104: Intracranial Hemorrhage

An intracranial hemorrhage (ICH) can occur in term and preterm infants. An ICH occurring in term infants tends to be subdural, subarachnoid, or subtentorial and is most related to birth trauma, hypoxic-ischemic events, coagulopathies (eg, thrombophilias or thrombocytopenia), and undetermined causes. The most common ICH in preterm infants is bleeding from the subependymal germinal matrix and may result in intraventricular or periventricular hemorrhage, either of which can potentially cause hemorrhagic infarctions of the white matter. This chapter reviews the following clinical conditions: subdural hemorrhage (SDH), epidural hemorrhage, subarachnoid hemorrhage (SAH), intracerebral parenchymal hemorrhage, intracerebellar parenchymal hemorrhage (ICPH), and germinal matrix and intraventricular hemorrhage (GM/IVH).

A subdural hemorrhage (SDH) is an accumulation of blood between the dura and the arachnoid membrane and involves tears of bridging veins of the subdural compartment. The vascular structures most affected are superficial cerebral veins, infratentorial posterior fossa venous sinuses, the inferior sagittal sinus, and tentorial sinuses and veins (eg, vein of Galen). Blood may accumulate and cause acute symptoms of intracranial pressure (ICP) or reside as a hematoma that slowly evolves as a chronic subdural hematoma with increasing fluid accumulation and increasing ICP.

A SDH is very common after birth: up to 50% of term asymptomatic infants may have SDH. It usually follows a traumatic delivery of a late preterm or term infant. Only on rare occasions does SDH become serious.

A SDH is typically related to traumatic birth events involving labor and delivery. Undue pressure on the skull and torsion may produce shear forces resulting in rupture of superficial cerebral bridging veins or tears in the dura or dural reflections (eg, either the falx cerebri or tentorium and associated venous sinuses). These events are usually found over the cerebrum or within the posterior fossa. Occasionally skull fractures accompany these findings. Timing of the onset of SDH and clinical findings may be acute or delayed. Clinical signs may be minimal to none with the SDH self-resolving, or subtle findings of slight irritability or a seemingly hyperalert state may foretell an underlying accumulating SDH with delayed onset of more serious neuropathic circumstances. Latent SDH can lead to a subdural hematoma and subdural effusion with increasing intracranial pressure.

Include precipitous labor and delivery, instrumented deliveries utilizing forceps or vacuum-assist devices, a large for gestational age infant with cephalopelvic disproportion, and coagulopathies including familial thrombophilias and vitamin K deficiency.

Signs include lethargy alternating with irritability or asymmetric hypotonia of upper and lower extremities on the contralateral side of the SDH. More specific to SDH is impaired third cranial nerve function ipsilateral to the SDH. Focal seizures may present at any time and are much more likely to occur in low birthweight infants. Signs of increasing ICP may include a bulging fontanel, deviations of eye movements, and increasing occipital-frontal head circumference. Additional clinical signs can be decreased feeding, intermittent vomiting, and failure to thrive, all of which are more often related to late post-SDH neuropathic events.

1. Laboratory

   1. Hematocrit. Unexplained anemia.

   2. Total serum bilirubin. Persistent newborn jaundice.

   3. Cerebral spinal fluid (CSF) studies are indicative but not diagnostic of SDH. Emphasis on ICH hemorrhage can be taken if a combination of CSF findings is seen: large numbers of red bloods cells (especially if crenated), xanthochromia, elevated protein content, and hypoglycorrhachia (ie, CSF glucose <50% of concomitant blood glucose).

2. Imaging studies

   1. Computed tomography (CT) readily identifies most SDH.

   2. Magnetic resonance imaging (MRI) is best for detailing posterior fossa lesions and accumulations of blood or effusion.

   3. Ultrasound (US) does not readily lend itself to SDH identification, except for possible midline shifts.

Documentation of risk factors and appropriate cogent observation are the most important first clinical steps. Careful and repeated neurologic examinations will reveal neurologic signs that should be followed by laboratory and imaging studies.

