Clinical manifestations

Presentation and course

All types of stroke, including ischemic stroke (IS), intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), dural sinus thrombosis, and cerebral venous thrombosis (CVT), can be seen during pregnancy and puerperium. The neurologic findings vary with the brain region affected and may include weakness, numbness, clumsiness, imbalance disturbance of vision, language, cognition, alertness, etc., or a combination of these.

In contrast, preeclampsia and eclampsia are unique to pregnancy. Along with pregnancy-induced hypertension and gestational diabetes, they are considered risk factors for cardiovascular disease (171; 101).

Preeclampsia is defined as the development of elevated systolic blood pressure (SBP) greater than 140 mmHg or diastolic blood pressure (DBP) greater than 90 mmHg, measured on 2 occasions at least 4 hours apart, associated with proteinuria (300 mg/24h) after the twentieth week of pregnancy in a woman previously normotensive. In absence of proteinuria, preeclampsia is diagnosed as hypertension associated with new onset of any of the following: thrombocytopenia (< 100,000/microliter), elevated liver enzymes more than twice the normal concentration, elevation of creatinine more than twice normal or greater than 1.1 mg/dL, pulmonary edema, new onset headache, or visual disturbances. Preeclampsia with severe features occurs when systolic or diastolic pressures are greater than 160 or 110 mmHg, respectively, and is associated with end organ damage as described above. In this situation, the confirmatory measurement should be performed after a few minutes to expedite treatment. When seizures or coma develop, the term eclampsia is applied. The neurologic deficits are usually reversible.

The primary lesion associated with preeclampsia is posterior reversible encephalopathy syndrome (PRES). It is characterized by headache, encephalopathy, visual disturbances, and seizures associated with reversible vasogenic edema seen on CT or MRI (81).

Postpartum angiopathy or reversible cerebral vasoconstriction syndrome (RCVS) is seen in pregnant women, sometimes with preexisting preeclampsia, and occasionally mimics eclampsia due to the associated seizures (62; 65). These patients usually complain of multiple thunderclap headaches, seizures, confusion, visual changes, focal neurologic deficits, and coma.

Preeclampsia may be associated with HELLP syndrome (hemolysis, elevated liver function, and low platelets). This consists of generalized edema, right upper quadrant or epigastric pain, and nausea and vomiting that are typically worse at night (17). The focal neurologic deficits are related to intracerebral hemorrhage (Martinez et al 1998; 187), severe cerebral edema (76), ischemic stroke from vasospasm (75), or carotid artery occlusion during rebound thrombocytosis (96). Subarachnoid hemorrhage without evidence of aneurysm or vasospasm on angiography may occur (169).

Cerebral vasculitis is characterized by less severe headaches and focal neurologic deficits due to ischemic stroke or intracerebral hemorrhage.

Prognosis and complications

Preeclampsia occurs in approximately 3% to 3.5% of pregnancies and 2.6% of these may develop eclampsia. The relative risk of death within 12 months following preeclampsia and eclampsia is 5.1 compared to normotensive women (180), increasing the cost of care between 40 to 100 times a term pregnancy (172).

In the United States, analysis of a Nationwide Inpatient Sample for the years 2000 to 2001 showed a mortality rate of stroke of 1.4 per 100,000 deliveries. Twenty-two percent of survivors were discharged to another facility (88). In the UK, the mortality rate was 0.3 per 100,000 deliveries, and the case fatality rate was 20% of all strokes and 50% of hemorrhagic strokes (167). Subarachnoid hemorrhage may account for 5% to 10% of nonobstetric deaths in pregnant women (60). The fetal death rate was 12%, and 35% of infants delivered prematurely (171).

Aside from obstetrical considerations, the complications of stroke in pregnancy are similar to those seen in nonpregnant patients. In the short term, the most serious risk is herniation from cerebral edema. Seizures, status epilepticus, aspiration pneumonia, sepsis, decubiti, deep vein thrombosis, and pulmonary embolism contribute to poor outcomes. Long-term disability includes difficulty with childcare and return to work.

Over 5 years, the absolute risk of ischemic stroke recurrence outside pregnancy was 0.5% and rose only slightly to 1.8% during pregnancy and the puerperium. About half of the patients with a history of ischemic stroke received antiplatelet therapy during at least part of their subsequent pregnancy. Most women with a history of cerebral venous thrombosis received no therapy during pregnancy and heparin in the postpartum period (107).

Cerebral venous thrombosis has a good long-term prognosis. In a pooled systematic review of 66 patients presenting with cerebral venous sinus thrombosis, 59% had modified Rankin scale (mRS) of 0, and 94% had mRS of 0 to 2 at follow-up. Headache alone was associated with excellent outcome, whereas obtundation and coma with poor outcome (94).

A pooled analysis of 13 studies showed that the risk of pregnancy-related venous thrombosis is low. The relative risk of noncerebral venous thromboembolism is 16-fold higher (2.7%) and the recurrence of cerebral venous thrombosis is 80-fold higher (0.9%) than in the general population. The rate of miscarriage is similar to the general population (Aguiar et al 2016). Although counseling is important, a history of pregnancy-related stroke should not be a contraindication for subsequent pregnancy.

In a retrospective cohort of 70 patients with PRES, 94% had altered consciousness, 81% had seizures, and 14% had ischemic stroke or intracerebral hemorrhage. Although most patients (56%) had good recovery at 90 days, 16% died and 37% had marked functional impairment. However, the patients with preeclampsia may have a better prognosis (109). A few patients with RCVS have neurologic sequelae after stroke, 9% had severe deficits, and 2% mortality (174). The prognosis of HELLP is similar to severe preeclampsia except for the hematologic variables (73).

