke are as likely to have an infarct as an intracerebral hemorrhage. Compared with the nonpregnant state, intracerebral hemorrhage occurs 2.5 times more often during pregnancy and almost 30 times more often during the 6 weeks postpartum. Up to 44% of these hemorrhages are associated with eclampsia and preeclampsia. In France, nearly half of women with intracerebral hemorrhage associated with eclampsia die. Additional diagnostic considerations include bleeding diatheses, cocaine toxicity, bacterial endocarditis, sickle cell disease, moyamoya disease, and metastatic choriocarcinoma. In approximately one-third of patients who have intracerebral hemorrhage, no specific cause is uncovered. The frequency of intracranial hemorrhage and its contributing causes is different in different populations (Jeng et al., 2004).

Subarachnoid hemorrhage is a common cause of nonobstetric maternal death. Hemorrhage from aneurysms and vascular malformations account for 25% to 35% of intracranial hemorrhages. Management strategies are generally the same as those applied outside of pregnancy. Usually, definitive therapy for AVMs may be postponed until after delivery, whereas surgery or obliterative therapy for an aneurysm usually is urgent. The effects of some treatment agents for subarachnoid hemorrhage (e.g., nimodipine) on the human fetus are well studied. However, the potentially fatal consequences of vasospasm associated with subarachnoid hemorrhage make their use reasonable during pregnancy.

Anticonvulsants, specifically phenytoin, are unnecessary and ineffective in nongravid patients who have had an intracranial hemorrhage but have not had a seizure. Data during pregnancy are unavailable. Prophylactic anticonvulsants probably are best reserved for those women whose hemorrhage has caused a mass effect sufficient to increase the risk of cerebral herniation associated with sudden increase in intracranial pressure that sometimes occurs with a seizure. Anticonvulsant use is reasonable after a first convulsion.

Although many physicians recommend cesarean section for patients with gestational intracranial hemorrhage, mode of delivery did not affect outcome in the studies available. Vaginal delivery with epidural anesthesia is acceptable.

Ischemic Stroke

Most women who have a stroke before gestation complete uneventful pregnancies with excellent outcomes (Coppage et al., 2004). A Chinese study found no difference in fetal outcomes when strokes occurred during pregnancy (Kang and Lin, 2010). Women in whom emboligenic cardiac disease, SLE, antiphospholipid antibody syndrome, or other coagulopathies have the added risk for stroke during pregnancy associated with those conditions. Stroke during one pregnancy by itself is not a risk factor for stroke in subsequent pregnancies. A review of a U.S. national database found a stroke risk per 100,000 deliveries of 34.2 (9.2 for ischemic stroke, 8.5 for cerebral hemorrhage, 0.6 for CVT, and 15.9 for the category of “pregnancy-related cerebrovascular event”) excluding subarachnoid hemorrhage. This same study suggests that the risk for gestational and peripartum stroke increases among women diagnosed with migraine, thrombophilia, SLE, heart disease, sickle cell disease, hypertension, thrombocytopenia, age over 35, being African American, and patients receiving blood transfusions (James et al., 2005). Cesarean delivery and gestational hypertension place women at increased risk for stroke.

Transitory neurological symptoms and preeclampsia are common during the pregnancies and puerperia of women with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL). For these women, there is no difference in the miscarriage rate, mode of delivery, neonatal birth weight, or Apgar score (Roine et al., 2005).

Two-dimensional echocardiography may be the test of greatest importance when evaluating a woman with gestational stroke. CT and selective angiography are associated with a small risk to the fetus and require adequate shielding and hydration. The lack of X-irradiation may make magnetic resonance angiography preferable to selective angiography in the diagnosis of aneurysm, AVM, arteritis, venous thrombosis, or vasospasm (see Imaging, earlier).

Eclampsia and preeclampsia are associated with approximately one-fourth of pregnancy-related infarcts in the United States and approximately half of such infarcts in France. In 25% to 35% of patients suffering ischemic stroke, the cause remains unclear. In the remainder, the stroke is symptomatic of a systemic illness such as premature atherosclerosis, hypertension, cardiac disease, hyperlipidemia, diabetes, arterial dissection, Takayasu disease, vasculitis, antiphospholipid antibody syndrome, SLE, sickle cell disease, thrombotic thrombocytopenic purpura, CVT, coagulopathies, tobacco, cocaine, and other drug use. Etiologies of stroke unique to pregnancy include choriocarcinoma, postpartum cerebral angiopathy, and postpartum cardiomyopathy. Stroke during labor or shortly after vaginal delivery may result from an amniotic fluid embolus. Air embolus to the heart, with subsequent generalized and focal cerebral ischemia, occurs with vaginal air insufflation during abortions performed outside medical standards, obstetric procedures including cesarean section, and orogenital sex.

