Although a rare event, the birth of conjoined twins has always fascinated both the physician and the layperson. The first well-documented case was reported in A.D. 1100 and described the Biddenden maids who were joined at the hips and shoulders. In 1134, when the maids had lived together for 34 years, Mary was suddenly taken ill and died. Eliza, her sister, died six hours later (Bondeson, 1992). The most famous conjoined twins were Eng and Chang Bunker, born in Siam in 1811. The inappropriate term “Siamese twins” was coined by P.T. Barnum, who promoted the exhibition of Chang and Eng Bunker. An early medical description of these most famous conjoined twins, who lived unseparated until they died at age 63, can be found in the works of Warren (1829). Many conjoined twins are stillborn. In one series 40% of conjoined twins were stillborn and an additional 35% survived only 1 day (Edmonds and Layde, 1982). Konig recorded the first successful separation of conjoined twins in 1689. These twins were joined at the umbilicus, and the division was accomplished by necrosing the band of tissue between the two children with a constricting ligature. Kiesewetter (1966) reviewed 24 surgical attempts at separation that appeared in the literature from 1689 to 1962. There are now over 100 reports of successful separations in the medical literature or lay press. Conjoined twins may be joined at a variety of anatomic sites, and classifications have been developed to describe all the possibilities (Guttmacher and Nichols, 1967). The nomenclature in use clinically is derived from the most prominent site of conjunction. The common twin types include thoracopagus, xiphopagus or omphalopagus, pygopagus, ischiopagus, and craniopagus.

Thoracopagus is the most common type of conjoined twin and with omphalopagus (or xiphopagus) represents about 75% of cases reported (Malone and D’Alton, 2000). The two individuals lie face to face and share a common sternum, diaphragm, and upper abdominal wall from xiphoid to umbilicus. An extensive review of the anatomyofthoracopa-gus twins has been published by Nichols et al. (1967). In data from 32 cases, 75% have conjoined hearts. Because of the abnormal ventricular arrangements and associated anomalies of the great arteries and veins, successful surgical division is usually not possible. In about half of these cases, the intestinal tracts are also joined. Occasionally, the esophagus and stomach are single, but usually the union starts in the distal duodenum and ends in a pouch at the site of a Meckel diverticulum. The biliary tree is joined in 25% of cases.

Xiphopagus or omphalopagus twins usually considered a subgroup of thoracopagus, also face one another and usually have the least complicated union of all conjoined twins. They are joined at the anterior abdominal wall from xiphoid to umbilicus. The peritoneal cavity of one communicates with that of the other, but the upper intestinal tracts are usually separate. A bridge of liver connects the infants in the majority of cases. Evaluation of the single umbilical cord in twins joined at the umbilicus has revealed the presence of two to seven umbilical vessels. An omphalocele is often present at the umbilical cord insertion.

Pygopagus twins represent about 20% of cases. They are joined at the buttocks and perineum, and face away from each other. A significant length of sacrum may be fused, and as a result, the twins often share the sacral spinal canal. A single lower rectum and anus is common, and often the lower genital tract and external genitalia are fused.

Ischiopagus accounts for 5% of cases. These twins are united at a single bony pelvis. Four normal lower extremities (ischiopagus tetrapus) may be attached to the pelvis, but often two of the four lower extremities are fused into one malformed limb (ischiopagus tripus). The intestinal tracts usually join at the terminal ileum, which empties into a single colon.

Craniopagus is the least common type of conjoined twins and accounts for 2% of cases. There is always fusion of the skull, and often the twins share large dural sinuses and vascular structures. A classification into partial or total forms, having a junction at brow, vertex, or parietal bone, has been devised by O’Connell (1976). In the partial forms, the brains are separated by bone or dura and each brain has separate leptomeninges. In the total form, the brains of each twin are connected, or they are separated only by arachnoid. Separation of the total type is extremely difficult, and feasibility is often determined by the presence of a superior sagittal sinus for each brain that will provide adequate venous drainage.

