The term holoprosencephaly describes a spectrum of cerebral and facial malformations that result from absent or incomplete division of the embryonic forebrain, the prosencephalon. The abnormality occurs during the 3rd week of gestation (Müller and O’Rahilly, 1989). Two separate sets of terms are used to describe the facial and brain anomalies. DeMyer (1964) proposed a subclassification of holoprosencephaly based on the extent of sagittal division of the cerebral cortex, thalamus, and hypothalamus. In the most severe form, alobar holoprosencephaly, midline structures are absent and there is no division of the hemispheres. A single common ventricle is present and the thalami are fused. In semilobar holoprosencephaly, incomplete division of the forebrain results in partial separation of the hemispheres. In lobar holoprosencephaly, there is normal cortical division and two thalami, but abnormalities exist in the corpus callosum, septum pellucidum, or olfactory tract or bulbs. More recently, a fourth subtype of holoprosencephaly was described, known as the middle interhemispheric variant (MIH) (Simon et al., 2002; Pulitzer et al., 2004). In MIH, the posterior frontal and parietal lobes fail to separate, but the poles of the frontal and occipital lobes are well separated.

The facial abnormalities accompanying holoprosencephaly range from subtle to grotesque (Figure 14-1). In general, the more severe facial malformations are associated with alobar holoprosencephaly, but exceptions do occur (Table 14-1). The most severe facial malformation is cyclopia, a single or fused double eye and absent nasal structures (Figures 14-1A and 14-1B). A proboscis, a cylindrical protuberance, may also be present (Figure 14-1C). In ethmocephaly, the eyes are separate but closely placed (hypotelorism); a proboscis is present (Figure 14-1D). Ethmocephaly is the rarest of the facial malformations seen in holoprosencephaly. In cebocephaly, ocular hypotelorism is present along with a nasal structure that has a single nostril (Figure 14-1E). In the milder forms of holoprosencephaly, ocular hypotelorism, flat nose, and a median cleft lip occur (Figure 14-1F). Arrhinencephaly refers to the absence of the olfactory tracts and bulbs. Subtle, often missed forms of holoprosencephaly include mild hypotelorism, eye abnormalities (iris or retinal colobomas), mild midface hypoplasia, bifid uvula, and single central maxillary incisor tooth (Figure 14-1G).

The true incidence of holoprosencephaly is unknown, as presumably there is a high incidence of death in early embryonic life. In a study of 36,380 spontaneous abortions, Matsunaga and Shiota (1977) noted 150 embryos with holoprosencephaly. This equals a prevalence of 40 per 10,000. Using data from a population-based birth defects registry from California, Croen et al. (1996) identified 121 cases among a cohort of 1,035,386 livebirths and fetal deaths. This equaled a prevalence of 1.2 per 10,000 births. Of all cases, 41% (50 of 121) had a chromosomal abnormality, most commonly, trisomy 13. Among cases with normal chromosomes, increased risks are observed among Hispanic and Pakistani women (Ong et al., 2007). There is an excess of female conceptuses in alobar (3:1), as opposed to lobar (1:1) holoprosencephaly (Cohen, 1989a). Prevalence is also increased in twins (Suslak et al., 1987). A more recent U.K. population-based survey included all cases in which the pregnancy was terminated after prenatal diagnosis (Bullen et al., 2001). The total prevalence (including terminations) was 1.2 cases per 10,000 registered births. The birth prevalence (affected livebirths and stillbirths at >24 weeks) was 0.49 cases per 10,000 births. Prenatal detection was 86% of cases with a program of routine second-trimester anatomic survey.

Maternal diabetes is the only confirmed human teratogen associated with holoprosencephaly; the incidence of holoprosencephaly among diabetics is increased 200-fold as compared with controls (Barr et al., 1983; Cohen and Shiota, 2002). Other postulated teratogens include ethanol (Jellinger et al., 1981; Ronen and Andrews, 1991; Cohen and Shiota, 2002), salicylate (Benawra et al., 1980), cytomegalovirus (byrne et al., 1987), retinoic acid, and products of the cholesterol biosynthesis pathway (Cohen and Shiota, 2002).

