The overall incidence of PV-IVH has declined over the past three decades; however the occurrence of severe hemorrhage remains substantial, particularly in the tiniest of the very-low-birth-weight population [1–3] (Table 11.1). Thus, approximately 26 % of infants of a gestational age between 23 and 25 weeks and 12 % between 26 and 27 weeks still develop the most severe forms of hemorrhage [3]. This is highly relevant because as survival of infants born at the cutting edge of viability continues to increase, severe hemorrhage rates are likely to increase, and with them long-term neurodevelopmental deficits.

The primary lesion is bleeding from vessels within the periventricular subependymal germinal matrix located between the caudate nucleus and thalamus at the level of the foramen of Monro [2] (Fig. 11.2). The matrix is a transient gelatinous region that provides poor support for a large, immature network of blood vessels primarily supplied by Heubner’s artery, a branch of the anterior cerebral artery [4]. The venous drainage also appears to be important with regard to the risk for bleeding. Thus the drainage includes the terminal, choroidal, and thalamo-striate veins that lead to the internal cerebral vein. The blood flow then makes a “U-turn” in the subependymal region at the level of the foramen of Monro, where most of hemorrhage originates [2]. This creates the potential for obstruction of the venous drainage resulting in venous distention with obstruction of the terminal and medullary veins, with subsequent rupture. The primary vascular source of PV-IVH has not been clearly established. Thus the initial descriptions suggested a venous origin of hemorrhage [2] while subsequent pathologic studies described it as emanating from capillary or arterial vessels [4]. Based on studies conducted over the years it is most likely, that PV-IVH results from forces acting upon both the arterial and venous circulations [1, 2].

The germinal matrix is an active site of active cellular proliferation, and is a source of both neuronal precursors early in gestation as well as glial elements that become oligodendroglia and astrocytes in the third trimester [2]. The matrix involutes after approximately 32 weeks and by term gestation, this region is essentially absent. Destruction of the matrix as a result of hemorrhage may result in impairment of myelination, brain growth, and subsequent cortical development [2].

The hemorrhage may be confined to the germinal matrix region (grade 1 IVH) or it may extend and rupture into the adjacent ventricular system (grade II or III IVH) depending on the extent of blood, or may extend into the adjacent white matter (termed a grade IV IVH or intraparenchymal echogenicity (IPE)) [2, 5, 6] (Fig. 11.1). The latter lesion, which is often unilateral, represents an area of hemorrhagic necrosis of varying size within periventricular white matter, dorsal and lateral to the external angle of the lateral ventricle [2, 6, 7].

In general terms, three factors appear to be central to the genesis of hemorrhage. The first relates to inherent vulnerability of the germinal matrix, with immature vessels and poorly supportive gelatinous matrix as described above [2]. The second relates to the concept of a pressure passive circulation often referred to as loss of autoregulation, and the third relates to perturbations in cerebral blood which are common in the sick premature infant.

Autoregulation is defined as a state whereby CBF remains constant as cerebral perfusion pressure varies over a certain range [8]. The mean arterial pressure at which CBF decreases during hypotension is termed the lower limit, and the arterial pressure at which CBF increases is called the upper limit of autoregulation. In newborn animals, the range of blood pressures at which CBF remains constant is lower than in adult animals [9, 10]. More importantly, in fetal lambs, the resting blood pressure is only slightly higher than the lower limit of the autoregulatory curve [10]. If a similar situation exists in newborn infants, moderate hypotension could result in reduced CBF and create the potential for ischemic brain injury. Evidence has indicated that the autoregulatory response remains intact in human neonates with mild asphyxiation or who have minimal respiratory distress [2]. By contrast seminal data of Lou and colleagues [11] derived from 19 newborn infants with varying degrees of respiratory distress studied in the first hours of life were the first to indicate that CBF varied considerably with spontaneous variations in blood pressure, suggesting that autoregulation was, in fact, lacking. Subsequent studies over the subsequent years utilizing different methods to assess CBF or cerebral blood flow velocity including Doppler, near-infrared spectroscopy, and xenon have supported the concept of a pressure passive cerebral circulation in the sick infant; that is, CBF varies directly with changes in systemic blood pressure [12–15].

To summarize, although intact cerebral autoregulation has been documented in the premature infant, it appears to function within a limited blood pressure range, and is likely to be absent particularly in the sick hypotensive preterm infant [12–15]. This clinical state places the developing brain at great risk for injury during times of hypotension and or elevated blood pressures .

Evidence has pointed to a critical role for intravascular factors and specifically perturbations in ABP as a major mechanism of capillary rupture and hemorrhage. First, the cerebral circulation of the sick infant was considered pressure passive (see above) [12–15]. Second, experimental studies indicated that germinal matrix hemorrhage could be produced by systemic hypertension with or without prior hypotension [16, 17]. Third, infants with lower mean ABP in the first postnatal days and infants who received rapid volume expansion to correct hypotension were more likely to develop IVH [18–20].

In the original pathologic description of IVH, elevations in venous pressure were proposed as an important source of hemorrhage, based in part on the previously described venous drainage of germinal matrix and white matter [2]. Indeed simultaneous increases in venous pressure were observed in infants who exhibit variability in ABP, such as with respiratory distress syndrome (RDS) and associated complications, e.g., pneumothorax or pulmonary interstitial emphysema, or with mechanical or high-frequency ventilation [21–23].