Vascular brain lesions, arteriovenous malformations, and cavernous hemangioma occur in 5.3 per 100,000 deliveries. Their presence does not seem to carry additional risk to mother and fetus. However, 79% of deliveries were performed via caesarean section (115).

Clinical vignette

A 23-year-old woman had sudden onset of blurred vision, occipital headaches, and dysarthria 5 days after delivery following an uneventful pregnancy. On physical examination, she was normotensive. She had dysarthria, left central facial weakness, subtle left arm weakness, and an extensor plantar response on the left. CT scan of the brain was normal. MRI showed areas of restricted diffusion in the frontal parietal and temporal lobes bilaterally. MRA was normal. MRV showed a filling defect in the superior sagittal sinus. CBC and complete metabolic panel were normal. Heterozygosity of the factor V Leiden mutation was found. Anticoagulation, initially with intravenous unfractionated heparin followed by oral warfarin, was started. One month later her symptoms had completely resolved. Warfarin was continued for 6 months and then switched to aspirin. Prophylactic heparin therapy during subsequent pregnancies was recommended.

Biological basis

Etiology and pathogenesis

The Baltimore-Washington Cooperative Young Stroke Study of 1,051,113 young women showed that diabetes mellitus, hypertension, ischemic heart disease, smoking, and drug use were similar in pregnancy-related stroke compared with stroke outside pregnancy (101). Another smaller study shows that stroke during pregnancy was associated less with the typical vascular risk factors and more with cerebral venous sinus thrombosis and RCVS (126). Additionally, black and Hispanic pregnant women had higher risk of stroke than non-Hispanic whites (127).

Cervical artery dissection is uncommon, but occurs more frequently in the postpartum period. In a small case series and literature review, the only risk factor found was advanced maternal age (97).

The pregnancy-related conditions associated with stroke are cesarean delivery; fluid, electrolyte, and acid-base disorders; and hypertension. For intracranial cerebral venous thrombosis, these conditions include cesarean delivery, hypertension, infections other than pneumonia and influenza (108), preeclampsia, greater parity, and cesarean delivery (153).

Complications of pregnancy leading to stroke are preeclampsia, eclampsia, postpartum infection, postpartum hemorrhage, blood transfusions (88), and gestational diabetes (111). The risk was highest during the third trimester, around delivery, and up to 6 weeks postpartum (Salonen et al 2001).

Additionally, peripartum infection is associated with increased 30-day postpartum readmission rate for ischemic but not hemorrhagic stroke (125).

In the United States, approximately 3.4% of pregnancies develop preeclampsia (07). Major risk factors include history of preeclampsia, hypertension, diabetes mellitus, antiphospholipid syndrome, and obesity. Other risk factors are nulliparity, assisted reproductive technology, advanced age, and chronic kidney disease. There is also a genetic predisposition in those with a family history (146).

Inpatient data from the New York State Department of Health reveal that 0.2% of women suffering from preeclampsia had a pregnancy-related stroke. These women were more likely to have severe preeclampsia and eclampsia, infection on admission, prothrombotic state, coagulopathy, or chronic hypertension (123). However, most patients had reversible deficits caused by vasogenic edema, not by infarction (170; 163). A few patients with restricted diffusion on MRI scan and irreversible deficits can be confidently diagnosed with stroke.

Multiple pregnancy loss (> 2) and stillbirth also increase the risk of stroke and myocardial infarction even after correction for other cardiovascular risk factors (114).

Most cardioembolism is due to prosthetic heart valves or atrial fibrillation. The prothrombotic state of pregnancy, in combination with changing intrathoracic pressures during pregnancy and delivery, may increase the risk of venous thrombosis and paradoxical embolism through a patent foramen ovale (104; 68).

Dilated cardiomyopathy is rare and develops in the second trimester. Ventricular arrhythmias, heart failure, stroke, and death may occur in up to 60% of high-risk patients. Hypertrophic cardiomyopathy leads to similar complications (164). Peripartum cardiomyopathy is uncommon during the last month of pregnancy and up to 5 months postpartum. Associated risk factors include older age, hypertension, and tocolytics use during pregnancy. Only about half of women recover good cardiac function, and up to 10% may require heart transplantation (54). At 4.7 years follow-up, freedom from all causes of death was 96.7% (23).

Ischemic stroke attributed to cerebral venous thrombosis has been reported in 10 to 20 per 100,000 deliveries in the Western world, with a higher incidence in the developing countries (33). More than 80% of patients with pregnancy-related cerebral venous thrombosis had a hypercoagulable state (30). Increased blood viscosity associated with sickle cell anemia, malignancy, polycythemia, and paroxysmal nocturnal hemoglobinuria can also precipitate cerebral venous thrombosis. Additional risk factors include systemic infection, anemia, and severe dehydration and help explain the significantly higher frequency of this condition in the developing world.

Postpartum hemorrhage with prolonged arterial hypotension may cause watershed infarctions, postpartum pan-hypopituitarism, and posterior ischemic encephalopathy.

PRES may mimic ischemic stroke and requires imaging for diagnosis. PRES also contributes to the clinical picture of preeclampsia and eclampsia and may lead to seizures, status epilepticus, ischemic stroke, or intracerebral hemorrhage.

RCVS occurs in patients with preeclampsia or eclampsia, as well as in normal pregnancy and delivery (173). It was also described in patients with cervical arteries dissection (09). RCVS should be differentiated from cerebral vasculitis, a rare cause of stroke in pregnant women.

