Abnormalities of the cerebrovascular circulation include strokes—both ischemic and hemorrhagic, as well as anatomical anomalies, such as arteriovenous malformations and aneurysms. Cerebral ischemia is caused by reduction in blood flow that lasts longer than several seconds. Early, neurological symptoms may manifest. After a few minutes, however, infarction often follows. Hemorrhagic stroke is caused by bleeding directly into or around the brain. It produces symptoms by its mass effect, by toxic effects of blood, or by increasing intracranial pressure. Of strokes in pregnant women, roughly half are ischemic and the other half hemorrhagic (Zofkie, 2018).
The current obesity endemic in this country, along with concomitant increases in rates of heart disease, hypertension, and diabetes, has increased the prevalence of strokes (Centers for Disease Control and Prevention, 2012). Women have higher lifetime risk of stroke than men and greater associated mortality rates (Martínez-Sánchez, 2011; Roger, 2012). Moreover, pregnancy increases the immediate and lifetime risk of both ischemic and hemorrhagic stroke (Jamieson, 2010; Jung, 2010).
Stroke is relatively uncommon in pregnant women, occurring in 10 to 40 per 100,000 births, but it contributes disparately to maternal mortality rates (Leffert, 2016; Miller, 2016; Yoshida, 2017). The incidence is rising as measured by pregnancy-related hospitalizations for stroke (Callaghan, 2008; Kuklina, 2011). Importantly, most are associated with hypertensive disorders or heart disease. Of the pregnancy-related mortality rate in the United States, 6.6 percent is due to cerebrovascular accidents, and 7.4 percent is associated with preeclampsia (Creanga, 2017). Of maternal deaths after 42 days postpartum, 9.8 percent were attributable to cerebrovascular accidents.
Most strokes in pregnancy manifest either during labor and delivery or in the puerperium. In a study of 2850 pregnancy-related strokes, approximately 10 percent developed antepartum, 40 percent intrapartum, and almost 50 percent postpartum (James, 2005). In contrast, Leffert (2016) reports a timing of 45 percent antepartum, 3 percent intrapartum, and 53 percent postpartum in 145 women. Several risk factors—unrelated and related to pregnancy—have been reported from studies that included more than 10 million pregnancies. These include age; migraines, hypertension, obesity, and diabetes; cardiac disorders such as endocarditis, valvular prostheses, and patent foramen ovale; and smoking. Those related to pregnancy include hypertensive disorders, gestational diabetes, obstetrical hemorrhage, and cesarean delivery. By far, the most common risk factors are pregnancy-associated hypertensive disorders. A third of strokes are associated with gestational hypertension, and hypertensive women compared with normotensive counterparts have a three- to eightfold greater risk of stroke (Scott, 2012; Wang, 2011). Women with preeclampsia undergoing general anesthesia may be at higher risk of stroke compared with those given neuraxial anesthesia (Huang, 2010). Another risk factor for peripartum stroke is cesarean delivery, which raises the risk 1.5-fold compared with vaginal delivery (Lin, 2008).
Pregnancy-induced effects on cerebrovascular hemodynamics include enhanced autoregulation that maintains blood flow despite changes in systemic blood pressure (van Teen, 2016). Although cerebral blood flow decreases by 20 percent from midpregnancy until term, it increases significantly with gestational hypertension (Zeeman, 2003, 2004b). Such hyperperfusion at least intuitively would be dangerous for women with certain vascular anomalies.
Acute occlusion or embolization of an intracranial blood vessel causes cerebral ischemia, which may result in death of brain tissue (Fig. 60-2). The more common associated conditions and etiologies of ischemic stroke are shown in Table 60-3. A transient ischemic attack (TIA) is caused by reversible ischemia, and symptoms usually last less than 24 hours. Approximately 10 percent of these patients have a stroke by 1 year (Amarenco, 2016). Patients with a stroke usually have a sudden onset of severe headache, hemiplegia or other neurological deficits, or occasionally seizures. In contrast, focal neurological symptoms accompanied by an aura usually signify a first-episode migraine (Liberman, 2008).
