Hydrocephalus is a pathologic increase in intracranial cerebrospinal fluid (CSF) volume, whether intraparenchymal or extraparenchymal, independent of hydrostatic or barometric pressure (Raimondi, 1994). CSF is formed within the ventricular system—50% from the choroid plexus and 50% from the cerebral capillaries. The circulation of CSF is unidirectional (Vintzileos et al., 1983). It flows from the lateral ventricles through the foramen of Monro into the third ventricle. It then flows from the third ventricle to the aqueduct of Sylvius through the fourth ventricle into the spinal subarachnoid space (foramen of Magendie) or to the basal cisterns (foramen of Luschka) over the cerebral hemispheres. CSF is reabsorbed by arachnoid villi in venous sinuses. The flow of CSF is partially derived by arterial pulsations of the choroid plexus (Vintzileos et al., 1983).

Hydrocephalus may result from either fluid production that exceeds absorption or primary atrophy of the cerebral parenchyma. Most cases are due to mechanical obstruction to the flow of CSF at some level (DeLange, 1977). The site of obstruction may be inside the ventricular system (noncommunicating or internal hydrocephalus) or outside the ventricles (communicating or external hydrocephalus). Aqueductal stenosis comprises one-third of the cases of hydrocephalus in postnatal series; it is less common in prenatal studies. The aqueduct of Sylvius is the narrowest portion of the spaces through which the CSF flows, and aqueductal stenosis can be diagnosed following the finding of dilation of both lateral, and the third, ventricles (Raimondi, 1994; Davis, 2003).

Ventriculomegaly is a descriptive term of a pathologic process that has many causes. It may occur due to obstruction of CSF flow, or as a consequence of maldevelopment of the ventricle in anomalies such as agenesis of the corpus callosum (colpocephaly) or as an ex vacuo (destructive) phenomenon secondary to cerebral atrophy (Cardoza et al., 1988a). Ventriculomegaly is an indicator of underlying central nervous system (CNS) anomalies. It may also be the first sign of associated extra-CNS anomalies. The main causes of ventriculomegaly are aqueductal stenosis, Chiari II malformation (associated with meningomyelocele), Dandy–Walker malformation, agenesis of the corpus callosum, and fetal aneuploidy (D’Addario et al., 2007).

The hereditary nature of hydrocephalus was first appreciated by Bickers and Adams in 1949. It is now known that mutations in L1CAM, which result in a collection of X-linked conditions known as the L1 spectrum disorders, account for the majority of inherited cases of hydrocephalus (Zhang et al., 2006).

The incidence of isolated fetal ventriculomegaly is 0.5 to 1.5 per 1000 pregnancies. (Wilson et al., 1989; Wiswell et al., 1990; Davis, 2003). In a retrospective analysis of all liveborn and stillborn infants in U.S. Army hospitals between 1971 and 1987, 370 of 763,364 pregnancies had hydrocephalus (Wiswell et al., 1990). This equaled an incidence of 0.5 per 1000 total births, or 1 per 2063 total births. Of the infants with hydrocephalus, 37% had additional anomalies unrelated to the primary defect. No significant racial differences were seen, but an increased incidence of affected males was noted (64% male, 36% were female). A predominance of males is consistently observed in all studies of hydrocephalus due to the inherited X-linked forms of the abnormality.

Hydrocephalus is generally easily diagnosed prenatally due to the striking sonographic appearance of enlarged ventricles (Figure 16-1). It is important to note that the fetal biparietal diameter may not necessarily be increased when the ventricles are dilated. The lateral ventricles can be visualized as early as 12 weeks of gestation. Early in the second trimester, the choroid plexus is very echogenic and large relative to the volume of the cerebral hemispheres, normally filling the entire lateral ventricle posterior to the foramen of Monro. This tends to obscure the lateral ventricular wall to which it is closely applied (Fadel, 1989). This, together with the hypoechoic brain mantle, may falsely be interpreted as dilated ventricles filled with CSF. In second trimester fetuses with hydrocephalus, the first recognizable abnormality is generally the relative shrinkage of the normally prominent choroid plexus. Additionally, the choroid plexus appears to hang away from the ventricular walls, an appearance that is referred to as “dangling choroid.”

Several different methods have been proposed to quantitatively evaluate increased CSF. These methods include measurement of the ratio of the lateral ventricular width (LVW) to the hemispheric width (HW), (the LVW:HW ratio), or measurement of the ventricular atria. However, the LVW:HW ratio is not sensitive in early pregnancy. The first measurement of lateral ventricular dilation is the displacement of the medial wall of the lateral ventricle toward the midline, which will not change the LVW:HW ratio (Fadel, 1989). Furthermore, the echogenic outer lines originally thought to represent the lateral walls of the lateral ventricle in an axial scan of the fetal head have been shown by Hertzberg et al. (1987) to originate from deep cerebral veins. These reflections from small venous structures and deep fetal white matter may be displaced in the presence of hydrocephalus. The position of the fetal choroid plexus relative to the ventricular walls is dependent on gravity, and so the choroid angle (the angle between the long axis of the choroid plexus and the linear midline echo on a transverse axial sonogram through the body of the lateral ventricle), may be a useful indicator of ventricular size (Cardoza et al., 1988a). In normal-sized ventricles, this angle varies from 6 to 22 degrees, while in fetuses with ventriculomegaly the angle ranges from 29 to 90 degrees (Cardoza et al., 1988a). The choroid angle appears to increase with the severity of hydrocephalus.