HELLP syndrome is associated with preeclampsia and may cause either ischemic stroke or intracerebral hemorrhage, depending on the number of platelets present in circulation.

Choriocarcinoma arises from fetal trophoblastic tissue and frequently metastasizes to the brain (60). This highly vascular tumor often presents as intracerebral or subarachnoid hemorrhage (58). Usually located at the gray-white matter junction, choriocarcinoma resembles hemorrhagic transformation of an embolic stroke. Intracranial aneurysms have also been described (130). Treatment is by surgical resection or chemotherapy (60; 140; 168).

Amniotic fluid embolism occurs in about 1 in 8000 to 80,000 pregnancies during or shortly after labor, cesarean section, abortion, or trauma (41). It contributes to up to 13% to 30% of maternal deaths (183). Presentation is typically with acute hemodynamic collapse associated with seizures and focal neurologic deficits caused by cerebral hypoperfusion, hemorrhage, or thrombosis. Consumptive coagulopathy develops in most patients.

Air embolism manifests as cardiovascular collapse, focal neurologic deficits, coma, and even death (134). Air can enter the venous circulation at any time during pregnancy but mostly during delivery, especially by cesarean section, following intrauterine manipulations, or orogenital sex (24; 133; 158).

Intracerebral hemorrhage has similar causes as in nonpregnant patients. Hypertension (HTN), pregnancy-induced hypertension, preeclampsia, and eclampsia are the most common causes (101). Intracerebral hemorrhage occurs more commonly during the third trimester and first 12 weeks postpartum (119). Migraine and RCVS are more commonly seen in pregnancy-associated hemorrhagic stroke than the typical cerebrovascular risk factors and underlying lesions present in nonpregnant women (124). Pathology shows fibrinoid necrosis of small penetrating vessels (151).

There is no increased risk of rupture of arteriovenous malformations (AVM) and aneurysms in pregnant versus nonpregnant women (100; 113; 43). Considering, however, the amount of exposure to pregnancy, 40 weeks per pregnancy, 6 weeks for each puerperium, and 6 weeks for each abortion, the annual hemorrhage rate was 1.3% in nonpregnant women versus 5.7% in pregnant women. Analysis for reproductive age patients (15 to 50 years) shows a rate of 1.3% versus 7.0% (143).

Also, pregnancy does not increase the risk of hemorrhage of cavernous malformations (93). Intracranial aneurysms usually bleed in the third trimester, in particular during labor and delivery. The incidence of ruptured aneurysms is 1 in 10,000 to 5 in 10,000 pregnancies. They are more likely to rupture in patients with advanced gestational age, primiparity, and hypertension (Dias et al 1990; 135).

The risk of moyamoya-related stroke during pregnancy, delivery, and postpartum period appears to be similar to or lower than in the studies of natural history previously published (138). This suggests a neuroprotective effect of pregnancy in moyamoya. In a small study of 20 patients with moyamoya disease, intracerebral hemorrhage tends to occur antepartum whereas ischemic stroke postpartum (85). In another study of 71 pregnancies in 54 women with moyamoya disease, the risk of stroke was higher before bypass surgery than after (35).

Intracerebral or subarachnoid hemorrhages are also triggered by bleeding diatheses, disseminated intravascular coagulation, drug abuse, and septic emboli.

The mechanism of stroke during pregnancy and puerperium is unclear. Hormonal shifts, more pronounced in the later stages of pregnancy and postpartum period, when stroke is most common, play a role, mediating damage to vascular tissue structure, hypercoagulable state, and hemodynamic changes. Physiologic changes during pregnancy create a slightly hypercoagulable state. Prothombotic factors I, V, VII, VIII, IX, and X as well as von Willebrand Factor, fibrinogen, and thrombin are increased and antithrombotic protein S level is decreased (78; 159; 25). Increased platelet aggregation and occasionally development of activated protein C resistance have also been noted (194).

The maternal hemodynamic changes induced by the hormonal changes include systemic vasodilatation, increased blood volume, and decreased mean arterial pressure. This is counteracted by sympathetic activation and increased cardiac stroke work. However, the vasoconstrictor response is blunted during normal pregnancy (61).

It is unclear if preeclampsia is a reaction to placental pathology or a consequence of cardiovascular disease. Preeclampsia is triggered by the presence of placenta but may occur even after delivery and is associated with long-term cardiovascular morbidity. Subclinical cardiovascular dysfunction is present before and after delivery, suggesting a more pervasive disruption of the cardiovascular system homeostasis, namely a gestational cardiorenal syndrome (72).

In early stages, poor placentation due to ineffective trophoblast invasion of the uterine wall may be caused by certain natural-killer cells and HLA-C combinations (128). Placental hypoperfusion results from inadequate remodeling of uterine spiral arteries (148). In later stages, placental ischemia leads to release of cytokines and other agents causing endothelial dysfunction (149). TNF-alpha stimulates endothelin-1, the most potent vasoconstrictor, and IL-6, the renin-angiotensin system (106). Matrix metallo-proteinases, soluble fms-like tyrosine kinase 1 (sFlt-1), soluble endoglin, and agonistic autoantibodies to the angiotensin type 1 receptor (AT1-AA) stimulate the production of endothelin-1, which seems to be the common pathway in pathogenesis of preeclampsia and a potential target for treatment (12). Hypertension is related to increased cardiac output and mild systemic vasoconstriction (42). During late pregnancy, the cardiovascular parameters are determined by the status of the systemic circulation rather than by inadequate placentation or spiral arteries remodeling (139). Glomerular endothelial dysfunction leads to renal injury, proteinuria, and hypertension (175).