TABLE 60-3Some Associated Disorders or Causes of Ischemic and Hemorrhagic Strokes During Pregnancy or the Puerperium ||Download (.pdf) TABLE 60-3 Some Associated Disorders or Causes of Ischemic and Hemorrhagic Strokes During Pregnancy or the Puerperium
|Ischemic Stroke ||Hemorrhagic Stroke |
|Preeclampsia syndrome ||Chronic hypertension |
|Arterial thrombosis ||Preeclampsia syndrome |
|Venous thrombosis ||Arteriovenous malformation |
|Lupus anticoagulant ||Saccular aneurysm |
|Antiphospholipid antibodies ||Angioma |
|Thrombophilias ||Cocaine, methamphetamines |
|Migraine induced ||Vasculopathy |
|Paradoxical embolus || |
|Cardioembolic || |
|Sickle hemoglobinopathy || |
|Arterial dissection || |
|Vasculitis || |
|Moyamoya disease || |
|Cocaine, amphetamines || |
Illustrations of a brain showing various types of strokes seen in pregnancy: (1) subcortical infarction (preeclampsia), (2) hypertensive hemorrhage, (3) aneurysm, (4) embolism or thrombosis in middle cerebral artery, (5) arteriovenous malformation, and (6) cortical vein thrombosis.
Evaluation of an ischemic stroke includes echocardiography and cranial imaging with CT, MR, or angiography. Serum lipids are measured with the caveat that their values are distorted by normal pregnancy (Appendix, Serum and Blood Constituents). Tests to detect antiphospholipid antibodies and lupus anticoagulant are performed. These underlie up to a third of ischemic strokes in otherwise healthy young women (Chap. 59, Antiphospholipid Syndrome). Also, sickle-cell syndromes are evaluated when indicated (Buonanno, 2016).
With a thorough evaluation, most causes of embolism can be identified, although treatment is not always available. Some of these include cardiac- associated embolism, vasculitis, or vasculopathy such as Moyamoya disease (Ishimori, 2006; Miyakoshi, 2009; Simolke, 1991). Outcomes of embolic strokes were reported to be favorable and similar to those of nonpregnant women (Leffert, 2016). Thrombolysis for ischemic stroke during pregnancy has been reported (Tversky, 2016).
In reproductive-age women, a significant proportion of pregnancy-related ischemic strokes are caused by gestational hypertension and preeclampsia (Jeng, 2004; Miller, 2016). As shown in Figure 60-2, areas of subcortical perivascular edema and petechial hemorrhage may progress to cerebral infarction (Aukes, 2007, 2009; Zeeman, 2004a). Although these are usually clinically manifest by an eclamptic convulsion, a few women will suffer a symptomatic stroke from a larger cortical infarction (Chap. 40, Eclampsia).
Other conditions with findings similar to preeclampsia include thrombotic microangiopathies (Chap. 56, Thrombotic Microangiopathies) and the reversible cerebral vasoconstriction syndrome (Chap. 40, Long-Term Consequences). The latter, also termed postpartum angiopathy, can cause extensive cerebral edema with necrosis and widespread infarction with areas of hemorrhage (Edlow, 2013; Katz, 2014; Miller, 2016).
These strokes usually involve the middle cerebral artery (see Fig. 60-2). The diagnosis can be made with confidence only after thrombosis and hemorrhage have been excluded and is more certain if an embolic source is identified. Hemorrhage may be more difficult to exclude because embolization and thrombosis are both followed by hemorrhagic infarction. Paradoxical embolism is an uncommon cause, even considering that more than a fourth of adults have a patent foramen ovale through which right-sided venous thromboemboli are deported (Scott, 2012). Foraminal closure may not improve outcomes in these patients, however, this procedure has been performed during pregnancy (Dark, 2011). Assorted cardioembolic causes of stroke include arrhythmias—especially atrial fibrillation, valvular lesions, mitral valve prolapse, mural thrombus, infective endocarditis, and peripartum cardiomyopathy.
Management of embolic stroke in pregnancy consists of supportive measures and antiplatelet therapy. Thrombolytic therapy and anticoagulation in pregnancy are controversial issues (Li, 2012).
Cerebral Artery Thrombosis
Most thrombotic strokes affect older individuals and are caused by atherosclerosis, especially of the internal carotid artery. Many are preceded by one or more TIAs. Thrombolytic therapy with a recombinant tissue plasminogen activator (rt-PA) is recommended. Alteplase is one of these and given within the first 3-hour window if there is measurable neurological deficit and if neuroimaging has excluded hemorrhage. This recombinant enzyme can be used in pregnancy. A principal risk is hemorrhagic transformation of an ischemic stroke in 3 to 5 percent of treated patients (Smith, 2015; van der Worp, 2007).
Cerebral Venous Thrombosis
In one study in the United States, 7 percent of cerebral venous thromboses were associated with pregnancy (Wasay, 2008). But, in the Nationwide Inpatient Sample of more than 8 million deliveries, James and associates (2005) reported that venous thrombosis caused only 2 percent of pregnancy-related strokes (Saposnik, 2011). There are numerous predisposing causes, and for gravidas, late pregnancy and the puerperium are times of greatest risk.