The use of aspirin at a low dose (60-150 mg/day) throughout pregnancy is demonstrably safe in the second and third trimesters. No longitudinal studies have confirmed the efficacy of aspirin in preventing stroke during pregnancy. The safety and efficacy of clopidogrel and ticlopidine are unknown during pregnancy, and their use is not advisable.

Thrombolytic therapy for stroke during pregnancy seems unimpressive and has theoretically significant but unexplored potential risks for the fetus (Murugappan et al., 2006). Researchers trumpet a first case of clinical improvement associated with use of a thrombolytic agent (urokinase) during the immediate postpartum period (Méndez et al., 2008).

Guidelines from the AAN on the use of thrombolytics (Report of the quality standards subcommittee of the American Academy of Neurology, reaffirmed February 18, 2003, last accessed 2010) do not recommend a specific time interval after cesarean section or vaginal delivery before considering the use of thrombolytics to be safe. Relevant guidelines within this protocol suggest waiting 2 weeks after a surgical procedure and 3 weeks for areas where compression to stop bleeding might be difficult (urinary or gastrointestinal hemorrhage).

Cardiac Disease and Anticoagulation

Prosthetic heart valves are associated with embolization; rates depend on whether the valves are aortic, mitral, bioprosthetic, or mechanical. Mechanical valves have the highest rate. Bioprosthetic valves are associated with better pregnancy outcomes, but some researchers postulate that gestation accelerates the natural rate of calcification and degeneration of these prostheses, ultimately leading to valve failure. We lack convincing studies that demonstrate an effect of pregnancy on these valves.

Attempts to resolve the most pressing issue of anticoagulation in women with mechanical heart valves have been unsatisfactory. The reported risk for thromboembolism of some mechanical valve prostheses during pregnancy is 25% to 35%, much higher than the annual risk in the nonpregnant state of about 1.25% to 5.4%. In the nonpregnant state, target values for warfarin therapy and the use of aspirin 80 to 100 mg/day vary, depending on whether the valves are caged ball, tilting disk, or bileaflet. Aortic or mitral valvular location, the presence of atrial fibrillation, left atrial size, or thrombus, and previous thromboembolic episodes also influence some recommendations for anticoagulation (Stein et al., 2001). Conflicting data and incomplete studies impede attempts to translate these recommendations for the pregnant patient.

Warfarin is teratogenic, with maximal effect during the first 6 to 12 weeks of gestation, and is associated with a high rate of fetal loss, congenital malformations, and mental and physical disability. Experts postulate that serious complications of warfarin use during the second and third trimesters are caused by microhemorrhages within the brain. Included among these are dorsal midline dysplasia such as corpus callosum dysgenesis/Dandy-Walker malformation or ventral midline dysplasia with optic atrophy, mental retardation, developmental delay, seizures, and microcephaly. Approximately 30% of living progeny suffer fetal embryopathy. Fetal warfarin syndrome includes nasal hypoplasia/stippled epiphyses, limb hypoplasia, low birth weight, hearing loss, and ophthalmic anomalies. One study found that women taking more than 5 mg of warfarin daily had an increased likelihood of fetal complications—an effect independent of the extent of anticoagulation as measured by the international normalized ratio. Insignificant quantities of warfarin appear in breast milk.

One systematic review and pooling of literature on fetal outcome for women with mechanical heart valves suggests that warfarin embryopathy occurs in 6.4% of live births when using warfarin throughout pregnancy (Chan et al., 2000). A logical therapeutic alternative to warfarin during pregnancy is heparin. Heparin does not cross the placenta and is not associated with teratogenic effects. The substitution of unfractionated heparin for warfarin at or before 6 weeks’ gestation eliminates the risk for warfarin embryopathy but increases the risk for thromboembolic complications over warfarin alone. Regardless of anticoagulant therapy, maternal mortality is approximately 3%, and fetal wastage is 16.3% to 44.4%. Additional carefully controlled studies are needed to clarify the validity of these data.

Some investigators recommend counseling against pregnancy for patients with mechanical heart valves who require warfarin. When a patient is pregnant, aggressive use of low-molecular-weight heparin (LMWH) or unfractionated heparin can be used throughout pregnancy. Other investigators suggest substituting unfractionated or LMWH for warfarin during 6 to 12 weeks’ gestation. During the 13th week, warfarin may be introduced and heparin discontinued. At the middle of the third trimester or at about 36 weeks, discontinue warfarin. Heparin may be given optionally until just before early induction of labor or cesarean section. Some European experts, based on a belief that the risks of warfarin are overstated, recommend use of warfarin throughout pregnancy (Verstraete et al., 2000). In the United States, the U.S. Food and Drug Administration categorizes warfarin as possessing fetal risk “which clearly outweighs any possible benefit to the patient (USFDA Category X).”