An umbrella review found serum iron level, chronic kidney disease, polycystic ovary syndrome, mental stress, bacterial and viral infections, smoking, oocyte donation, obesity, and primiparity highly suggestive of an association with preeclampsia (67). In women with preeclampsia, the risk of stroke is increased by infection on admission, prothrombotic state, coagulopathy, and chronic hypertension (123).

In eclampsia, PRES is an important component of pathogenesis (26). It is caused by altered cerebral autoregulation due to endothelial dysfunction (166), leading to vasogenic edema in the vulnerable regions of the posterior brain where the sympathetic innervation is less robust. This syndrome occurs more often in presence of hypertension, but not always (11). Status epilepticus is not uncommon in these patients (63).

RCVS can develop in the puerperium in absence of preeclampsia or hypertension. Two thirds of patients develop symptoms within a week postpartum, even after a normal pregnancy (48). The etiology is not clear, although several drugs have been implicated, such as ergot derivatives, cocaine, serotonin reuptake inhibitors, and immunosuppressive drugs (89; 71; 48). Pregnancy-associated arterial intimal hyperplasia (28) and arterial spasm have been suspected. CSF studies are usually normal or show mild pleocytosis. Mild hyperplasia without inflammation was found in 1 fatal case. Brain biopsy in 1 severe case found microangiopathic inflammatory infiltrates and fibrinoid necrosis without meningeal involvement (31). Although eclampsia, PRES, and RCVS are distinct entities that can occur independently, sometimes there is an overlap between them due to common pathophysiology.

Cerebral vasculitis in pregnant women is rare; it may be isolated to the central nervous system or part of a systemic illness. Arterial stenosis due to inflammation leads to ischemic stroke or intracerebral hemorrhage. Decreased exacerbation in pregnant patients with polyarteritis nodosa (PAN) and Bechets disease (102), and a protective role of multiple pregnancies against giant cell arteritis was described (49). One patient who complained of severe headaches during cesarean section was found to have intracerebral hemorrhage and focal narrowing of the cerebral arteries that did not respond to the initial course of steroids but to a combination of corticosteroid and cyclophosphamide (196). Another woman with idiopathic granulomatous angiitis of the nervous system during pregnancy had a good outcome with an expectant approach (16). The autopsy of a fatal case of vasculitis described in a woman who developed systemic lupus erythematosus during pregnancy showed polyarteritis nodosa-like necrotizing vasculitis of the small muscular arteries and arterioles, with acute and healing lesions in the leptomeninges, brain parenchyma, and visceral organs (178).

HELLP syndrome is characterized by microangiopathic hemolytic anemia, elevated liver enzymes, and thrombocytopenia. On liver biopsy, the classic lesion is periportal or focal parenchymal necrosis in which hyaline deposits of fibrin-like material can be seen in the sinusoids. Fibrin microthrombi and fibrinogen deposits in the sinusoids in areas of hepatocellular necrosis and in sinusoids of histologically normal parenchyma are also seen (17).

Any infection on the admission day for delivery increased the risk of stroke. The risk was even higher for genitourinary infections and sepsis. This suggests that infections may contribute to peripartum stroke (122).

Placental abruption is also associated with increased risk of ischemic and hemorrhagic stroke, particularly if delivery is before 34 weeks, when accompanied by placenta ischemia and in women with more than 1 abruption (06).

A certain proportion of strokes during pregnancy and puerperium do not have a clear etiology. However, these strokes must be categorized as stroke of unknown etiology rather than caused by the pregnant state.

Epidemiology

The incidence of stroke varies with its definition, the population studied, the duration of postpartum monitoring, and epoch.

A systematic review and metaanalysis of 11 studies demonstrate a pooled rate of stroke of 30 per 100,000 pregnancies, of which 19.9 were arterial and venous thrombosis and 12.2 were hemorrhagic. The crude stroke rate for antenatal and perinatal stroke was 18.3, and it was 14.7 for postpartum stroke (179).

In the United Kingdom, a nationwide population-based study showed a stroke incidence of 1.5 strokes (0.9 ischemic and 0.6 hemorrhagic) per 100,000 deliveries. The risk factors were migraine, gestational diabetes, and preeclampsia or eclampsia (167). In another open cohort study from England of 2,046,048 women aged 15 to 49 years, nonpregnant women had an incidence rate of first stroke of 25.0 per 100,000 person-years. Antepartum, the incidence rate was lower (10.7 per 100,000 person-years) but peripartum was 9-fold higher (161.1 per 100,000 person-years). Early postpartum (first 6 weeks) incidence was 3-fold higher (47.1 per 100,000 person-years). Rates of ischemic and hemorrhagic stroke were both increased peripartum and early postpartum (14).

Analysis of National Inpatient Sample revealed that stroke occurred in 1 of 2222 pregnancy-related hospitalizations and has not changed between 2007 and 2015. During the same period, the risk factors for stroke have increased but in-hospital mortality among pregnant women with stroke has decreased from 5.5% in 2007 versus 2.7% in 2015 (53).

The risk of recurrent ischemic stroke is not increased during pregnancy, but is in the postpartum period (107). The risk of stroke recurrence during pregnancy is less than 1%, similar to the risk of women without risk factors but is higher in women with known cerebrovascular risk factors like type 1 diabetes mellitus, large artery atherosclerosis, heart failure, previous transient ischemic attack, and increasing age (144). Unfortunately, the epidemiological studies suffer not only from variability of methods used but also from the use of an outdated definition of stroke introduced several decades ago, which was only recently updated (154).

