Intracranial Hemorrhage: Subdural, Subarachnoid, Subpial, Intraventricular (Term Infant), Miscellaneous

“Traumatic” Lumbar Puncture

The finding of bloody CSF in a newborn is common (occurring in greater than one-third of all CSF samples within a neonatal intensive care unit) and often is attributed to “traumatic” lumbar puncture. This conclusion primarily relates to the relative difficulty of performing the puncture in the newborn but also to the relative frequency of finding bloody CSF in infants without overt neurological signs. It is, however, likely that traumatic lumbar puncture in the newborn is much less common than is generally thought. For example, in a study in which lumbar punctures were performed on the third postnatal day in all premature infants of less than 2000 g, the 76 infants who had grossly bloody CSF with elevated protein content were evaluated by CT scan ( Table 26.3 ). Only 6 (8%) had no increased attenuation consistent with blood. Subarachnoid blood was detectable in 22 (29%), and intraventricular hemorrhage was noted in 48 (63%). There are few studies that have attempted to correlate the findings of red cells on CSF with the nature and extent of intracranial hemorrhage in either the preterm or term-born infant.

Cerebrospinal Fluid Findings of Intracranial Hemorrhage

CSF findings that indicate intracranial hemorrhage are, primarily, xanthochromia of the centrifuged fluid and elevations of the number of red blood cells (RBCs) and the protein content. Particular emphasis should be placed on the occurrence of combinations of findings rather than on a single, isolated abnormality.

Xanthochromia of the CSF develops within several hours after hemorrhage in older children and adults. In one particularly large study of adults with subarachnoid hemorrhage, nearly 90% exhibited xanthochromia within 12 hours of the ictus. The evolution of xanthochromia in newborns has not been studied systematically, although our impression is that it appears to occur more slowly than in older patients. This slower evolution may relate to a delay in the induction of the enzyme, heme oxygenase, which is located in the arachnoid and is responsible for the conversion of heme to bilirubin, the major pigment accounting for xanthochromia of the CSF. In adult rats, the activity of heme oxygenase reaches peak values 6 to 12 hours after injection of heme into the subarachnoid space. These data are closely comparable to the clinical observations with adult patients cited. Determination of the significance of xanthochromia in newborns is occasionally difficult in the presence of elevated serum bilirubin levels.

The number of RBCs that should be considered significant is difficult to state conclusively, in part because of the remarkably wide range of values considered normal (see Chapter 13 ). In studies of infants in neonatal intensive care facilities, median values of 100 to 200 RBCs/mm have been observed. A more recent study reported even higher values for mean RBC when the lumbar puncture was undertaken by residents. In a study of 184 cases, 64% of infants had RBC counts greater than 100,000. In the only report with ultrasonographic correlates, among 43 infants of less than 1500 g birth weight, the median value was 112, but the mean value was 785, and 20% of CSF samples had more than 1000 RBCs/mm. These infants did not exhibit ultrasonographic evidence of intracranial hemorrhage. However, exclusion of minor subarachnoid hemorrhage by cranial ultrasonography is not reliable. Thus the data indicate that findings of more than 100 RBCs/mm in the newborn are common, and in the very-low-birth-weight infants, values greater than 1000 occur in a substantial minority in the absence of apparently clinically significant intracranial hemorrhage. Again, the combination of findings is important in the evaluation.

Values for CSF protein are higher in newborns in an intensive care nursery than in older children. In the series of Sarff and coworkers, an average protein content in CSF of 90 mg/dL was observed for term infants, and a content of 115 mg/dL was observed for preterm infants. We have obtained similar data. In general, values for CSF protein are higher in the most premature infants; in one series, the mean value at 26 to 28 weeks of postconceptional age was 177 mg/dL, and at 35 to 37 weeks, it was 109 mg/dL. Values in intracranial hemorrhage are usually severalfold or higher than these. A recent study found that CSF protein concentrations increased by approximately 2 mg/dL for every 1000 CSF RBCs. In a recent meta-analysis of the normative values for CSF concentrations, four studies (one in preterm infants) investigated the protein concentration in CSF. The protein concentration was higher in preterm ( n = 53) compared with term infants ( n = 79), with the preterm mean protein level 144.6 mg/dL (range 52.0 to 300.0), and the term mean protein of 71.4 mg/dL (range 8.0 to 140.0).

