Brain Injury in the Preterm Infant

Pathogenesis

IVH is a common injury in the preterm brain, originating in the subependymal germinal matrix, a highly vascularized region with highly active angiogenesis during this period in development. Cortical neuronal and glial progenitors develop from the germinal matrix and adjacent ventricular germinal zone during the late second and early third trimesters. The consequences of IVH are thus potentially several fold. IVH triggers degeneration of progenitor populations that are actively establishing cerebral connectivity. Hemorrhage disrupts the blood-brain barrier exposing the preterm brain to toxic and pro-inflammatory substances, which may potentiate the risk for further hemorrhage by disrupting pathways that stabilize the vasculature. Moreover, disruption of the germinal matrix disrupts neonatal neurogenesis, with a particular impact on later maturing populations of neurons including interneurons, which are key to the regulation of synaptic activity. Severe IVH with parenchymal white matter extension additionally targets vulnerable glial progenitors leading to disturbances in myelination.

Involution of the germinal matrix occurs with advancing gestation. The subependymal germinal matrix derives its arterial supply from the anterior and middle cerebral arteries as well as the anterior choroidal artery. These arteries feed an elaborate capillary network of thin-walled vessels that is continuous with a deep venous system that terminates in the vein of Galen. The terminal, choroidal, and thalamostriate veins course anteriorly to form the internal cerebral vein, which courses posteriorly to join the vein of Galen. Several of these veins make a pronounced U-shaped turn as they join with the internal cerebral vein. This turn may influence venous drainage and pressure in the germinal matrix. Visualization of subependymal venous anatomy using susceptibility-weighted venography recently identified five anatomical variants that differ from the classical anatomical pattern and were associated with increased risk for IVH.

The predisposition of the preterm infant to IVH is due to several hemodynamic factors. A pressure-passive state exists because of the lack of autoregulation of blood flow in the cerebral arterioles of the preterm brain. In the presence of a highly vascularized subependymal germinal matrix, the risk of IVH is enhanced by the lack of a supporting basement membrane for the germinal matrix blood vessels, an increased amount of fibrinolytic activity, and a decrease in extravascular tissue pressure in the first few days of extrauterine life. Thus, IVH may occur in the setting of elevated venous pressure or an increase in fluctuations in cerebral blood flow (CBF) velocity triggered by factors that include respiratory distress, pneumothorax, asphyxia, myocardial failure, patent ductus arteriosus, arterial hypotension, hypothermia, and hyperosmolarity. Fluctuating pressure passivity is common in preterm infants and may be associated with and precede IVH.

IVH has been produced experimentally when hypotension is followed by reperfusion. These studies support the finding that IVH is more likely when an early period of prolonged hypotension is followed by an increase in blood pressure. Isolated hypertension associated with seizures, intubation, and suctioning also predisposes the brain to IVH. Even gavage feeding and surfactant administration can lead to changes in cerebral hemodynamics, as measured by near-infrared spectroscopy, that lead to IVH. Box 54.1 lists key factors that may interact to produce IVH.

Cellular injury in infants with grade III or grade IV (periventricular hemorrhagic infarction [PVHI]) IVH may occur from antecedent ischemic injury, a decrease in cerebral blood flow, increased intracranial pressure, or vasospasm. More severe IVH is associated with cerebral WMI, cerebellar injury, and less frequently, pontine neuronal necrosis. In this setting, venous infarction leads to neuronal as well as glial death. In a contemporary multi-center cohort, grade II-III intraventricular hemorrhage was associated with a higher risk of punctate WMI, suggesting links to the less severe spectrum of brain injury.

The contribution of coagulation disorders and genetic factors to the pathogenesis of fetal or neonatal IVH remains unresolved. Fetal IVH occurs infrequently and in association with a wide variety of conditions, which include fetal bleeding disorders, anticoagulation, and conditions associated with thrombocytopenia including alloimmunization and hypoxemia. Fetal IVH is also associated with apparent remote perinatal strokes in term newborns who present with hemiparesis related to periventricular venous infarction. Mechanisms for coagulopathies in neonatal IVH may include thrombocytopenia related to platelet consumption or destruction in the setting of sepsis or pro-inflammatory states. Roles for factor V Leiden or prothrombin variants in IVH also are of unclear significance. Currently under study are many potential IVH risk genes including those involved in collagen-mediated vascular stabilization (e.g., COL4A1), nitric oxide-mediated vasomotor function (e.g., NOS3, encoding endothelial nitric oxide synthase), and vasoconstriction-mediated cerebral autoregulation (e.g., END1, encoding endothelin 1).