A systematic review of pregnant patients with moyamoya disease (MMD) shows that the mean gestational age at diagnosis due to stroke was 28.7 weeks, and 69.5% presented with cerebral hemorrhage. In those diagnosed with moyamoya disease postpartum, 46.6% had a stroke within 3 days of delivery, of which 78.3% were ischemic (116).

Cerebral venous thrombosis rate was 9.1 per 100,000 pregnancies in a systematic review and metaanalysis of 11 studies published between 1990 and 2017 (179). Another systematic review of women with a history of cerebral venous thrombosis shows that the absolute risk of pregnancy-related venous thrombosis is low. However, the relative risk of noncerebral venous thromboembolism is 16-fold higher and the recurrence of cerebral venous thrombosis is 80-fold higher than in general population. The rate of miscarriage is not significantly increased (03). Cerebral venous thrombosis is a more important cause of pregnancy-associated stroke in developing countries (136).

Prevention

Most pregnancy-related strokes are not preventable, as the contributing factors are usually identified after the event or are inherent to the pregnancy itself.

Screening for preeclampsia by measuring blood pressure throughout pregnancy is recommended (185). Treatment is recommended if systolic blood pressure is 160 mmHg or higher and diastolic blood pressure is 105 mmHg or higher. If lower blood pressure values are not accompanied by end organ damage, no treatment is recommended. The following antihypertensive agents may be used: labetalol, nifedipine, and methyldopa. The medications not recommended include angiotensin-converting enzyme inhibitors, angiotensin receptor inhibitors, renin inhibitors, and mineralocorticoid receptor antagonists.

Although treatment of blood pressure values of 140-159 / 90-109 mmHg does not influence the risk of fetal death, preterm delivery, or small for gestational age, it may halve the risk of subsequent severe hypertension (02) and may be considered (29). However, a meta-analysis showed that for each 10 mmHg of mean arterial pressure reduction, birth weight decreased by 145 g, regardless of the agent used (188).

Additionally, magnesium sulphate is recommended for blood pressure greater than 160/110 mmHg or severe preeclampsia intrapartum or postpartum. A Cochrane review demonstrated a prevention of eclampsia by more than 50% with parenteral magnesium sulphate versus placebo (50).

If there is a history of early-onset preeclampsia and preterm delivery in more than one pregnancy, a daily low-dose aspirin (60 to 80 mg) beginning in the late first trimester is suggested (51).

After discharge, blood pressure should be monitored at 3 days and at days 7 to 10. Women should be given instructions regarding the signs and symptoms to monitor regardless if they had preeclampsia or not.

Preeclampsia increases the risk of cardiovascular disease twice compared to uneventful pregnancy. This risk increases 8-fold to 9-fold if preeclamptic women gave birth prematurely (ie, before 34 0/7 weeks of gestation).

The California Teachers Study, a prospective cohort study including 83,749 women, showed that hypertensive disorder of pregnancy was associated with an increased risk for stroke before the age of 60, but not in aspirin users. Statin use did not modify this risk (121). Women with a history of recurrent preeclampsia or a combination of preeclampsia and preterm delivery should have yearly evaluations of blood pressure, lipids, fasting blood glucose, and body mass index (05).

Vitamin D3 deficit may be associated with increased risk of preeclampsia; however, there are no data to support supplementation (21). Calcium supplementation greater than 1 g/day during pregnancy was also found to reduce the relative risk of hypertension (82).

The recommendations for anticoagulation during pregnancy developed by the writing group of the ACCP are based on 2 scenarios: (1) high risk condition that requires anticoagulation, or (2) low risk condition that requires antiplatelet therapy.

Artificial heart valves represent one of the highest stroke risks and require anticoagulation throughout pregnancy. Anticoagulation can be achieved in 3 ways: (1) low molecular weight heparin (LMWH) or unfractionated heparin (UFH) until the 13th week with substitution of vitamin K antagonists until close to delivery when LMWH or UFH is resumed; (2) adjusted-dose UFH administered subcutaneously every 12 hours to keep mid-interval aPTT at least twice control throughout pregnancy; or (3) adjusted-dose of LMWH twice daily throughout pregnancy. If the risk of thromboembolism is very high, eg, older mitral prosthesis or history of thromboembolism, vitamin K antagonists throughout pregnancy except close to delivery when administration of UHF or LMWH is preferred. Addition of 75 to 100 mg aspirin to anticoagulation is also suggested (19).

Women requiring long-term anticoagulation for other indications should be switched as soon as pregnancy is determined to UFH or LMWH for the duration of pregnancy. A discussion of risks and benefits with the patient is important.

Because aspirin crosses placenta, its use is divided before and after the first trimester. Some but not all studies found an increased risk of birth defects with low dose aspirin (105). Most birth defects were not associated with aspirin, although there was a small to moderate risk of anophthalmia, microphthalmia, anencephaly, craniorachischisis, encephalocele, amniotic bands/limb body wall defect, and pulmonary valves stenosis (80). However, low dose aspirin may be considered for prevention of recurrent ischemic stroke if the benefit is thought to be greater than harm. During the second and third trimesters, low dose aspirin (less than 150 mg/day) is safe for both mother and fetus (84; 37).