Thrombosis of the lateral or superior sagittal venous sinus usually occurs in the puerperium and often in association with preeclampsia, sepsis, or thrombophilia (see Fig. 60-2). It is more common in patients with inherited thrombophilias or antiphospholipid antibodies (Chaps. 52, Inherited Thrombophilias and 59, Antiphospholipid Syndrome). Headache is the most frequent presenting symptom, neurological deficits are common, and up to a third of patients have convulsions (Wasay, 2008). The diagnosis is made using MR venography (Saposnik, 2011).
Management includes anticonvulsants for seizures, and although heparinization is recommended by most, its efficacy is controversial (Saposnik, 2011; Smith, 2015). Antimicrobials are given if there is septic thrombophlebitis, and fibrinolytic therapy is reserved for those women failing systemic anticoagulation. The acute prognosis for venous thrombosis in pregnant women is better than in nonpregnant subjects, and mortality rates are less than 10 percent (McCaulley, 2011).
In women with a prior cerebral venous thrombosis, one systematic review found only one recurrence in 217 pregnancies and five noncerebral venous thrombotic events in 186 pregnancies (Aguiar de Sousa, 2016). In a study of 52 women on prophylactic anticoagulation with prior cerebral venous thrombosis, there were no cases of recurrent thrombosis or bleeding, however 24 percent had late obstetrical complications (Martinelli, 2016).
Recurrence Risk of Ischemic Stroke
Women with prior ischemic stroke have a low risk for recurrence during a subsequent pregnancy unless a specific, persistent cause is identified. During a 5-year follow-up of 373 women with arterial ischemic strokes, there were 187 pregnancies in 125 women. Thirteen women had a recurrent ischemic stroke, and of these, only two were associated with pregnancy. The authors concluded that the risk of stroke recurrence is low and a previous ischemic stroke is not a contraindication to pregnancy (Lamy, 2000). In one study of 1770 nonpregnant women with antiphospholipid-related ischemic stroke, investigators reported no difference in the recurrence risk as long as preventative treatment was given with warfarin or aspirin (Levine, 2004).
Currently, no firm guidelines define prophylaxis in pregnant women with a stroke history (Helms, 2009). The American Heart Association stresses the importance of controlling risk factors such as hypertension and diabetes (Furie, 2011). Women with antiphospholipid syndrome or certain cardiac conditions should be considered for prophylactic anticoagulation as discussed in Chapter 49 (Surgically Corrected Heart Disease) and 52 (Acquired Thrombophilias).
The two distinct categories of spontaneous intracranial bleeding are intracerebral and subarachnoid hemorrhage. The symptoms of a hemorrhagic stroke are similar to those of an ischemic stroke, and their differentiation is only possible with CT or MR imaging (Morgenstern, 2010; Smith, 2015).
Bleeding into the brain parenchyma most often is caused by spontaneous rupture of small vessels previously damaged by chronic hypertension (see Fig. 60-2). Thus, pregnancy-associated hemorrhagic strokes such as the one shown in Figure 60-3 are often associated with chronic hypertension and superimposed preeclampsia (Cunningham, 2005; Martin, 2005). Because of its location, this type of hemorrhage has much higher morbidity and mortality rates than does subarachnoid hemorrhage (Smith, 2015). Pressure-induced rupture causes bleeding into the putamen, thalamus, adjacent white matter, pons, and cerebellum. In the 28 women described by Martin and associates (2005), half died and most survivors had permanent disabilities. This cautions for the importance of proper management for gestational hypertension—especially systolic hypertension—to prevent cerebrovascular pathology (Chap. 40, Management Considerations).
A 37-year-old gravida with intrapartum eclampsia at term. A noncontrast computed tomography axial head image demonstrates a large intraparenchymal hemorrhage.
In a study of 639 cases of pregnancy-related subarachnoid hemorrhage from the Nationwide Inpatient Sample, the incidence was 5.8 per 100,000 pregnancies, with half occurring postpartum (Bateman, 2012). A remarkably similar incidence was reported in Japanese women (Yoshida, 2017). These bleeds are more likely caused by an underlying cerebrovascular malformation in an otherwise normal patient (see Fig. 60-2). Ruptured saccular or “berry” aneurysms cause 80 percent of all subarachnoid hemorrhages. The remaining cases are caused by a ruptured arteriovenous malformation, coagulopathy, angiopathy, venous thrombosis, infection, drug abuse, tumors, or trauma. Such cases are uncommon, and a ruptured aneurysm or angioma or bleeding from a vascular malformation has an incidence of 1 in 75,000 pregnancies. Although this frequency is not different from that in the general population, the mortality rate during pregnancy is reported to be as high as 35 percent (Yoshida, 2017).