The most common method for assessing ventricular size today is the atrial diameter measured on an axial sonogram through the fetal brain. In a study of 100 healthy fetuses between 14 and 38 weeks of gestation, the normal atrial diameter remained relatively constant throughout gestation despite growth of the surrounding brain (Cardoza et al., 1988b). These measurements were compared with 38 fetuses in whom ventriculomegaly had already been diagnosed. The mean diameter of the normal atrium was 7.6 ± 0.6 mm, and it was suggested that atrial diameters >10 mm, (greater than 4 SD above the mean), indicated the presence of ventriculomegaly. This measurement is quick to perform, is reproducible, and does not vary by gestational age (Cardoza et al., 1988b).

While an atrial diameter of >10 mm is considered abnormal, the term borderline ventriculomegaly is often used to refer to an atrial measurement of between 10 and 12 mm. Borderline ventriculomegaly is associated with an increased risk of CNS and non-CNS abnormalities, and suggests the need for a more detailed fetal anatomic examination. Several studies have substantiated this recommendation. Mahony et al. (1988) performed a prospective study of 20 fetuses with apparently isolated borderline ventriculomegaly (defined as a 3-to 8-mm separation existing between the choroid plexus in the atrium of the lateral cerebral ventricle and the adjacent ventricular wall on an axial scan). Of the 20 fetuses identified, 8 had a normal outcome (40%). The remaining 12 fetuses had additional sonographic abnormalities. Of these 12, 4 had an uncertain prognosis and 8 died. In another study of 55 fetuses with borderline ventriculomegaly, 13 had isolated ventriculomegaly and 42 had associated abnormalities (Goldstein et al., 1990). Of 15 living children who could be identified, 9 (60%) were normal at 6 to 30 months of postnatal age, 3 (20%) were abnormal, and 3 (20%) were lost to follow-up. These authors concluded that borderline ventriculomegaly, when isolated, was associated with a better prognosis than more substantial ventriculomegaly. A literature review in 1993 found 109 cases of mild ventriculomegaly, 92 of which were isolated, and 11 of these had an abnormal karyotype (12%) (Achiron et al., 1993). In another retrospective study of 44 fetuses with borderline ventriculomegaly (10–12 mm), 17 fetuses (39%) had other sonographic abnormalities, 6 of which were in the CNS (Bromley et al., 1991). Of 36 liveborn neonates in this study, 26 were developmentally and clinically normal at 3 to 18 months of postnatal age, while 10 of the 36 were developmentally impaired, including 5 cases from apparently isolated borderline ventriculomegaly.

Unilateral hydrocephalus is extremely uncommon. This condition carries a better prognosis than bilateral hydrocephalus. Of eight reported prenatal cases, five were due to congenital absence or stenosis of the foramen of Monro, one was due to transient obstruction of CSF flow by an intracranial hematoma, one was due to holoprosencephaly, and one was due to unknown causes (Patten et al., 1991). It appears that unilateral hydrocephalus is not associated with the presence of extracranial anomalies, and such patients generally have required postnatal placement of a ventriculoperitoneal shunt (Patten et al., 1991; Anderson et al., 1993; Chari et al., 1993).

The differential diagnosis for hydrocephalus and ventriculomegaly includes aqueductal stenosis, meningomyelocele (Chiari II malformation), Dandy–Walker malformation, agenesis of the corpus callosum, aneuploidy, intrauterine infection (cytomegalovirus, toxoplasmosis, or syphilis), intracranial hemorrhage, CNS tumor, hydranencephaly, porencephaly, and holoprosencephaly. Meningomyelocele can be the underlying reason for hydrocephalus due to herniation of the cerebellum through the foramen magnum resulting in the Chiari II malformation. The fetal spine should therefore be carefully examined in multiple planes in all such cases (see Chapter 19).

In hydranencephaly, only remnants of the cortex remain, with fluid-filled sacs that are lined by leptomeninges replacing the rest of the brain (see Chapter 15). In porencephaly, the parenchyma of the brain contains one or more fluid-filled cavities (see Chapter 21). The Dandy–Walker malformation includes ventricular dilation (see Chapter 11). Demonstration of the absence of falx cerebri and fusion of the thalami can distinguish holoprosencephaly from other causes of hydrocephalus (see Chapter 14). Consideration should also be given to the possibility of underlying single-gene disorder.

One of the major concerns with the antenatal sonographic finding of hydrocephalus is the fact that ventriculomegaly is frequently associated with additional anomalies within and outside of the fetal brain. Associated intracranial anomalies (such as agenesis of the corpus callosum or the Dandy–Walker malformation) are present in at least one-third of cases. Extracranial abnormalities have been demonstrated in two-thirds of cases (Fadel, 1989). In another series of 61 cases of hydrocephalus, 51 fetuses (84%) had associated abnormalities, 34 of which were extracranial and 27 had multiple malformations (Nyberg et al., 1987). In this series, false-negative diagnoses were common, and included esophageal atresia, spinal abnormalities, lung hypoplasia, and cardiac defects. Of 61 fetuses, 41 ultimately died from associated conditions such as asphyxiating thoracic dystrophy (Jeune syndrome), Apert syndrome, in which hydrocephalus is considered a major associated malformation (Hyon et al., 1986), and the Walker–Warburg syndrome. This latter recessively inherited condition includes hydrocephalus, multiple CNS malformations, microophthalmia, severe mental retardation, congenital myopathy, and a limited life span (Crowe et al., 1986). Finally, inborn errors of metabolism, such as fumarase deficiency, can cause polyhydramnios and hydrocephalus in utero (Remes et al., 1992). In this condition, enlargement of the cerebral ventricles is due to cerebral atrophy, while the associated polyhydramnios is due to intrauterine hypotonia, causing poor swallowing.