The safety of other antiplatelet agents, including ticlopidine, clopidogrel, and dipyridamole, has not been determined and they are best avoided during pregnancy. There is no consensus among stroke specialists on the optimal preventive therapy (79). Therefore, low-dose aspirin, UFH, LMWH or no treatment, are all acceptable approaches during the first trimester (98).

In the postpartum period, nursing mothers may safely take warfarin, but they may also take UFH (19) and even LMWH (152). During breastfeeding, low dose aspirin seems to be safe for the infant (87; 15).

There are no data on closing patent foramen ovale in women who desire pregnancy or during pregnancy. In selected patients with cryptogenic stroke and large shunts, patent foramen ovale closure prevented more strokes compared to medical therapy alone. Rarely, serious device-related adverse events and atrial fibrillation occurred after patent foramen ovale closure (38; Gore REDUCE Clinical Study Investigators 2017; 150).

A Cochrane analysis in women with antiphospholipid antibodies and recurrent pregnancy loss has not determined the best preventive strategy: a) aspirin with LMWH versus placebo or IVIG, or b) aspirin plus high-dose LMWH versus aspirin with low-dose LMWH or UFH. Further adequately powered studies are needed (13).

Of the patients diagnosed with moyamoya disease, 80% delivered via cesarean section. Maternal mortality was 13.6% and fetal death rate was 23.5%. Moyamoya disease does not seem to contraindicate pregnancy. It is unclear if bypass surgery before pregnancy or cesarean delivery after diagnosis of moyamoya disease improve outcome (116).

Pregnancy and puerperium do not seem to increase the risk of hemorrhage of cavernous hemangioma (93), unruptured arteriovenous malformations (AVMs) (113), or aneurysm (Tiel et al 2009; 100; 43). An incidental lesion is not a contraindication for pregnancy or reason to terminate it. The risk of hemorrhage from an unruptured AVM during pregnancy is approximately 3.5%, not significantly increased compared to 1% to 2% in the general population with AVMs (83; 59). Moreover, interventional therapy of unruptured AVMs was not superior to conservative management in patients followed up to 33 months (129). The risk of rebleeding of AVM is higher; however, the number of events during pregnancy is low, and there are not enough data to guide a preventive approach.

In cases of asymptomatic aneurysms, no data exist on the balance between the risks and benefits of intervention. The ISUIA-I and ISUIA-II studies on unruptured aneurysms suggest that stable asymptomatic anterior circulation aneurysms smaller than 7 mm have a low risk of rupture (86; 192). Whenever treatment for AVM or aneurysm is considered, the effect of radiation on the fetus needs to be weighed against the benefit of lower risk of rebleeding. Measures taken to avoid strenuous labor are reasonable.

Differential diagnosis

Other disorders seen during pregnancy may cause focal neurologic deficits. Metabolic disorders, seizures, migraine, and psychogenic disorders can be easily discerned based on history, basic laboratory, and imaging. The most important differential consideration specific to pregnancy is eclampsia. The focal neurologic deficits and T2 hyperintensities on MRI are usually transient and there is no evidence of ischemia on diffusion-weighted MRI imaging. However, a small percentage of patients can develop ischemic or hemorrhagic strokes in the context of PRES. PRES is also characterized by reversible T2 hyperintensities in the posterior cerebrum, but also in other subcortical regions and transiently narrowed blood vessels. Typical for RCVS is thunderclap headache, often associated with convexity subarachnoid hemorrhage, ischemic infarcts, intracerebral hemorrhage, and transient vasospasm, that usually resolves within 3 months. Cerebral vasculitis is characterized by moderately intense headache, focal neurologic symptoms and signs, and arterial narrowing that is not readily reversible. Abdominal pain, hemolytic anemia, elevated liver enzymes, and thrombocytopenia should raise suspicion for HELLP syndrome.

Diagnostic workup

With a few exceptions, the diagnostic approach to suspected stroke in a pregnant woman is similar to that of a nonpregnant young woman. Although radiation exposure is concerning, head CT is reasonably safe in pregnancy if the uterus is shielded (165; 47). CT angiography gives a dose of radiation to the fetus comparable to digital subtraction angiography and carries a lower risk of maternal complications (147). Use of iodinated contrast during the third trimester is considered safe, with only a slight risk of treatable fetal hypothyroidism (118).

Although most studies have shown no adverse effects of MRI exposure on fetal development, there are a few concerns based on animal studies of ocular deformities and growth retardation (77; 184). Nevertheless, radiologists have become more comfortable with the use of MRI during pregnancy, particularly in the second and third trimesters (45). Gadolinium contrast, however, should be avoided as it crosses the placenta and has unknown effects on development.

Catheter cerebral angiography is the gold standard for the diagnosis of vasculitis and for the evaluation of subarachnoid hemorrhages, aneurysms, and arteriovenous malformations. It also offers the option of nonsurgical intervention for ruptured aneurysms, which may be safer for the fetus than surgery (165).

History is the most important form of screening for preeclampsia. If preeclampsia without severe symptoms is present, daily assessment of symptoms and fetal movements, biweekly blood pressure measurements, and weekly platelet count and liver enzymes is recommended. Ultrasound and antenatal testing for fetal growth monitoring and status is suggested. If restricted fetal growth is suspected, umbilical artery Doppler velocimetry is recommended. Women with gestational hypertension should have weekly blood pressure and proteinuria measurements (05).

In patients with PRES, brain MRI shows T2 hyperintensities, especially in the occipital regions, without restricted diffusion on diffusion-weighted imaging, suggesting presence of vasogenic rather than cytotoxic edema (170; 163). Edema can be seen in other regions of the brain, including the brainstem of comatose patients (99). Catheter angiography or MR or CT angiography may reveal areas of arterial constriction alternating with dilatation and even a beading appearance (18).