Approximately 1 to 2 percent of adults have this lesion (Lawton, 2017). Fortunately, only a small percentage rupture. The rate approximates 0.1 percent for aneurysms <10 mm and 1 percent for those >10 mm (Smith, 2015). Most aneurysms identified during pregnancy arise from the circle of Willis, and in 20 percent of case, there are multiple lesions. Pregnancy does not raise the risk for aneurysmal rupture. However, because of their high prevalence, they are more likely to cause subarachnoid bleeding than other etiologies (Hirsch, 2009; Tiel Groenestege, 2009). A systematic review of 44 women with 50 aneurysms in pregnancy reported that 72 percent ruptured during pregnancy, and 78 percent of these did so during the third trimester (Barbarite, 2016). This proclivity for rupture late in pregnancy was also reported by Yoshida and colleagues (2017).
The cardinal symptom of a subarachnoid hemorrhage from an aneurysm rupture is sudden severe headache that is accompanied by visual changes, cranial nerve abnormalities, focal neurological deficits, and altered consciousness. Patients typically have signs of meningeal irritation, nausea and vomiting, tachycardia, transient hypertension, low-grade fever, leukocytosis, and proteinuria. Prompt diagnosis and treatment may prevent potentially lethal complications. The American Heart Association recommends noncontrast cranial CT imaging as the first diagnostic test, although MR imaging may be superior (Connolly, 2012; Smith, 2015).
Treatment of subarachnoid hemorrhage includes bed rest, analgesia, and sedation, with neurological monitoring and strict blood pressure control. Repair of a potentially accessible aneurysm during pregnancy depends in part on the risk of recurrent hemorrhage versus the surgical risks. At least in nonpregnant patients, the risk of subsequent bleeding with conservative treatment is 20 to 30 percent for the first month and then 3 percent per year. The risk of rebleeding is highest within the first 24 hours, and recurrent hemorrhage leads to death in 70 percent.
Early repair after the sentinel hemorrhage is done by surgical clipping of the aneurysm. Also, an endovascular coil can be placed using fluoroscopic angiography, while attempting to limit fetal radiation exposure. Barbarite and colleagues (2016) report lower complication rates with coil embolization than clipping. For unruptured aneurysms, surgical management resulted in a third fewer complications than no treatment. For gravidas remote from term, repair without hypotensive anesthesia seems optimal. For women near term, cesarean delivery followed by aneurysm repair is a consideration, and we have successfully done this in several cases.
For aneurysms repaired either before or during pregnancy, most allow vaginal delivery if labor ensues remote from aneurysmal repair. Problems arise in defining “remote,” and although some recommend 2 months, the time for complete healing is unknown. For women who survive subarachnoid hemorrhage, but in whom surgical repair is not done, we agree with Cartlidge (2000) and recommend against bearing down—put another way, we favor cesarean delivery.
These are congenital focal abnormal conglomerations of dilated arteries and veins with subarteriolar disorganization (see Fig. 60-2). They lack capillaries and have resultant arteriovenous shunting. Although unclear, the risk of bleeding may rise with gestational age. When arteriovenous malformations (AVMs) bleed, half do so into the subarachnoid space, whereas half are intraparenchymal with subarachnoid extension (Smith, 2015). They are uncommon and are estimated to occur in 0.01 percent of the general population. Of 65 identified cases of AVM in pregnancy, 83 percent ruptured during pregnancy or postpartum, and more than 80 percent of these ruptured in the second or third trimester. Hemorrhage upon presentation is associated with poor maternal outcome (Lu, 2016).
Bleeding does not appear to be more likely during pregnancy. Although these malformations are correspondingly rare during pregnancy, AVM bleeding accounted for 17 percent of hemorrhagic strokes in one study (Yoshida, 2017). At Parkland Hospital in a 33-year period during which there were about 466,000 births, 57 women had a CVA, and five of these strokes were due to a bleeding AVM (Simolkie, 1991; Zofkie, 2018).
Treatment of AVMs in nonpregnant patients is largely individualized. No consensus guides whether all accessible lesions should be resected. Factors include AVM symptoms; its anatomy and size; presence of an associated aneurysm, which is found in up to 60 percent of cases; and especially, prior AVM bleeding. After hemorrhage, the risk of recurrent bleeding in unrepaired lesions is 6 to 20 percent within the first year, and 2 to 4 percent per year thereafter (Friedlander, 2007; Smith, 2015). The mortality rate with a bleeding AVM is 10 to 20 percent. In pregnancy, the decision to operate is usually based on neurosurgical considerations, and Friedlander (2007) recommends strong consideration for treatment if bleeding occurs. Because of the high risk of recurrent hemorrhage from an unresected or inoperable lesion, we favor cesarean delivery.