On the basis of limited evidence, a consensus statement from the American Stroke Association and the American Heart Association recommends that: “For pregnant women with ischemic stroke or TIA and high-risk thromboembolic conditions such as hypercoagulable state or mechanical heart valves, the following options may be considered: adjusted-dose UFH throughout pregnancy, for example, a subcutaneous dose every 12 hours with monitoring of activated partial thromboplastin time; adjusted-dose LMWH with monitoring of anti-factor Xa throughout pregnancy; or UFH or LMWH until week 13, followed by warfarin until the middle of the third trimester and reinstatement of UFH or LMWH until delivery (Class IIb; Level of Evidence C). In the absence of a high-risk thromboembolic condition, pregnant women with stroke or TIA may be considered for treatment with UFH or LMWH throughout the first trimester, followed by low-dose aspirin for the remainder of the pregnancy (Class IIb; Level of Evidence C),” (Furie et al., 2011).

Like unfractionated heparin, LMWH does not cross the placenta and offers additional benefits including reduced incidence of heparin-induced thrombocytopenia, osteoporosis, and bleeding complications. No blood test is required to monitor its safety. Relative safety has been demonstrated in small studies that include women with antiphospholipid antibody syndrome and active lupus disease, but clear indications for the use of these preparations are still under investigation. Each variety of LMWH has its own therapeutic level and properties. Case reports document thromboembolic complications in women with mechanical valves who receive certain preparations of LMWH during pregnancy.

The frequency of cerebral embolism associated with chronic atrial fibrillation during pregnancy is 2% to 10%. When atrial fibrillation is associated with cardiac disease, experts recommend anticoagulation throughout gestation, commonly with high-dose subcutaneous heparin.

Antiphospholipid Antibody Syndrome

Women with circulating antiphospholipid antibodies and without a history of pregnancy loss do not require treatment to prevent stroke during pregnancy. Successful pregnancies without treatment are common. Women with very high antibody titers, habitual first-trimester abortion, a single miscarriage in the later trimesters, or a symptomatic antiphospholipid antibody syndrome (APS), particularly with previous stroke, usually receive treatment. Various studies have examined the use of monotherapy or polytherapy in widely ranging doses and combinations of aspirin, prednisone, subcutaneous unfractionated heparin (UFH), IVIG, and occasionally placebo. Studies on LMWH are optimistic but incomplete. Stroke in women with symptomatic APS may occur at doses of heparin sufficient to produce a usually therapeutic serological anticoagulation effect. Some researchers advise women seriously ill with APS to avoid pregnancy.

The broad range of therapies offered as consensus recommendations for these patients, including those who have suffered stroke, underline the need for additional investigation in this area (Bates et al., 2004; Lim et al., 2006a; Tincani et al., 2004). Women with antiphospholipid antibodies and monosymptomatic habitual pregnancy loss should receive subcutaneous high-dose UFH together with low-dose aspirin until 34 weeks’ gestation. A combination of low-dose aspirin and LMWH during pregnancy appears to be current practice in the United Kingdom (Shehata et al., 2001) for women with antiphospholipid antibodies and a history of fetal loss after 16 weeks’ gestation, intrauterine growth retardation, early-onset preeclampsia, placental abruption, or stillbirth.

Postpartum Stroke

Debate continues over classification of syndromes described as postpartum cerebral angiopathy, delayed peripartum vasculopathy, reversible posterior leukoencephalopathy, reversible cerebral vasoconstriction syndrome, or postpartum stroke. These patients present with puerperal focal neurological signs and symptoms, often with headache, and have hypertension without edema or proteinuria. Brain MRI scanning depicts ischemia primarily in the parieto-occipital region, and angiography commonly demonstrates vasospasm. The course often is benign, but permanent deficits may occur. When these occurrences include headache, altered sensorium, seizures, or visual loss without hemorrhage, consider the possibility of a reversible posterior leukoencephalopathy syndrome (RPLS). Other authors might suggest the diagnostic consideration of a reversible segmental cerebral vasoconstriction (RSCV) syndrome (see Chapter 51F), particularly when vasospasm demonstrated by arteriography involves large blood vessels within the circle of Willis (Singhal et al., 2009). Patients have been treated with calcium channel antagonists, corticosteroids, and antihypertensive medication. Some researchers believe postpartum stroke is a form of eclampsia/preeclampsia and suggest relaxation of requirements for diagnosis of eclampsia/preeclampsia.

Primary angiitis of the central nervous system presenting with stroke is less commonly associated with pregnancy than the cerebral vasoconstriction syndromes noted. It is associated with milder headache or none at all, but with abnormal cerebrospinal fluid in 80% to 90% of patients with pleocytosis and elevated protein. Confirmation is through biopsy and less commonly through diagnostic arteriography, which may have difficulty visualizing the disease of the small blood vessels that characterizes this condition. Prompt treatment with cyclophosphamide and prednisone is recommended (Birnbaum et al., 2009).