During RCVS, imaging reveals convexity subarachnoid hemorrhage, brain infarcts, intracerebral hemorrhage that may occur days after the initial negative scan, and cerebral edema. MRI may show hyperintense signal abnormalities on T2-weighted images and fluid-attenuated inversion recovery (FLAIR), without restricted diffusion changes, predominantly in the posterior parietal and occipital lobes (103). These lesions are often reversible, although patients may suffer permanent deficits or may have a fatal course (176). Intracerebral hemorrhages may occur (66). Angiography demonstrates diffuse smooth narrowing of large and medium sized cerebral arteries. Transcranial Doppler may be used for monitoring (34). An overlap of vasoconstriction associated with vasculitis characterized by smooth arterial narrowing associated with perivascular inflammation was also described (31).

In patients with cerebral vasculitis, brain MRI may show several small, deep, or superficial infarcts of different ages, white matter abnormalities, and occasional enhancement of leptomeninges and intracranial lesions. Angiography may show alternating areas of smooth or irregular narrowing with arterial dilatation. High-resolution MRI may show enhancement with Gadolinium of the inflamed blood vessel wall (74).

CSF studies are reserved for diagnosis of subarachnoid hemorrhage or vasculitis. Brain biopsy is indicated in patients with suspected cerebral vasculitis for 2 reasons: to confirm the diagnosis and to exclude an alternative diagnosis that may exist in 39% of cases (04).

An echocardiogram with saline contrast bubble study should be obtained as for any young stroke patient, with particular attention paid to the presence of a patent foramen ovale or right to left shunt.

Laboratory studies for metabolic derangements, drug abuse, or coagulopathy should be routinely obtained, particularly in cases of cerebral venous thrombosis. Testing for protein C, protein S, or antithrombin III deficiencies, activated protein C resistance, factor V Leiden, MTHFR mutation, prothrombin gene G20210A mutation, elevated factor VIII, plasminogen activator inhibitor, and homocysteine levels, as well as anticardiolipin antibodies and lupus anticoagulant, is reasonable, but the impact on treatment is not clear (27). The impetus for routine testing for hypercoagulopathy has diminished as it does not provide useful information or alter the thromobophilia treatment (56; 08; 10).

Management

It is important to understand that preeclampsia may evolve even in the absence of proteinuria. Nonsteroidal antiinflammatory agents prescribed for pain should be avoided as they increase blood pressure. Antihypertensive medication is indicated for persistent systolic blood pressure greater than 160 mmHg and for diastolic blood pressure greater than 110 mmHg.

Delivery decision should not be based on amount of proteinuria. Delivery is recommended at 37 0/7 weeks or if severe preeclampsia at 34 0/7 weeks is complicated by any of the following: uncontrollable hypertension, eclampsia, pulmonary edema, abruption placentae, disseminated intravascular coagulation, nonreassuring fetal status, or intrapartum fetal demise. If severe preeclampsia and preterm rupture of membranes, labor, thrombocytopenia, fetal growth restriction, severe oligohydramnios, new or worsening of renal dysfunction, or reversed end-diastolic flow on umbilical artery Doppler occur before 33 6/7 weeks, delivery should be postponed for 48 after administration of corticosteroids.

There is no need for routine invasive hemodynamic monitoring in patients with severe preeclampsia. Corticosteroids may be administered to promote fetal lung maturation at less than 43 0/7 weeks. Administration of magnesium sulphate is recommended for eclampsia.

After discharge, if headaches are associated with blurry vision or preeclampsia with severe hypertension, parenteral magnesium sulphate is suggested. The treatment of moderate hypertension of systolic blood pressure greater than 150 mmHg and diastolic blood pressure greater than 100 mmHg should be initiated if measures taken 4 to 6 hours apart are still elevated during postpartum. Persistent systolic blood pressure greater than 160 mmHg and diastolic blood pressure greater than 110 mmHg should be treated within 1 hour (05).

Management of intracranial hemorrhages in pregnancy is similar to a nonpregnant patient. Coagulopathy or thrombocytopenia should be corrected urgently and blood pressure controlled with labetalol, hydralazine, or nifedipine. Seizures should be treated but prophylactic antiepileptic medication should not be used in absence of seizures. If Glasgow Coma Scale is less than 8, or if there is evidence of herniation or acute hydrocephalus, intracranial pressure monitoring might be considered. Evacuation of intracerebral hemorrhage can be performed in life-threatening situations such as cerebellar hemorrhages with deteriorating neurologic status, brainstem compression, or hydrocephalus. If the intracerebral hemorrhage is supratentorial, volume is greater than 30 cc, and the distance from surface is less than 1 cm, hematoma evacuation may be beneficial (131).

In aneurysmal subarachnoid hemorrhage, clipping of the aneurysm or endovascular coiling can be performed. After the aneurysm is secured, vaginal delivery is safe (46; 141). If labor or fetal distress occurs before the aneurysm is secured, cesarean section followed by or coincident with treatment of the aneurysm is preferred (Wilterdink and Feldman 2002).

Predictors of untreated arteriovenous malformation (AVM) bleeding are increased age, deep location, and exclusive deep venous drainage. Without these risks, the annual rate of first bleeding is 0.9% and with all 3 predictors as high as 34.4% (177). The risk of rebleeding of an arteriovenous malformation during pregnancy in a small study was estimated at 27% (155). Management of bleeding from arteriovenous malformation is similar in pregnant and nonpregnant women. Endovascular treatment, however, should be avoided prior to the twelfth week of gestation due to potential radiation risk to the fetus (142). Excision of bled arteriovenous malformation during pregnancy does not seem to prevent rebleeding (46).