In most fetuses, a midline echo is normally generated by the reflection of sound waves from acoustic interfaces at the interhemispheric fissure. This echo is absent in alobar holoprosencephaly (Figure 14-2) (Pilu et al., 1987). In addition, in the standard transverse view of the fetal head obtained during measurement of the biparietal diameter (BPD), ventriculomegaly may be noted. The extent of thalamic fusion is best evaluated on a coronal scan (Figure 14-3). When a large cystic abnormality of the fetal head is detected, evaluation of the midline structures of the fetal face is recommended. Greene et al. (1987) have described the use of two sonographic criteria for the diagnosis of holoprosencephaly: intracranial abnormalities and structural abnormalities of the face. In any case of suspected holoprosencephaly, the bony orbits should be visualized. Normal standards exist for the distance between orbits; this is important in the diagnosis of ocular hypotelorism (Mayden et al., 1982). Facial views will also demonstrate the presence of a proboscis. First trimester holoprosencephaly has been diagnosed by two and three-dimensional transvaginal sonography (Bronshtein and Wiener, 1991; Hamada et al., 1992; Stagiannis et al., 1995; Tongsong et al., 1999; Lai et al., 2000; Chen et al., 2005). The first trimester cross-sectional view of the fetal brain demonstrating both choroid plexuses (the so-called “Butterfly sign”) has been advocated as a screen for holoproscencephaly (Sepulveda et al., 2004).

Several studies have shown that extracranial abnormalities occur in approximately 50% of cases (Berry et al., 1990; McGahan et al., 1990). The most common anomalies are meningomyelocele, renal dysplasia, cardiac defects, and polydactyly.

The main considerations in the differential diagnosis of holoprosencephaly include hydrocephalus, midline cerebral defects, such as septo-optic dysplasia, hydranencephaly, and porencephalic cysts. Holoprosencephaly may be distinguished from fetal hydrocephalus by demonstration of the absence of the midline echo and fusion of the thalami. Hydrocephalus due to aqueductal stenosis or an Arnold–Chiari malformation displays an intact falx cerebri, distinct and separate ventricles, and splayed thalami (Nyberg et al., 1987). In hydranencephaly, a thinned cerebral cortex may be noted. Finally, although both hydranencephaly and porencephalic cyst can demonstrate an absent or deviated falx, the thalami should be distinct with these diagnoses (Nyberg et al., 1987). Holoprosencephaly can occur as an isolated finding or in combination with other anomalies in a single-gene disorder, such as Smith–Lemli–Opitz syndrome (Peebles, 1998).

Holoprosencephaly is highly lethal during fetal life. It has been estimated that only 3% of holoprosencephalic conceptuses are liveborn (Cohen, 1989b). In a retrospective study, 40% of cases were associated with vaginal bleeding and were considered threatened miscarriages (Berry et al., 1990). The perinatal mortality rate is on the order of 89% (McGahan et al., 1990). It is important to recognize, however, there is a correlation between the severity of the condition and the outcome.

Holoprosencephaly is a profound fetal brain anomaly that cannot be altered or treated. Considerations for management of pregnancy include elective termination if the diagnosis is made earlier than 24 weeks, determining the cause of the holoprosencephaly, and planning the route of delivery. Because approximately 30% to 50% of fetuses with holoprosencephaly have chromosomal abnormalities, prenatal karyotype is strongly recommended. A chromosomal abnormality is more likely to be present if extrafacial abnormalities are detected on sonography (Berry et al., 1990; Bullen et al., 2001). Trisomy 13 accounts for ∼75% of the chromosome abnormalities. A wide variety of other chromosomal abnormalities have also been described (Aronson et al., 1987; Gillerot et al., 1987; Münke, 1988, 1989; Urioste et al., 1988; Helmuth et al., 1989; Estabrooks et al., 1990; Isada et al., 1990; Lurie et al., 1990; Hamada et al., 1991; Hatziioannou et al., 1991; Kuchle et al., 1991; Petit et al., 1991; Van Allen et al., 1993).