The outcome of SDH ranges from early death to minimal or no disabling conditions. Much of the neurologic outcome of SDH depends on accompanying conditions soon after birth (eg, prematurity, birth asphyxia, shock, hypoxic-ischemic encephalopathy, or infection). Massive tentorial tears and hemorrhage dictate death or severe and long-term handicap. Infants with major SDH can have mortality rates upward of 45%. Conversely, in most cases SDH can be limited, produce few clinical signs, and have a good outcome. More than 50% of infants with minimal early clinical findings and later good outcomes have been largely found to have had small cerebral convexity SDH hemorrhages.

An epidural hemorrhage, blood between the inner skull and dura, is extremely rare in newborns. It is usually caused by injury to the middle meningeal artery, which is less susceptible to injury because it moves freely. Causes include birth trauma or an infant being dropped at delivery. Diagnosis is by CT or MRI. Affected infants usually have a skull fracture and cephalhematoma, and treatment is supportive with possible surgical/needle aspiration.

A subarachnoid hemorrhage (SAH) is an accumulation of blood between the arachnoid mater and the pia mater. The arachnoid mater is an avascular membrane situated below the dura mater and, together with the pia mater, constitutes what is known as the leptomeninges. Unlike adult SAH, which is arterial, infant SAH is venous in origin, coming from bridging veins within the subarachnoid space; however, on rare occasions it may be arterial, coming from leptomeningeal arteries of the subarachnoid space. SAH may be primary, coming from the vessels of the subarachnoid space, or secondary, occurring when the blood extends from existing intraventricular, cerebral, or cerebellar hemorrhages.

A small SAH is commonly seen in preterm and term newborn infants. It is of limited significance unless other conditions are present, such as prematurity, coagulopathies, birth trauma, or asphyxia. Primary bleeding within the subarachnoid space is usually self-limited, and it is the second most common ICH seen in newborns.

Rupture of the small vessels of the subarachnoid space can be associated with trauma in term infants, or birth asphyxia in preterm infants, and is most often idiopathic and insignificant.

See Pathophysiology.

In term infants, a SAH is mostly asymptomatic. Mild to intermittent irritability or lethargy may herald the onset of seizures on the second to third day of life.

1. Laboratory. CSF findings in SAH mirror those already discussed for SDH.

2. Imaging studies. CT and MRI studies establish the existence of primary SAH or identify other lesions that may be the source of secondary SAH.

Close observation and repeated neurologic examinations suffice for those infants at risk but without signs of SAH. Anticonvulsant medication and intravenous fluid therapy are needed if the infant has lethargy and/or seizure activity. Serum electrolyte and urine output monitoring for possible syndrome of inappropriate antidiuretic secretion is recommended if a significant amount of SAH has been identified. Regular sequential head circumference measurements will identify suspect cases of posthemorrhagic hydrocephalus. The latter is due to obliteration of CSF resorption sites by the organizing blood. Follow-up cranial imaging studies will also be needed.

A primary isolated SAH is mostly uncomplicated. Infants who have seizures that resolve before discharge from the hospital have an expected 90% uncomplicated outcome. Infants who develop long-term complications are mostly within those who had coexisting problems associated with birth trauma or perinatal asphyxia.

An intracerebral parenchymal hemorrhage occurs deep within the brain tissue after venous infarction and is commonly referred to as periventricular hemorrhagic infarction (PVHI). Periventricular leukomalacia in preterm infants and porencephalic cysts in term infants are not uncommon complications of PVHI.

The occurrence of PVHI may be as much as 10–15% among infants with ICH.

It is postulated that venous thrombosis and/or stasis causes PVHI through increased intravascular pressure that leads to rupture of parenchymal vessels; however, the exact mechanism remains unclear. Affected preterm infants have hemorrhagic venous infarction of subcortical and periventricular white matter, whereas affected term infants develop subcortical hemorrhage with infarction of overlying cortex.

PVHI is seen most often after a perinatal hypoxic-ischemic event.

Clinical signs of PVHI follow those of severe neonatal encephalopathy and overlap with clinical signs as seen with SDH, SAH, or IVH.