The treatment of acute ischemic stroke during pregnancy was not studied in randomized trials and pregnancy is considered a relative contraindication (90). Recombinant tissue plasminogen activator (rt-PA) is currently the only drug approved for acute treatment of ischemic stroke within 4.5 hours from onset. tPA is category C, does not cross placenta, and is not associated with animal teratogenicity. Several case reports showed successful use of rt-PA (52; 191; 91; 110; 132; 193). The complication rate following thrombolytic therapy was similar in pregnant and nonpregnant women. The fetal fatality rate was estimated at 8%. In patients younger than 60 years of age, the risk of symptomatic intracerebral hemorrhage is lower than in general population, approximately 2.8% (120). The earlier the treatment, the higher the benefit and the lower the risk of complications (161). For a large vessel occlusion, intraarterial clot extraction is as effective as intravenous rt-PA (36). Several clinical trials demonstrated the benefit of endovascular thrombectomy in patients with large vessel occlusion but pregnancy was an exclusion criterion (20; 32; 70; 92; 162). In a case series of 7 patients, endovascular thrombectomy was safe and effective (112). Both penumbra and stent retriever were used successfully (01; 22). Although the intraarterial intervention is associated with increased upfront costs, the potential life year gains make this intervention cost effective (64).

Cerebral venous thrombosis, even when associated with hemorrhage, should be treated in the acute state with intravenous heparin followed by 3 to 6 months of anticoagulation with warfarin. In cases with a persistent prothrombotic state, long-term treatment is warranted. A small systematic review of 26 patients treated with systemic thrombolysis showed that 88% of patients regained their independence, but 2 died from intracerebral hemorrhage (186). Patients with severe edema, herniation, including comatose with bilaterally dilated pupils, may benefit from decompressive surgery. Comatose patients were less likely to become independent (57). Isolated cortical vein thrombosis is usually treated with anticoagulation (40).

Prophylactic treatment with low dose heparin beginning after delivery and lasting for 6 weeks postpartum should be considered for women with a prior history of cerebral venous thrombosis.

Posterior reversible encephalopathy syndrome (PRES) and eclampsia are treated with intravenous magnesium sulphate that promotes vasodilation, protects the blood-brain barrier, prevents edema formation, as well as acts as a potential anticonvulsant (55). Gradual treatment of elevated blood pressure with labetalol or nicardipine is advised in order to avoid placental hypoperfusion. The target levels are 140 to 155 mmHg systolic and 90 to 105 mmHg diastolic (39). If cytotoxic medications are the cause of PRES, they should be stopped or dosage reduced. Seizures should be treated with an appropriate antiepileptic unless eclampsia is present, which is generally treated with magnesium sulphate.

RCVS management was not tested in randomized trials. It is important to differentiate it from CNS vasculitis. Bed rest and avoidance of sexual activity, Valsalva maneuvers, vasoactive medications, or other potential offending drugs is generally recommended. Headache responds to analgesics and seizures to antiepileptic drugs. Intravenous hydration and vasodilator drugs like nimodipine, verapamil, or magnesium sulphate for 4 to 12 weeks have been tried. Although nimodipine improves headache, some patients still developed new hemorrhages, transient ischemic attack, or ischemic stroke several days after inception of therapy. Glucocorticoids should be avoided. In severe cases, angioplasty or intraarterial nimodipine was attempted with variable success (48). In case reports RCVS resolved soon after cesarean delivery (95; 44). Upshaw-Schulman syndrome is RCVS associated with anemia and thrombotic thrombocytopenic purpura and responds to plasmapheresis (182).

Cerebral vasculitis treatment was not studied in randomized trials. The response to corticosteroids with or without cyclophosphamide is good in most cases (157). A careful assessment of the risks and benefits as well as the optimal timing of therapy is essential in pregnant women.

HELLP syndrome responds to delivery. The urgency depends on the degree of maternal or fetal distress and gestational age (05).

Amniotic fluid embolism requires hemodynamic stabilization and correction of any metabolic derangements or coagulopathy. Air embolism is treated with hyperbaric oxygen and supportive care (137).

Special considerations

Anesthesia

Neurosurgical intervention and neuroanesthesia in pregnancy remain a complex medical dilemma in the absence of guidelines regarding surgical treatment options, timing of neurosurgical interventions, and the lack of evidence-based recommendations for neuroanesthesia.

The focus of neuroanesthesia is to prevent fetal distress while protecting the central nervous system of the mother. The choice of anesthetic agents and techniques (conscious sedation vs. general anesthesia) largely depends on causes of stroke and type of neurosurgical interventions.

Cerebral perfusion pressure should be maintained and increased intracranial pressure avoided. Intraoperative neuromonitoring should be considered in high-risk patients (145; 190).

The risk of intracerebral hemorrhage is not significantly reduced by cesarean section. Epidural anesthesia is preferred during vaginal delivery (Dias et al 1990; Stoodley et al 1998; 189).

In moyamoya disease, vaginal delivery under epidural anesthesia is an option. Transient ischemic attacks may occur (160).

A strong multidisciplinary collaboration between neurologists, neurosurgeons, and anesthesiologists and consideration of selected treatment strategies are essential to achieve favorable outcome for both the fetus and the mother.