Site, Incidence, and Timing of Hemorrhage

In preterm infants, germinal matrix hemorrhage is most commonly seen at the junction of the terminal, choroidal, and thalamostriate veins in the germinal matrix overlying the body of the caudate nucleus at the level of the foramen of Monro. Parenchymal hemorrhage occurs most commonly in the frontoparietal regions, where it appears not to be an extension of IVH but rather a separate process—a hemorrhagic infarction. The hemorrhage is more often unilateral or, in less than a third of cases, asymmetrically bilateral.

The incidence and severity of IVH increase with decreasing gestation and peak at the limit of viability. IVH occurs infrequently in term newborns, but often in association with birth-related complications. The overall incidence of grade III IVH and PVHI in VLBW infants varies among centers and generally ranges from 10% to 20% for the majority of population-based studies worldwide. Recent data from over 50,000 VLBW infants found an overall incidence of IVH of 24.6% and 8.1% for grade III IVH and PVHI. Importantly, in very preterm neonates, even low-grade IVH is associated with an increased risk of WMI. Although there has been a decline in the incidence of most grades of IVH, the increased survival of VLBW infants has resulted in an increase in the absolute number of infants with IVH who are at risk for adverse neurodevelopmental outcomes, which include cerebral palsy, developmental delay, hydrocephalus, and epilepsy.

The high-risk period for IVH is the first 3 or 4 days of life. Hemorrhage is rarely seen at birth, although it has been reported as early as the first hour of life. Around half of neonatal hemorrhages occur by the sixth hour of life, and only about a third occur after the first 24 hours of life. Less than 5% of newborns develop IVH after the fourth or fifth day of life with a small percentage occurring by 7 to 10 days of life. Both early and later onset of hemorrhage may occur because of systemic hypoperfusion that results in disturbances in cerebral blood flow.

Clinical Presentation

The clinical presentation of IVH in the newborn depends on the extent of the hemorrhage. It may range from asymptomatic to a sudden and catastrophic deterioration that manifests itself with neurologic signs such as stupor or coma, seizures, decerebrate posturing, or apnea. A tense fontanel together with a sudden drop in hematocrit, hyperglycemia, hyperkalemia, hypotension, or bradycardia may herald an IVH. Inappropriate secretion of antidiuretic hormone may occur. The more common presentation, however, is that of a gradual clinical deterioration with an altered level of consciousness, hypotonia, abnormal extremity, or eye movements. In 25% to 50% of cases, clinical signs are lacking and IVH is identified on routine cranial ultrasound screening.

Grading of Intraventricular Hemorrhage

Ultrasound examination is a reliable and sensitive bedside technique for the evaluation of IVH. Papile et al. adapted the standard grading system originally applied to computed tomographic images of IVH to ultrasound images. They classified IVH into four grades of severity related to the location and extent of the hemorrhage ( Fig. 54.1 , Table 54.1 ). This classification system was previously used widely for outcome studies. Currently, the more widely used grading system is that proposed by Volpe et al., which relies on the cranial ultrasound examination to define the extent of ventricular hemorrhage. Because parenchymal involvement is a distinct process, it is not included in the continuous grading of IVH severity ( Table 54.1 ). PVHI appears to arise from venous infarction of the periventricular white matter rather than from a direct extension of the IVH into the parenchyma. Hence the presence of intracerebral hemorrhage or parenchymal lesions is described separately and is not designated as grade IV. Intraparenchymal hemorrhage is followed in 1 to 8 weeks by tissue destruction and the formation of a porencephalic cyst. Serial ultrasound examinations are especially important in linking the severity of IVH with neurodevelopmental outcomes.

In addition to various forms of cerebral WMI, IVH may cause graded injury to the cerebellum ( Fig. 54.2 ) that manifests as magnetic resonance imaging (MRI) defined changes in microstructure and reductions in cerebellar growth. This reduced growth appears to be related to a large population of proliferative external granule cells that are the progenitors that generate the internal granule cell layer, and which account for most of the cells in the human cerebellum. These cerebellar progenitors may be particularly vulnerable to the toxicity of IVH-derived blood products, which are detected by MRI as hemosiderin deposition on the surface of the brainstem and cerebellum. The heightened vulnerability of cerebellar progenitors is consistent with serial neuroimaging studies of human preterm survivors that demonstrated disrupted cerebellar growth in response to postnatal glucocorticoid exposure. Furthermore, with the increasing use of MRI with “blood-sensitive” sequences such as susceptibility weighted images, primary hemorrhage in the cerebellar germinal matrix is increasingly recognized. Neurodevelopmental assessments at 4.5 years show that the size and location of cerebellar hemorrhages are associated with dose-dependent disturbances in motor and visuomotor function and increased externalizing behaviors. Given this, the volumetric quantification and localization of a cerebellar hemorrhage, even when small, may allow for targeted rehabilitation interventions.