1. CT is most useful for detecting recent hemorrhage.

2. MRI findings of hypodensities suggest evolving areas of brain injury.

3. Cranial US is particularly useful in identifying PVHI. An echodense lesion of periventricular white matter with an associated germinal matrix (GM) hemorrhage or IVH usually identifies coexisting PVHI. Whether the PVHI is unilateral or bilateral, it is always an asymmetric lesion.

PVHI requires observational and supportive care, much as severe SDH or SAH. If imaging studies suggest a midline shift, neurosurgical consultation should follow. Subsequent posthemorrhagic hydrocephalus (PHH) is always a threat, which also requires neurosurgical consultation.

Developmental studies of preterm infants with PVHI have shown that significant cognitive and/or motor delays complicate overall recovery in at least two-thirds of survivors. Careful follow-up is thus indicated in every case of PVHI.

An intracerebellar parenchymal hemorrhage (ICPH) is most often seen in preterm infants with complications of labor and delivery and for whom intense respiratory management is required. ICPH in term infants is almost always associated with birth trauma.

Reports vary by gestational age. In neuropathologic reports, preterm infants <1500 g have an occurrence of 15–25%. A report using cranial US has defined the incidence as 2.8% in a population of <1500 g infants and an incidence of 8.7% for the smallest infants weighing <750 g. In another report using brain MRI in infants <34 weeks' gestation, the incidence was ∼10%.

Traumatic delivery.

Four mechanisms for intracerebellar parenchymal hemorrhage are possible:

1. Primary hemorrhage into either cerebellar hemisphere or into the vermis.

2. Venous infarction.

3. Supratentorial IVH and SAH is associated with reduction of preterm cerebellar growth, which may reflect concurrent cerebellar injury or direct effect of the blood on cerebellar development.

4. Direct trauma to the posterior fossa with rupture of cerebellar bridging veins or the occipital sinuses. This is seen primarily in term infants. Most ICPH is unilateral and focal with a predilection to the right cerebellar hemisphere.

An ICPH is unique in causing unexplained motor agitation, in addition to respiratory compromise, apnea, and breathing irregularities. Otherwise, the general symptoms of ICH are present as well.

CT and MRI have superior ability over US in diagnosing ICPH. However, US done via the mastoid fontanel can provide additional information.

All management modalities presented for other ICH patients apply to infants with confirmed or suspected ICPH. For infants at risk for ICPH, the added combination of shock and acidosis are closely related and warrant diagnostic efforts specific for ICPH.

Infants with ICPH usually require longer periods of mechanical ventilation. They will need close neurodevelopmental assessment, much as required with other types of ICH. In general, cerebellar hemorrhages seen only on MRI have much better prognosis than those detectable by US.

IVH is the most common CNS complication of a preterm birth. The occurrence is greatly associated with the immaturity of the germinal matrix of the lateral ventricles. Acidosis, birth asphyxia, shock, blood pressure fluctuations, and hypoxia are common related problems.

The germinal matrix, located between the caudate nucleus and the ependymal lining of the lateral ventricle, is normally not seen on cranial US. When germinal matrix hemorrhage occurs, it becomes readily identified by US and is seen as subependymal bleeding originating between the groove of the thalamus and the head of the caudate nucleus. Bleeding may be confined to the germinal matrix or it may rupture into either lateral ventricle and thereby become a unilateral or bilateral GM/IVH.

By 36 weeks' postconceptional age, the germinal matrix has involuted in most infants, although some residual may persist. If IVH occurs in term infants, it originates most often in the choroid plexus; however, residual subependymal germinal matrix may also be a point of origin. Following IVH, further insult by venous thromboses may result in thalamic infarction.

The overall occurrence of IVH in preterm infants <1500 g is ∼13–15%. Rates vary by gestation, with the greatest risk of GM/IVH in preterm infants with birthweights <750 g. Because IVH is rarely seen in term infants, their incidence rates are exceptionally low and associated with birth-related injury or asphyxia. Curiously, 2–3% of seemingly normal term infants, when studied prospectively, have been noted to have asymptomatic IVH.