To summarize, the cumulative data suggested that there is a significant contribution from both arterial and venous perturbations to the development of IVH. Over the years, studies have indicated that these intravascular responses can be modulated by additional factors, i.e., inflammation associated with chorioamnionitis (negative influence) [24] or the administration of glucocorticoids to the mother (positive influence) [25, 26].

The cause of the WMI associated with hemorrhage still remains unclear. The original hypothesis was that the intraparenchymal lesion represented an “extension” of hemorrhage from the germinal matrix or lateral ventricle into previously normal periventricular white matter. However subsequent neuropathologic data indicate that the intraparenchymal lesion represents an area of hemorrhagic necrosis [2, 5, 7]. The WMI appears to be closely linked to the adjacent hemorrhage and two potential pathways have been proposed to explain this intricate relationship. The first suggests a direct relationship to the hemorrhage based on several clinical observations. First, the WMI is always noted concurrent with or following a large GM and/or IVH, and is rarely if ever observed prior to the hemorrhage, and second the WMI is always observed ipsilateral to the side of the larger hemorrhage with bilateral bleeding [1, 2]. This consistent relationship between the GM and WMI may in part be explained by the venous drainage of the deep white matter as described previously. A second explanation is that the WMI evolves de novo and the IVH and WMI occur concurrently. Since both the GM and the periventricular white matter are border zone regions, the risk for ischemic injury increases during periods of systemic hypotension, particularly in the face of a pressure passive cerebral circulation [1, 2]. Hemorrhage in these regions may then occur as a secondary phenomenon i.e. reperfusion injury. In support of this theory is the fairly consistent observation of the simultaneous detection of PV-IVH and WMI by cranial ultrasound imaging. Moreover elevated hypoxanthine and uric acid levels, perhaps as markers of reperfusion injury, have been observed on the first postnatal day in infants who subsequently developed WMI [27, 28].

The infant continued to exhibit moderate respiratory distress and developed a left sided pneumothorax requiring chest tube placement. The arterial blood gases showed moderate hypercarbia (pCO₂ 55 to 58 mmHg and mild acidosis—pH 7.20–7.26). The hematocrit continued to fall to 22 % on DOL 4. A repeat cranial sonogram revealed more prominent “echoes” within the left fronto-parietal white matter, and the lateral ventricles were more distended as a result of increased hemorrhage. Serial sonograms over the first 6 weeks demonstrated progressive enlargement of the lateral, third, and fourth ventricles (Fig. 11.3).

Question

The potential contribution of hypercarbia has only in recent years been clearly delineated in premature infants [31]. Thus in a study undertaken in the first week of life in VLBW infants, increasing PaCO₂ resulted in an increase in CBF and progressive impairment of cerebral autoregulation [31]. Hypercapnia defined by a maximum PaCO₂ recorded during the first 3 days of life has also been associated with severe IVH. In a retrospective cohort study of 574 VLBW infants, as the maximum PaCO₂ increased from 40 to 100 mmHg, the probability of severe IVH increased from 8 to 21 % [32]. In a single-center retrospective review of 849 infants weighing <1250 g extremes in PaCO₂, both low and high, as well as fluctuations in PaCO₂ during the first 4 days of life increased the risk of severe IVH [33]. These cumulative observations indicate that extremes in PaCO₂ should be avoided during the period in which infants are at high risk of IVH. This is highly relevant since permissive hypercapnia has been advocated as a ventilator strategy to minimize barotrauma to the lungs of preterm infants, and thus prevent the evolution to chronic lung disease [34] (Fig. 11.4).

The introduction of surfactant administration to reduce the severity of RDS in the late 1980s was the most important development in the management of sick premature infants. Since surfactant reduced the severity of RDS as well as the rate of pneumothorax, it was postulated that it would also lead to a decrease in the rates of IVH. However, this was not borne out in the many surfactant trials [35]. This lack of effect has been attributed to the impact of surfactant administration on fluctuations of CBF in sick preterm infants who have pressure passive circulation and the rapid changes of PaCO₂ and PaO₂ that may subsequently lead to brain injury [36, 37].

PHH complicates approximately 40–50 % of severe IVH [2]. The progression of PHH may be rapid, i.e., days, and in such cases, it is usually associated with raised intracranial pressure . The obstruction under such circumstances appears to be secondary to an impairment of cerebrospinal fluid absorption caused by particulate clot usually demonstrated by ultrasound scan. This may be at the foramen of Monroe or the aqueduct of Sylvius. Treatment is usually immediate, via external drainage above the site of obstruction. More commonly, a communicating hydrocephalus evolves from 1 to 4 weeks following the diagnosis of hemorrhage. The hydrocephalus is commonly secondary to an obliterative arachnoiditis distal to the outflow of the fourth ventricle. The clinical criteria of evolving hydrocephalus, i.e., full anterior fontanel and rapid head growth, do not appear for days or weeks after ventricular dilation has already been present. Possible explanations for this discrepancy include the relative excess of water in the centrum semiovale and the relatively large subarachnoid space. Thus, serial scans are critical to follow for the evolution of PHH.