In one case series, Witlin et al. (2000) characterized postpartum stroke as an uncommon and unpreventable complication of pregnancy. After excluding trauma, neoplasm, infection, or eclampsia, they described women suffering ischemic or hemorrhagic stroke without specific warning; hemorrhage may follow initial ischemia with injury to blood vessel walls. In general, these events were most common around the eighth day after delivery (range 3-35 days) and were associated with cesarean delivery. Seizures in half of these patients, an increase in mean arterial pressure to 1.5 times above baseline, and headache heralded the onset. Two of 20 patients died of severe intracerebral hemorrhage. Included in this series were patients with CVT and a ruptured AVM. The investigators postulated that the hypercoagulable or thrombophilic state of pregnancy may emerge as a major risk factor, perhaps interacting with underlying coagulopathies.

Cerebral Venous Thrombosis

Aseptic thrombosis of the cerebral venous system, in its most obvious clinical state, presents with puerperal headache that worsens over several days, a change in behavior or personality, convulsive seizures, and neurological deficits. The patient may be emotionally regressed, anxious, or lethargic, with mild to obvious neurological signs and (occasionally) papilledema. Initial symptoms generally begin 1 day to 4 weeks postpartum and peak in occurrence at 7 to 14 days postpartum.

CVT has been associated with hypercoagulable states, infection, sickle cell disease, dehydration, and ulcerative colitis, in addition to gestation. Women diagnosed with peripartum CVT have no known increased risk for CVT in subsequent pregnancies (Mehraein et al., 2003). Differential diagnoses include eclampsia, meningitis, and cerebral mass. A large Swedish cohort study indicated that CVT was the most common cerebrovascular disorder of women in the period between 2 days before and 1 day after delivery. The standardized incidence rate ratio increased about 115-fold for CVT, 95-fold for intracerebral hemorrhage, 47-fold for subarachnoid hemorrhage, and 34-fold for cerebral infarction. Rates between 2 days post delivery and 6 weeks postpartum continued elevated: 15-fold for CVT, 12-fold for intracerebral hemorrhage, 8.3-fold for cerebral infarction, and 1.8-fold for subarachnoid hemorrhage (Salonen Ros et al., 2001).

Brain MRI scanning with magnetic resonance venography is the initial imaging procedure of choice. Although it detects occlusion of major sinuses with high sensitivity, when smaller veins are involved, detection may be more difficult. The MRI may show multiple small infarcts involving the gray and white matter, sometimes with minor hemorrhage. More sensitive CT venography may be considered in some patients when potential benefits outweigh the risks of radiation and intravenous contrast. Digital subtraction and selective angiography/venography remain additional options.

Geographic location influences the frequency of CVT. India reports a high rate of puerperal CVT, estimated at 400 to 500 per 100,000 births. This high rate is primarily attributable to dehydration and has a predilection for women delivering at home. Incidence in the United States is comparatively low, approximately 9 per 100,000 deliveries. CVT associated with pregnancy is relatively more benign than that occurring without pregnancy. Researchers in Mexico describe a mortality rate of approximately 10% for gestational CVT and 33% for CVT not associated with pregnancy. In the United States, estimates of the mortality rate of CVT from all causes suggest that approximately 1 of 10 patients dies. However, death did not occur in a national survey of 4454 patients with peripartum CVT in the United States. Cesarean section and age older than 25 are risk factors for CVT, while preeclampsia and eclampsia are not (Cantu-Brito et al., 2010).

Researchers find antiphospholipid antibodies and diminished activity of protein S in women with gestational and puerperal CVT. Occasionally, multiple defects in coagulation occur in the same woman. Current theory suggests that levels of proteins such as C4b-binding protein increase during pregnancy and the puerperium, creating a hypercoagulable state. Some women have conditions that predispose them to hypercoagulation during pregnancy, such as activated protein C resistance (factor V Leiden mutation), protein S deficiency, or antithrombin III antibodies. Homocystinuria (hyperhomocysteinemia) may increase risk.

Heparin anticoagulation may be of benefit, particularly if the patient has concurrent thrombophlebitis in the pelvis and legs. Some observers recommend heparin therapy only when clinical indicators suggest that a poor prognosis is likely. Therapy with antifibrinolytic agents is controversial. Thrombolysis of an occluded dural sinus with urokinase or with tissue plasminogen activator via selective venography has been used as initial therapy without significant deterioration. Physicians performing this procedure claim therapeutic success, even for some patients with cerebral hemorrhage. Heparin and antifibrinolytic therapies are relatively contraindicated when the patient has had an intracranial hemorrhage. Long-term anticonvulsant therapy usually is unnecessary.