The germinal matrix is a marginally supported and highly vascularized area. The blood vessels (as arterioles, venules, or capillaries) in this area of immature cerebrum are especially prone to hypoxic-ischemic injury. The vessels are irregular, with large luminal areas, and readily rupture. The GM begins to involute after 34 weeks' postconceptional age, and thus the peculiar vulnerability and predilection for GM/IVH for preterm infants lessens, but is not totally removed. Late preterm infants (34–37 weeks' gestation) may have IVH that reflects those of early preterm infants. Fluctuations in cerebral blood flow (CBF) play an important role in the pathogenesis of GM/IVH because sick premature infants have pressure-passive cerebral circulation. A sudden rise or fall in systemic blood pressure can result in an increase in CBF with subsequent rupture of the germinal matrix vessels. Decreases in CBF can also result in ischemic injury to the GM vessels and surrounding tissues, making them prone to secondary rupture following reperfusion.

The unique deep venous anatomy at the level of the foramen of Monroe and the open communication between the GM vessels and the venous circulation contribute to the danger of abrupt or sharp fluctuations in cerebral venous pressure. Given this anatomic proximity, rupture through the GM subependymal layer results in entrance of blood into the lateral ventricles in nearly 80% of affected infants.

1. Neuropathologic consequences of IVH

   1. GM ventricular-subventricular zone contains the migrating cells of origin for the cerebral cortex. It is the site of production of neurons and glial cells of the cerebral cortex and basal ganglia. GM destruction may result in impairment of myelination, brain growth, and subsequent cortical development. Moreover, in preterm infants, GM-IVH leads to reduction of cerebral perfusion in the first 2 weeks after the hemorrhage. The reduction was found to be most severe around day 5 and was irrespective of the IVH grade.

   2. Periventricular hemorrhagic infarction (PVHI) is venous in origin, associated with severe and usually asymmetric IVH, and invariably occurs on the side with the larger amount of intraventricular blood. It is a distinct pathologic event following venous stasis; it is often mistakenly described as an “extension” of IVH, of which it is not. Moreover, PVHI is neuropathologically distinct from periventricular leukomalacia (PVL). See preceding discussion under PVHI.

   3. Posthemorrhagic hydrocephalus (PHH) is more common in those infants with the highest grade of hemorrhage. It is most frequently attributable to obliterative arachnoiditis either over the convexities of the cerebral hemispheres with occlusion of the arachnoid villi or in the posterior fossa with obstruction of outflow of the fourth ventricle. Rarely, aqueductal stenosis is caused by an acute clot or reactive gliosis.

   4. Periventricular leukomalacia (PVL) is a frequent accompaniment of IVH but is not directly caused by IVH. PVL is an ischemic brain injury followed by necrosis of periventricular white matter adjacent to the lateral ventricles. It is usually a nonhemorrhagic event resulting from hypotension, apnea, and other hypoxic-ischemic events known to decrease CBF. Most PVL lesions are symmetrical in distribution.

Prematurity and respiratory distress syndrome have remained as the most closely related clinical circumstances to GM/IVH. As mentioned earlier, the immature cerebral vascular structures of preterm infants are extremely vulnerable to volume and pressure changes and to hypoxic and acidotic changes. Secondarily, respiratory distress, and its attendant limitations for oxygenation, further attenuates the immature vasculature of the preterm brain. Birth asphyxia, pneumothorax, shock/hypotension, acidosis, hypothermia, and therapeutic volume and/or osmolar overloads all serve to multiply the risk for GM/IVH. Even procedures that we perceive as routine in the care of premature infants may also be contributory, such as tracheal suctioning, abdominal examination, and handling to reposition or to instill mydriatics for an eye examination.

Of growing importance for the understanding of preterm GM/IVH is the possible role of fetal and neonatal inflammatory responses. Chorioamnionitis and funisitis may be harbingers of postnatal cerebral vascular events leading up to GM/IVH. Fetal inflammatory responses and subsequent neonatal hypotension and sepsis are closely related processes to IVH. Mediators of an inflammatory response, such as cytokines, have vasoactive properties that may be the source of exaggerated blood pressure changes that overwhelm the pressure passive state of the germinal matrix.

