When congenital malformations occur in a multiple gestation, management decisions can be difficult for both parents and physicians because the fates of such sibling fetuses are necessarily linked. Given that the incidence of multiple gestations is increasing in industrialized countries, primarily because of assisted reproductive technologies (Jewell and Yip, 1995), the management dilemma for twins with malformations will also inevitably increase. Congenital malformations occur more commonly in twins as compared with singleton gestations and are an important contributor to the increased perinatal mortality associated with twin gestation. Some malformations in twins are inherent to the monozygotic twinning process, such as acardiac twinning, conjoined twins, and twin-to-twin transfusion syndrome. These malformations are discussed separately (see Chapters 119–121). Other malformations in twins are deformation abnormalities, such as clubfoot, which result from the crowding of the intrauterine environment.

Using data from the British Columbia Health Surveillance Registry, the incidence of congenital malformations in twin pairs was estimated at 6% (Schinzel et al., 1979). In this series, the incidence of congenital anomalies was 2.5 times more common in monozygotic twins than in dizygotic twins or singletons. The incidence of chromosomal abnormalities is increased twofold in dizygotic twins as compared with agematched singleton pregnancies, but the incidence of nonchromosomal abnormalities is not increased (Drugan et al., 1996).

In one series of 1424 twin pairs, 445 pairs were monozygotic, 26 of which (6%) had congenital malformations (Cameron et al., 1983). Even among monozygotic twin pairs with malformations, however, the majority of fetuses will be discordant for the abnormality, with only 6 of the 26 twin pairs (23%) concordant (Cameron et al., 1983). In another registry of 4490 twins, there was a 50% increase in incidence of congenital malformations among twins as compared with singletons, 4.9% versus 3.3%, respectively, with this increase almost entirely limited to same sex—and therefore presumably monozygotic—twin pairs (Layde et al., 1980).

The incidence of open neural tube defects in twins is controversial. In one study from California and Norway, anencephaly and encephalocele, but not meningomyelocele, were found more commonly in twins as compared with sin-gletons (Windham et al., 1982). In a national survey from England and Wales, there was a threefold increased incidence of anencephaly in twins as compared with singletons, while the incidences of encephalocele and meningomyelocele were similar in twins and singletons (Doyle et al., 1991). In a further study from Spain, the incidence of anencephaly was 0.1% for twins as compared with 0.03% for singletons, and this increase was almost entirely in same-sex twin pairs (Ramos-Arroyo, 1991). However, in a study from Northern Ireland, the incidence of anencephaly was decreased in twins, while the incidence of meningomyelocele was similar to that of singletons (Little and Nevin, 1989a). The incidence of congenital cardiovascular malformations was also found to be modestly increased in twins as compared with singletons, 0.9% versus 0.7%, with the increase again being confined to same-sex twin pairs (Little and Nevin, 1989b).

The ability of sonographic examination to detect congenital malformations in twin pregnancies has not been adequately evaluated, with no large patient series available for review. In one series of 33 fetuses from twin pregnancies with anomalies, none of the 8 cardiac anomalies were diagnosed prenatally, while 11 of 20 (55%) other major anomalies and none of the 12 minor anomalies were diagnosed prenatally (Allen et al., 1991). In another series of 24 fetuses with anomalies, it was concluded that by using serial ultrasonography in a tertiary care center it may be possible to achieve an 88% detection rate, with 100% specificity, in the prenatal diagnosis of anomalies in twin pregnancies (Edwards et al., 1995). However, almost 40% of fetuses with anomalies in this series were not diagnosed until 24 weeks of gestation or later, when options for management are more limited. It has also been suggested that up to 88% of cases of Down syndrome in twin pregnancies could be detected by combining risks derived from maternal age and sonographic nuchal translucency thickness measurement at 10 to 14 weeks of gestation (Sebire et al., 1996).

The differential diagnosis following the prenatal diagnosis of a congenital malformation in a twin pregnancy is extensive and is dependent on the particular abnormality that is suspected. For differential diagnoses for individual malformations, the corresponding chapters in this textbook should be consulted. Judging from the few published series of sonographic diagnosis of congenital malformations in twins, there is a high degree of specificity involved in the correct prenatal identification of anomalies (Allen et al., 1991; Edwards et al., 1995).

