Epidemiology

The incidence of neonatal stroke is estimated to be between 1:1600 and 1:5000 live births. Neonatal stroke accounts for approximately 25% of arterial strokes in the pediatric population. There are no apparent differences by infant sex or racial group, although data are somewhat inconsistent between study cohorts. Mortality was estimated in one study as 13.4% (8 died of hemorrhagic stroke, 9 died of arterial ischemic stroke, and 1 died of asphyxia and cerebral sinovenous thrombosis).

Risk Factors

Risk factors for PAIS can be grouped according to maternal, infant, or placental origin, and there is substantial overlap among these categories. Frequently, multiple concurrent risk factors are present and, in one study, the presence of multiple risk factors before delivery significantly increased the risk of perinatal stroke. Notably important is that most infants with one or more of these risk factors are entirely normal, and conversely, no risk factor is identified in many cases of neonatal stroke. Further understanding of the genetic factors that predispose newborns to stroke may help explain the cause in a proportion of newborns with no discernible etiology.

Clinical Manifestations

The clinical manifestations of neonatal stroke depend on the timing of presentation (fetal, neonatal, or during infancy and childhood), as well as the location and extent of brain involvement.

Neonatal Arterial Stroke

Term newborn infants with arterial stroke most commonly present with persistent focal seizures in the first hours of life. Infants may otherwise appear well. However, a study from the International Pediatric Stroke Study found that a quarter of the babies presenting in the newborn period with PAIS actually are systemically ill. Motor asymmetry, if present, is usually subtle, although the infant’s general movements are frequently abnormal. In the setting of concurrent global hypoxic-ischemic brain injury, signs of encephalopathy also occur, characterized by depressed level of alertness and altered feeding.

Arterial infarcts are slightly more common in the left hemisphere than in the right hemisphere and frequently occur in the anterior circulation ( Figs. 6.1 and 6.2 ). The stroke may involve small- or large-artery territory, and multiple infarcts are present in up to 20% of cases. Recent data have demonstrated that noncortical strokes in regions supplied by perforator arteries have pathogenic mechanisms similar to large-artery territory infarcts, such as hypoxia-ischemia, embolism, and infection. Perforator artery strokes are frequently asymptomatic.

Arterial ischemic stroke in a 2-day-old term infant presenting with focal clonic seizures at 24 hours of life. Placental pathology showed thrombotic vasculopathy and chorioamnionitis. A, Apparent diffusion coefficient map shows hypointense signal in the territory of the left middle cerebral artery posterior division. B, Corresponding hyperintensity is seen on diffusion-weighted imaging. C, Magnetic resonance angiography findings were normal.

Serial imaging of an infant treated with therapeutic hypothermia who was subsequently diagnosed with arterial ischemic stroke and later development of hemorrhagic transformation. Initial magnetic resonance imaging (MRI) obtained on day 4 of life with (A) apparent diffusion coefficient map and (B) diffusion-weighted imaging showing acute infarct of the left middle cerebral artery. C, T2-weighted MRI shows hyperintensity in the area of acute stroke consistent with vasogenic edema. D, Cranial ultrasound image obtained 3 days later for recurrent seizures showed increased echogenicity on the left, raising concern for hemorrhagic transformation. (E) Susceptibility-weighted imaging shows areas of hypointensity consistent with hemorrhage. (F) Axial T2-weighted MRI 2 weeks after the acute infarct shows evolving encephalomalacia and residual hemorrhage.

Ischemic Stroke in the Preterm Infant

Increased use of imaging in premature infants has led to greater recognition of arterial stroke prior to term-equivalent postmenstrual age. The clinical presentation in premature infants can differ greatly from that of term infants, with less than one-third presenting with seizures or apnea. Most often premature infants with stroke are identified during routine cranial ultrasound and the diagnosis of stroke is later confirmed using MRI. As with term gestation infants, arterial stroke in premature newborns is more often unilateral, in the middle cerebral artery territory distribution, and left-sided. Lenticulostriate distribution is common, especially in infants born at 28 to 32 weeks’ gestation. Twin-to-twin transfusion syndrome, fetal heart rate abnormalities, and hypoglycemia appear to be risk factors for preterm arterial stroke.

Presumed Perinatal Ischemic Stroke

PPIS is diagnosed in a child older than 28 days who presents with nonacute neurologic signs or symptoms referable to focal, remote (gliosis, encephalomalacia, and/or atrophy and absent restricted diffusion) infarct(s) on neuroimaging ( Fig. 6.3 ). Presenting signs and symptoms include asymmetric motor development typically characterized by pathologic early hand preference or hemiplegia, as well as nonmotor delay, or seizures leading to a diagnosis of PPIS at a median age of 1 year. PPIS may be due to either arterial (80%) or venous infarction but excludes global injuries, such as periventricular leukomalacia, basal ganglia, or watershed injury caused by hypoxic-ischemic injury. Cases of children diagnosed with arterial PPIS reviewed retrospectively indicate that acute perinatal risk factors, such as fetal distress, emergency cesarean section, or a requirement for neonatal resuscitation, are often present.