The clinical presentation is diverse, and diagnosis requires neuroimaging for confirmation. Signs may mimic those of other ICH or common neonatal disorders such as metabolic disturbances, asphyxia, sepsis, or meningitis. IVH may be totally asymptomatic, or there may be subtle symptoms, for example, a bulging fontanel, a sudden drop in hematocrit, apnea, bradycardia, acidosis, seizures, changes in muscle tone, or changes in level of consciousness. A catastrophic syndrome may accompany an extensive IVH. It is characterized by a precipitous fall of hematocrit, a rapid onset of stupor or coma, respiratory failure, seizures, decerebrate posturing or a profound flaccid quadriparesis, and fixed pupils

1. Cranial US. Cranial US (see Chapter 11 for imaging examples) is the procedure of choice for screening and diagnosis. CT and MRI are acceptable alternatives but are more expensive and require transport to the imaging service. They are valuable for a more definitive diagnosis or documentation of static brain injury before discharge from the hospital. Two systems for classifying GM/IVH have been advanced for clinical use. The older and most time honored has been that of Papile, based originally on CT but adapted for interpretation of cranial US. The second is the classification promulgated by Volpe, based also on cranial US. The utility of classification schema resides in the ability of clinicians to communicate degrees of severity and to have a source of information for comparison of lesions as well as having a means to follow progression or regression and recovery of the initial insult of IVH. The GM/IVH classification by Papile (updated in 2002) gives classes I to IV. Classes I and II are small hemorrhages. Class I is hemorrhage only seen in the GM. Class II shows hemorrhage from the GM extending into the lateral ventricles but without ventricular dilation. Classes III and IV are moderate to severe lateral ventricular hemorrhages, with the former resulting in acute ventricular dilation and the latter marked by extension into parenchymal hemorrhages.

   IVH classification by Volpe offers a somewhat different perspective. His class I is confined to the GM with little or no IVH. Class II is an IVH seen on parasagittal view and extends into >50% of the lateral ventricles. Class III is IVH >50% on parasagittal view with distention of the lateral ventricles. Lastly, Volpe points out that a cranial US finding of any periventricular echo density is an obvious and more serious intracranial vascular insult, such as PVHI or PVL.

   A cranial US is indicated for screening sick preterm infants for IVH from the first day of life and throughout the hospitalization. Typically a cranial US is done between day 1 and day 7, depending on clinical presentation and institutional protocols, keeping in mind that 50% of GM/IVH may occur on day 1, but 90% have occurred by day 4 of life. Of all GM/IVH identified by day 4 of life, 20–40% will progress to more extensive hemorrhage. Most clinicians obtain a final cranial US, CT, or MRI before discharge, or at 36 weeks' postconceptional age.

2. Laboratory studies. CSF initially shows elevated red and white blood cells, with elevated protein concentration. The degree of elevation of CSF protein correlates approximately with the severity of the hemorrhage. It is frequently difficult to distinguish IVH from a “traumatic tap.” Within a few days after hemorrhage, the CSF becomes xanthochromic, with a decreased glucose concentration as in other forms of ICH. Often, the CSF shows a persistent increase in white blood cells and protein and a decreased glucose level, making it difficult to rule out meningitis

1. Prenatal prevention

   1. Avoidance of premature delivery.

   2. Transportation in utero.

   3. Data suggest that active preterm labor may be a risk factor for early IVH and there may be a protective role for cesarean delivery. However, Anderson et al showed that cesarean birth before the active phase of labor resulted in a lower frequency of severe IVH and less progression to severe IVH, although it did not affect the overall incidence of IVH. In a study by Durie et al, there was no correlation between mode of delivery and the incidence of IVH in very low birthweight infants who are presenting by the vertex.