Finally, determination of the CSF glucose level may be helpful in the diagnosis. In term and preterm infants evaluated in a neonatal intensive care unit and free of intracranial infection, the ratios of CSF to blood glucose levels are relatively high (i.e., 0.81 and 0.74, respectively). As with CSF protein levels, values for CSF glucose have been reported to be higher in the most premature infants; in one series, the mean value at 26 to 28 weeks was 85 mg/dL, and at 38 to 40 weeks, it was 44 mg/dL. However, in contrast to these findings a meta-analysis of CSF glucose concentrations from four studies did not report any difference in the mean glucose concentration between preterm (57.5 and 51.6 mg/dL, range 35.0 to 162.0) and term infants (51.2 mg/dL, range not specified). After neonatal intracranial hemorrhage, the CSF glucose level is frequently low ( Table 26.4 ). Indeed, in one study in which serial lumbar punctures were performed (for therapeutic purposes) on 13 infants with intraventricular hemorrhage, the CSF glucose concentration decreased on subsequent measurements in all the infants. Eleven of the 13 infants had CSF glucose values lower than 30 mg/dL at some point subsequent to the hemorrhage, and values of 10 mg/dL or less were common. The low values occurred as early as 1 day after the hemorrhage but usually became apparent between approximately 5 and 15 days after the hemorrhage. The depressed CSF glucose values persist for weeks and have been noted as long as 3 months after the hemorrhage.

The basis of hypoglycorrhachia is probably related to an impairment of the mechanisms of glucose transport into CSF. This impairment may occur at the level of the plasma membrane glucose transporter. Other proposed pathogeneses have included glucose use by RBCs or by contiguous brain. The former is ruled out by the lack of correlation between RBC number and CSF glucose level and by the negligible rates of glucose consumption observed when the cellular CSF is incubated in vitro. The possibility of excessive anaerobic use of glucose by contiguous brain rendered hypoxic-ischemic by hemorrhage, ventricular dilation, or other insult appears unlikely in view of simultaneous, serial determinations of CSF glucose and lactate. Thus, in 13 infants described with CSF hypoglycorrhachia, CSF glucose and lactate concentrations decreased pari passu; if anaerobic use of glucose had been operative, a concomitant increase in CSF lactate would have been expected. These observations favor the notion of a defect in glucose transport mechanisms.

An important practical problem arises when the low CSF glucose level is accompanied by pleocytosis and elevated protein content . This not uncommon occurrence is related presumably to meningeal inflammation from blood products and raises the question of bacterial meningitis. Although appropriate cultures are always indicated, and even initiation of antimicrobial therapy may be necessary (until results of cultures are known), the CSF formula of pleocytosis, depressed glucose, and elevated protein content is not infrequent after neonatal intracranial hemorrhage.

The optimal imaging procedure for diagnosis becomes apparent in the following discussions of the respective lesions, and the relative value of cranial ultrasonography, CT, and MRI and CT in diagnosis is reviewed in Chapter 13 . Suffice it to say here that cranial ultrasonography is often used as a screening procedure, MRI is the most effective methodology, and CT is used for a more rapid emergent approach. The features of MRI signal change over the days and weeks after neonatal parenchymal hemorrhage and are reviewed in Table 26.5 . The MRI changes relate primarily to changes in hemoglobin state, which proceed from predominately intracellular deoxyhemoglobin, to intracellular methemoglobin, to extracellular methemoglobin, and finally hemosiderin.

Tentorial Laceration

With major, lethal tears of the tentorium, hemorrhage is most often infratentorial. This finding is the case particularly with rupture of the vein of Galen or straight sinus or with severe involvement of the transverse sinus. The clots extend into the posterior fossa and, when large, very rapidly result in lethal compression of the brainstem. A massive infratentorial hemorrhage from a rupture of the vein of Galen also may occur without visible tear of the tentorium.

Lesser degrees of tentorial injury , with the advent of modern brain imaging techniques, are recognized now to be more common than the major lethal lacerations just described and probably are much more common than previously suspected. Thus several series, the largest and most recent described earlier, have documented a spectrum of intracranial hemorrhage, primarily subdural, associated with apparent or presumed tentorial injury. This spectrum, summarized in Table 26.7 , includes both infratentorial (usually retrocerebellar) subdural hemorrhage ( Fig. 26.5 ), secondary to inferior extension, and SDH , secondary to superior extension. It is important to note that the infratentorial, posterior fossa subdural hemorrhages may relate also to tear of cerebellar bridging veins, with or without accompanying overt tears of the tentorium. In addition to infratentorial or supratentorial extension, the hemorrhage of a tentorial tear may remain confined to the free edge of the tentorium , most often near the junction of the tentorium and falx ( Fig. 26.6 ), or it may extend anteriorly further into the subarachnoid space, velum interpositum, or ventricular system (see Table 26.6 ). The likely mechanism of tentorial hemorrhage after vacuum extraction can be understood with such vulnerability ( Fig. 26.6 ). Very minor varieties of this spectrum may account for the relatively high RBC counts in CSF in “normal” newborns (see earlier discussion).