Remote left middle cerebral artery stroke in an 11-month-old child presenting with congenital hemiparesis and early handedness. Axial T1-weighted magnetic resonance imaging (A, C, E) and T2-weighted images (B, D, F) show reduced volume of the left hemisphere and cystic encephalomalacia in the left middle cerebral artery territory. Note the diminished size of the left cerebral peduncle at the level of the midbrain, reflecting Wallerian degeneration of the left corticospinal tract (E, F).

Imaging Arterial Ischemic Stroke

Magnetic resonance imaging (MRI) is the study of choice for the diagnosis and evaluation of neonatal stroke. Cranial ultrasound with transcranial Doppler may detect obliteration of normal gyral patterns, echogenicity in an arterial territory distribution, mass effect, or decrease in cerebral artery flow velocities in the affected hemisphere. However, cranial ultrasound is less sensitive than other imaging modalities. When possible, ultrasound should be supplemented with additional imaging, and MRI is preferable compared with computed tomography (CT). CT has the advantage of rapid acquisition and it can often be performed without sedation, but it provides less detailed anatomy than MRI and exposes the infant to ionizing radiation.

MRI is highly sensitive to evolving injury caused by ischemia and provides the highest level of resolution to determine the location and extent of injury. Reduced water motion on diffusion-weighted imaging sequences is apparent within hours after the injury (see Figs. 6.1 and 6.2 ) but becomes falsely negative with pseudonormalization by approximately 7 days in newborns. Conventional T1- and T2-weighted MRI sequences may be normal in the first 48 hours, making 2 to 5 days of life an ideal time to performing imaging in the infant with suspected stroke. MR angiography (MRA) permits the evaluation of the intracranial and neck arterial vasculature. MRA is useful to reveal underlying vessel pathology such as dissection from birth trauma or arterial malformation. MRI and MRA are often performed with anesthesia to improve image quality by sedating the neonate to reduce motion artifact. Wrapping techniques have been developed that use vacuum devices to immobilize the unsedated neonate and improve image quality. MR venogram (MRV) is warranted if there is thrombosis or suggestion of venous distribution infarct on the MRI (see later text).

In cases of PPIS, conventional T1- and T2-weighted images may reveal signs of remote infarct, including cystic encephalomalacia, gliosis, focal ventricular dilation, and Wallerian degeneration of the descending corticospinal tracts ( Fig. 6.3 ).

Management of Perinatal Arterial Ischemic Stroke

During the acute phase of neonatal stroke, attentive supportive care is important to minimize secondary brain injury. While there are no data from human neonatal trials, animal and adult studies support active maintenance of physiologic homeostasis, including temperature and blood glucose levels. Expert opinion supports aggressive treatment of clinical and electrographic seizures that are frequent or prolonged.

The value of antithrombotic therapy in newborns with ischemic stroke is uncertain, and guidelines or recommendations are based on clinical experience in older populations, observational or case studies, and clinical consensus. Guidelines by the American College of Chest Physicians recommend 3 months of anticoagulation with unfractionated or low-molecular-weight heparin for infants with cardioembolic arterial stroke (or cerebral venous sinus thrombosis [CSVT]) and without large territory involvement or hemorrhage. Anticoagulation therapy is not suggested for noncardioembolic neonatal arterial stroke.

Long-term management of cerebral palsy involves traditional rehabilitation with passive stretching, splinting, and casting, as well as medical or surgical treatment for spasticity, including baclofen, tendon release surgery, or botulinum toxin A. Constraint-induced movement therapy (CIMT), which involves restraint of the unaffected limb and frequent repetition of manual therapeutic tasks with the affected limb, is a promising treatment approach for children with hemiplegic cerebral palsy. Emerging evidence from studies using functional MRI suggests CIMT is associated with changes in cortical activation ; however, a Cochrane review concluded that there is only limited evidence of the clinical effect and suggested further trials to evaluate the efficacy of CIMT for hemiplegic children. Transcranial magnetic stimulation (TMS) is a noninvasive brain stimulation and neuromodulation technique that is currently under investigation in children with neonatal stroke. Preliminary data indicate safety and tolerability of contralesional inhibitory TMS with CIMT and intensive rehabilitation. Children with neonatal stroke should also be monitored through school age for subtle signs of cognitive impairment and referred for psychoeducational testing as needed.

Recent data from animal models have demonstrated that erythropoietin is a promising therapy to promote repair of injury following neonatal stroke. Benders and colleagues have demonstrated safety and feasibility of erythropoietin for neuroprotection after PAIS, and a larger randomized controlled trial is needed to determine whether treatment improves repair after PAIS and, consequently, neurodevelopmental outcomes.