The exact incidence of conjoined twins is not known, but estimates have varied from 1 in 25,000 to 1 in 80,000 births (Siegel, 1950; Freedman et al., 1962). Other reports on the frequency of conjoined twinning show the incidence to be from 1 in 2,800 to 1 in 200,000 (Hanson, 1975). Three conjoined twins over a 10-month period among residents of South Glamorgan with approximately 5400 deliveries a year is the highest reported incidence of conjoined twinning so far described (Rees et al., 1993). In Rudolph et al., (1967) review, about 70% ofconjoined twins were female. Maternal age and parity do not appear to be factors that influence the occurrence of this type of twinning. However, use of assisted reproductive techniques may result in an increased risk for conjoined twins (Goldberg et al., 2000).

The first report of conjoined twins diagnosed ultrasonographically was in 1977 (Fagan, 1977). Since then prenatal diagnosis has been reported many times and can be made in the first trimester (Figure 121-1) (Hill, 1997; Hubinot et al, 1997; Lam et al, 1998; Maymon et al, 1998). Three-dimensional ultrasound examination may be a valuable adjunct to two-dimensional ultrasound examination when diagnosing conjoined twins (Figure 121-2) (Johnson et al., 1997; Bega et al., 2000; MacKenzie et al., 2004). A suspicion of conjoined twins should arise when a dividing membrane cannot be visualized. van den Brand et al., (1994) have suggested nine sonographic findings to diagnose conjoined twins (Table 121-1). Polyhy-dramnios is present in 75% of thoracopagus twins (Harper et al., 1980). Congenital anomalies are common in conjoined twins, even in the unshared organs. Congenital heart disease, renal and genitourinary abnormalities, intestinal duplication, and omphalocele have all been reported.

Ultrasound examination has been found useful in describing the extent ofjoining of the cardiovascular system in thoracoabdominally joined twins. Sanders et al. (1985) have reviewed their experience in four pairs of thoracoabdominally joined twins and demonstrated that prenatal echocardiography correctly diagnosed major cardiac anomalies, although they missed certain important features because of their inability to detect abnormal pulmonary venous connections. In addition, the investigators concluded that conjoined twins were more easily and thoroughly examined in utero because more views could be obtained as compared with the postnatal examination, which is hampered by the conjunction and the associated omphalocele. With the use of ultrasound equipment it is always possible to detect whether bone tissue is interposed between the two brains. However, it is impossible to detect if the two cerebral hemispheres are joined or separated by either the dura or pia mater. Color flow mapping may be useful in craniopagus twins to determine vascular connections (Loverro et al., 1991).

There are several pitfalls for the sonographic diagnosis of conjoined twins. Caution should be exercised in making a definite diagnosis of conjoined twins at less than 10 weeks of gestation because false-positive diagnoses have been documented (Usta and Awwaad, 2000; Weiss and Devine, 2002) Three-dimensional ultrasonography may help with the diagnosis (Babinski et al., 1999). Later in gestation, inseparable fetal skin contours must be a persistent finding at the same anatomic level to avoid the false-positive diagnosis of conjoined twins (Barth et al., 1990). Even discordant presentation does not exclude the diagnosis, particularly in omphalopagus twins. The joining bridge may be sufficiently small to allow for rotation of the twins. Finally, with severe conjoining, twins may be melded into a conglomerative tissue mimicking a single pregnancy (Weingast et al., 1984). The diagnosis of conjoined twins does not exclude the presence of other problems unique to twinning. For example, many cases have been reported of conjoined twins coexisting in triplet pregnancies (Sanjaghsaz et al., 1998; Sepulveda et al., 2003).

Little is known about the antenatal natural history of conjoined twins. Polyhydramnios has been reported in 75% of thoracopagus twins (Harper et al., 1980). In the older obstetric literature on the subject, stillbirth occurred in approximately 20% to 40% of cases (Harper et al., 1980). In one recent series of 14 sets of prenatally diagnosed conjoined twin pregnancies from a single center, the combination ofprenatal ultrasonography, echocardiography, and magnetic resonance imaging accurately defined the anatomy in all cases. In this series, 3 pregnancies were terminated, 1 resulted in intrauterine demise, and of the remaining 10 pregnancies delivered after viability, 5 individual fetuses survived (Mackenzie et al., 2002). Most of the contemporary literature has concentrated on the subject of surgical separation.