Specific information on the antenatal natural history of individual malformations is available in the corresponding chap-tersin this textbook. The history of such pregnanciesbecomes even more complicated when, as in the majority of cases, there is discordance for the malformation in question. In a series of 14 expectantly managed twin gestations in which only one fetus had a congenital malformation, Malone et al. (1996) found a significantly increased (20%) risk of preterm delivery in such pregnancies, in addition to the baseline risk of prematurity already seen with normal twin gestations. This was attributable to the presence of a fetus with an anomaly. Birth weight was significantly lower, and both cesarean delivery rate and perinatal mortality rate were increased in anomalous twin pregnancies as compared with control twins. These data were subsequently confirmed by Alexander et al. (1997) and Gul et al. (2005), both of whom found both lower gestational age at delivery and lower birth weight in their small series of expectantly managed anomalous twin pregnancies as compared with control twins.

When both fetuses in a twin pregnancy are concordant for malformations, subsequent management of that pregnancy is straightforward and should involve both the usual obstetric management of twin gestations and any required interventions for the particular malformation. However, pregnancy management becomes considerably more complex when one twin has a congenital malformation but the co-twin is normal. Counseling of parents depends on the type of abnor-mality and the prognosis for the twin with the anomaly as well as on the likely outcome for the normal co-twin. Three management options are available in this situation: expectant management, selective termination of the anomalous fetus, and termination of the entire pregnancy (Malone and D’Alton, 1997; Rustico et al., 2005). Selective termination of an anomalous twin is described in detail under “Fetal Intervention.”

For pregnancy management, we recommend karyotyping for twins with malformations. As already described, parents should be counseled that expectant management of an anomalous twin pregnancy is associated with an increased risk of preterm delivery. In addition to the risk of prematurity, expectant management can also be complicated by intrauterine death of the anomalous fetus, which can have profound implications for the well-being of the normal co-twin, especially in a monochorionic twin gestation. If one fetus in a monochorionic twin pair dies, there may be a greater than 20% risk to the remaining co-twin of developing multicystic encephalomalacia, leading to profound neurologic handicap (Pharoah and Cooke, 1997; Pharoah and Adi, 2000). This risk is present from the moment of the death of the first twin, and may not be predictable, even by intensive surveillance with sonographic examination or fetal heart rate monitor-ing. Therefore, serious consideration should be given to delivering the twins if an anomalous fetus in a monochorionic gestation appears to be in a premorbid condition. This deci-sion will depend on the gestational age, so that the risks of neurologic morbidity associated with expectant management are balanced against the iatrogenic risk of prematurity associated with delivery. If an anomalous twin in a monochorionic twin pair has already died in utero, close fetal surveillance for the surviving co-twin is recommended, although it must be realized that this may not prevent neurologic morbidity, which may already have occurred. Similarly, delivery immediately after diagnosis of the intrauterine death of one twin may not protect against neurologic morbidity in the surviving co-twin. Delivery at 37 weeks, or after measuring lung indexes consistent with maturity, is reasonable in such situations.

Antenatal surveillance of twin gestations complicated by congenital malformations should follow the usual recommended fetal surveillance practices for normal twin pregnancies. This includes serial sonographic examinations every 3 to 4 weeks, from approximately 18 weeks for fetal growth, or every 2 weeks if growth restriction or growth discordance greater than 20% is present. More intensive fetal surveillance with nonstress tests and Doppler velocimetry is recommended for cases of growth restriction or significant growth discordance. Additional fetal testing may also be indicated depending on the particular type of congenital malformation present.

The decision on location of delivery will depend entirely on the nature of the congenital malformation, presence of associated anomalies, and availability of postnatal therapies. With regard to mode of delivery, the choice between vaginal and cesarean delivery will also be dictated by the individual malformation and fetal prognosis. This is described in detail in the appropriate section in the chapters describing each abnormality. In addition, the usual obstetric indications for determining mode of delivery of normal twin gestations may also apply. Typically, all vertex/vertex twins are candi-dates for vaginal delivery, while most obstetricians perform a cesarean delivery if the presenting twin is nonvertex. When the presenting twin is vertex and the second twin is non-vertex, a vaginal breech delivery of the second twin is generally acceptable if the estimated fetal weight is greater than 1500 g. There are insufficient data to confirm the optimal mode of delivery when a nonvertex second twin weighs less than 1500 g.