   4. Antenatal steroid therapy. Several large multicenter trials have shown a clear efficacy of antenatal steroids in reducing IVH. In one study the incidence of GM hemorrhage or IVH was 2- to 3-fold lower in infants whose mothers received a complete course of antenatal steroids compared with those whose mothers received no steroids or an incomplete course (<48 hours). Moreover, this beneficial effect appears to be independent of the improvement in respiratory status. The prevention of IVH may be a composite effect of enhanced vascular integrity, decreased respiratory distress, and possibly altered cytokine production. Blickstein and coworkers have reported that antenatal steroids (given between 24 and 32 weeks' gestation) as a completed course (48 hours) resulted in a 2.5 times lower incidence of GM/IVH (7.7% vs 19.4%) for singleton and multiple births.

2. Postnatal prevention

   1. Avoid birth asphyxia.

   2. Avoid large fluctuations in blood pressure.

   3. Avoid rapid infusion of volume expanders or hypertonic solutions.

   4. Use prompt but cautious cardiovascular support to prevent hypotension.

   5. Correct acid-base abnormalities.

   6. Correct abnormalities of coagulation.

   7. Avoid poorly synchronized mechanical ventilation. Consider sedation and in difficult situations pharmacologic paralysis.

   8. Postnatal pharmacologic intervention with indomethacin. Ment et al reported in 1994 that low-dose prophylactic indomethacin significantly lowered the incidence and severity of IVH but did not appear to be of benefit in the prevention or extension of early IVH. Subsequent reviews and reports over the ensuing years have not confirmed indomethacin prophylaxis for prevention of IVH, and the approach remains controversial. There are reports of reduced cerebral blood flow after indomethacin as well as no difference in long-term neurologic outcome for treated versus nontreated infants. More recently, Ment et al have reported on long-term follow-up of school-age preterm infants, indicating improved vocabulary scores with a striking favorable sex response to indomethacin therapy for male subjects. In the same year, Miller et al reported less white matter injury after 3–6 doses of low-dose indomethacin therapy for infants of <28 weeks' gestation. Much of the controversy today surrounding low-dose prophylactic indomethacin therapy for the prevention or amelioration of GM/IVH involves the continuing concerns for indomethacin-related complications of necrotizing enterocolitis, spontaneous intestinal perforation, decreased renal function (albeit transient in most cases), and the threat of persistent pulmonary hypertension. Jones et al reported in a meta-analysis published in 2010 that there is no statistical difference in the incidence of IVH as a complication due to the use of either ibuprofen or indomethacin to treat patent ductus arteriosus.

3. Management of acute hemorrhage

   1. General supportive care to maintain a normal blood volume and a stable acid-base status.

   2. Avoid fluctuations of arterial and venous blood pressures.

   3. Follow-up serial imaging (cranial US or CT) to detect progressive hydrocephalus (see Chapter 98).

1. Short-term outcome of GM/IVH is directly related to birthweight, gestational age, and the ensuing severity of the hemorrhagic insult to the immature brain. For the years 1995–1996, 1997–2002, and 2003–2007, very low birthweight infants in the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network saw survival steady at 84–85%. Meanwhile, the occurrence of severe IVH remained unchanged at 12%. A broader view of the 1997–2002 infants weighing 501–1500 g reveals 15% mortality, 25% survival with short-term complications of bronchopulmonary dysplasia/chronic lung disease, severe IVH, and necrotizing enterocolitis, with ∼60% survival for infants without the short-term complications; however, long-term outcomes remain in question.

2. Long-term major neurologic sequelae of GM/IVH depend primarily on the extent of associated parenchymal injury, laterality, and any added effects of the short-term complications. From the NICHD studies noted previously, long-term outcomes of seemingly normal extremely low birthweight infants at time of hospital discharge are worrisome. At 8–9 years of age, and in comparison with control term infants, studies now reveal strikingly lower IQ scores, greater learning problems, poor motor skills, markedly increased behavioral problems, and considerable hearing loss. Also, bilateral IVH has been proved to correlate with severe cerebral palsy and Mental Developmental Index <70 after 12 months of age, as opposed to unilateral IVH, in a multicenter retrospective study done by Maitre et al in 2009.