Epidemiology

The incidence of venous ischemic stroke is estimated at 2.6 to 12 per 100,000 term newborns each year. Among reported cases of CSVT in pediatric patients, newborns are at highest risk. Mortality estimates range from 6% to 19% depending on the study. Male newborns appear to have an increased frequency of CSVT compared with female newborns.

Risk Factors for CSVT

Infants with CSVT frequently have comorbid risk factors, including dehydration, cardiac defects, sepsis, or meningitis, and extracorporeal life support requirement. In one study, hypoxic-ischemic encephalopathy was present in 20% of newborns diagnosed with CSVT. Maternal risk factors include premature rupture of membranes, chorioamnionitis, gestational diabetes, and preeclampsia with associated endothelial dysfunction. A difficult delivery with or without trauma to the skull and injury to superior sagittal sinus also increases the risk of CSVT.

Although prothrombotic disorders are present in up to two-thirds of patients with CSVT, these disorders are less common among newborns with CSVT. Few abnormalities were detected in a study of 52 newborns with CSVT, including 2 infants diagnosed with G20210A prothrombin gene mutation, as well as 13 with MTHFR C677T and A1298C mutations. Anticardiolipin antibody, lupus anticoagulant, prothrombin time, partial thromboplastin time, and fibrinogen testing should not be overlooked.

Clinical Manifestations

CSVT typically presents in the first weeks of life with nonspecific signs and symptoms, including encephalopathy, apnea, and seizures. A term infant presenting with intraventricular hemorrhage with thalamic hemorrhage and seizures is a common triad suggestive of venous sinus thrombosis. A full fontanelle and prominent scalp veins may also be seen.

Imaging CSVT

Cranial ultrasound in the term newborns with CSVT may show intraventricular and thalamic hemorrhage. In the hands of experienced technicians and radiologists, color Doppler ultrasound can demonstrate absent flow in the sinuses consistent with CSVT. Computed tomography, magnetic resonance, or conventional angiographic venography is much more sensitive for CSVT. Most often the sagittal, transverse, or straight sinus is involved, and in one study, the majority of newborns (80%) had multiple sinuses involved ( Fig. 6.4 ).

Cerebral sinovenous thrombosis in a newborn with congenital heart disease. A, Sagittal T1-weighted magnetic resonance imaging (MRI) shows extensive thrombosis with hyperintensity in the superior sagittal sinus, straight sinus, and great cerebral vein. B, Axial T1-weighted MRI shows thrombosis of the transverse sinuses and (C) thalamic hemorrhage and bilateral watershed injury with hemorrhage (white arrows). Thrombus is also seen in the superior sagittal sinus (black arrow).

Common sequelae include venous infarction with parenchymal and intraventricular hemorrhage. CSVT should be suspected when infarcts are not confined to known vascular territories or they are bilateral and involve deep gray nuclei. Hemorrhage is frequently associated with CSVT and can be seen on MRI with susceptibility-weighted image. In one study of 67 neonates for whom full imaging data were available, an infarct or hemorrhage was present in two-thirds. Serial imaging is required if a clot is present to evaluate for extension or propagation.

Management of CSVT

Management of neonatal CSVT includes supportive care, such as management of dehydration, treatment of infection, and seizures, as well as anticoagulation. The American Heart Association recommendations suggest considering anticoagulation in selected newborns with severe thrombophilic disorders, multiple cerebral or systemic emboli, or when there is clinical or radiologic evidence of propagating clot. The recommendations from the American College of Chest Physicians are more inclusive and suggest anticoagulation when there is not extensive intracranial hemorrhage. They further recommend consideration of treatment when clot is propagating even in the presence of extensive hemorrhage. In one retrospective study, slightly more than half of the 81 newborns with CSVT were treated with anticoagulation and there were no adverse events. In a prospective study of 104 neonates with CSVT, half received anticoagulation therapy with unfractionated heparin or low-molecular-weight heparin. Anticoagulation was initiated at diagnosis in three-fourths of treated infants and after confirmation of thrombus propagation on serial imaging in one-fourth. In the treated group, there were no serious adverse events such as systemic hemorrhage or anticoagulant-related deaths. Fourteen newborns had significant hemorrhage before receiving anticoagulation and none experienced worsening hemorrhage with treatment. Thrombus propagation occurred in one-third of the newborns who were not treated with anticoagulation but only occurred in one treated newborn. Venous infarction was common in infants with clot propagation. Thrombus propagation was clinically silent in all but one neonate who developed worsening seizures.

For infants with CSVT who are not treated with antithrombotic therapy, imaging should be repeated to look for propagation of the thrombosis. There is no consensus regarding the timing of repeat imaging, and 5 to 7 days is a reasonable approach in a clinically stable infant. Newborns with CSVT who are treated with antithrombotic therapy should have imaging performed again at 3 months to ensure complete recanalization of the venous sinuses.