Occipital Osteodiastasis

A prominent traumatic lesion in some infants who die after breech delivery is occipital diastasis with posterior fossa subdural hemorrhage and laceration of the cerebellum (see Table 26.6 and Chapter 40 ). The diastasis lesion consists of traumatic separation of the cartilaginous joint between the squamous and lateral portions of the occipital bone. In its most severe form, the dura and occipital sinuses are torn, resulting in massive subdural hemorrhage in the posterior fossa and cerebellar laceration. The bony lesion may be more common than has generally been recognized because it is missed easily at postmortem examination.

Falx Laceration

Laceration of the falx alone is distinctly less common than laceration of the tentorium and usually occurs at a point near the junction of the falx with the tentorium. The source of bleeding is usually the inferior sagittal sinus, and the clot is located in the cerebral fissure over the corpus callosum (see Table 26.6 ).

Superficial Cerebral Vein Rupture

Rupture of the bridging, superficial cerebral veins results in hemorrhage over the cerebral convexity, the well-known convexity subdural hematoma (see Table 26.6 ). The hematoma is usually more extensive over the lateral aspect of the convexity than near the superior sagittal sinus. Although convexity subdural hemorrhage is usually unilateral, bilateral lesions are not uncommon. Subarachnoid blood is a typical accompaniment. Convexity subdural hemorrhage is not a rare event, and, indeed, in small amounts, it is a frequent incidental finding at autopsy of the term infant. The trauma that leads to the hemorrhage may result also in cerebral contusion , which, in fact, may dominate the clinical picture.

Rapidly Lethal Syndromes

Tentorial laceration with massive infratentorial hemorrhage, an extremely rare disorder in the modern obstetrical era, is associated with neurological disturbance from the time of birth. The majority of the most severely affected infants weigh more than 4000 g at birth. Initially, the baby demonstrates signs of midbrain-upper pons compression (i.e., stupor or coma, skew deviation of eyes with lateral deviation that is not altered by doll’s eyes maneuver, and unequal pupils, with some disturbance of response to light). With such infratentorial hemorrhage, nuchal rigidity with retrocollis or opisthotonos may also be a helpful early sign. When these features are associated with bradycardia, a large infratentorial clot with brainstem compression should be suspected. Over minutes to hours, as the clot becomes larger, stupor progresses to coma, pupils may become fixed and dilated, and signs of lower brainstem compression appear. Ocular bobbing and ataxic respirations may occur, and finally, respiratory arrest ensues.

The severe clinical syndrome associated with occipital osteodiastasis resembles that described for major tentorial laceration. With occipital osteodiastasis, delivery is characteristically breech. A depressed Apgar score at 1 minute is common, and the course is one of rapid deterioration. In the six infants described by Wigglesworth and Husemeyer, the age at the time of death ranged from 7 to 45 hours.

Less Malignant Syndromes Associated With Posterior Fossa Subdural Hematoma

Less severe clinical syndromes accompany most examples of posterior fossa subdural hematoma currently encountered on obstetrical and neonatal services. These syndromes appear to result from smaller tears of the tentorium than those just noted, from rupture of bridging veins from superior cerebellum without tentorial tear, or, perhaps, from lesser degrees of occipital diastasis. The clinical syndrome consists of three phases. First, no neurological signs are apparent for a period that varies from several hours after birth (usually a difficult vacuum, forceps, or breech extraction or both) to as much as 3 or 4 days of age. Most commonly, the interval is less than 24 hours. Presumably, this period is associated with slow enlargement of the hematoma. Second, various signs develop referable to increased intracranial pressure (e.g., full fontanelle, irritability, “lethargy”). Most of these signs appear to relate to the evolution of hydrocephalus secondary to a block of CSF flow in the posterior fossa. Third, signs referable to disturbance of brainstem develop, including respiratory abnormalities, apnea, bradycardia, oculomotor abnormalities, skew deviation of eyes, and facial paresis. These deficits relate to direct compressive effects of the posterior fossa hematoma. In addition to brainstem signs, seizures occur in the majority of infants, perhaps because of accompanying subarachnoid blood. In infants who clearly worsen over hours or a day or more, as do approximately one-half, lethal brainstem compression may develop.

In more recent years, more common lesions of particularly small posterior fossa subdural hemorrhages have been identified by CT or MRI during the investigation of more subtle neurological abnormalities in term infants. In one carefully studied series of 26 small subdural hemorrhages detected by CT, 19 were infratentorial, and the leading clinical features were respiratory abnormalities (apnea, “dusky episodes”) in approximately 60% and neurological features (subtle seizures, hypotonia, apnea) in approximately 40%. None of the infants developed progressive neurological signs. Finally, two recent case reports describe the finding of vocal cord paralysis in infants who presented with stridor and respiratory distress and exhibited subdural hemorrhages diagnosed on subsequent MRI. Finally, as noted earlier, the commonest presentation of subdural hemorrhage in the term infant is to be completely asymptomatic to the clinical providers.