Ultrasound has revolutionized the care of women carrying twins. It begins with the diagnosis of twins and continues through the determination of chorionicity, cervical length measurement in predicting preterm delivery of twins, ultrasound screening for structural abnormalities and aneuploidy, management of complications unique to twins, and the intrapartum role of ultrasound in vaginal twin deliveries.13 The problems found in twins are also seen in the increasing number of higher order multiples born in the United States and throughout the world.

Upon learning of a twin gestation, most parents are struck with joy and awe. The obstetrician and sonographic practitioner, however, are faced with numerous problems including prematurity, deceased survival, impaired growth, increased structural abnormalities, and aneuploidy.16 Approximately 14% to 25% of twins are growth restricted and 25% require admission to neonatal intensive care units.7,8 The risk of cerebral palsy is four times greater for twins than singletons, and 17 times greater for triplets than singletons.9,10 The fetal death rate for twins of 15.5/1000 live births is almost four times higher than that of singletons, which is 4.3/1000 live births.11 The likelihood of not surviving the first year of life is 7-fold greater for twins than for singletons.12,13 Twin-specific problems include twin-to-twin transfusion syndrome (TTTS), monochorionic-monoamniotic twin placentation and variations of conjoined twins. Management decisions for one twin must consider the impact on the co-twin.14,15 Maternal complications such as preeclampsia and diabetes are two to three times more common in twin gestations than in singletons.16,17 Despite these risks, twins generally result in a successful pregnancy for both the parents and health care providers.

Twin gestations accounted for 3.2% of all live births in the United States in 2003.4 Multiple gestations have increased 1.33-fold in the United States over the past decade from 75,858 of 3,612,258 live births (1/48) in 1980 to 128,665 of 4,089,950 live births (1/31) in 2003. When considered as the proportion of pregnancies rather than live births (Fig. 8-1), twins comprised 1/94 pregnancies in 1980 and 1/62 in 2003. Figure 8-2 tracks the changes in the number of twin, triplet, and quadruplet live births in the United States from 1990 to 2003. All three categories of multiples increased from 1990 to 1998. From 1998 to 2003, however, the number of quadruplets declined, triplets leveled off, and the rate of increase in twins has slowed. Most multiple pregnancies are twins. They comprised 96.9% of all multiples in 1990, but this percentage decreased steadily to 95.7% in 2000 before increasing to 96% in 2001.


Twin embryology differs from that of singletons in that there are two embryos with one or two placentas. Twins are either monozygotic or dizygotic depending on the number of eggs fertilized at conception (Fig. 8-3).2 Monozygotic twins, often called “identical” by nonmedical personnel, result from one egg being fertilized by one sperm that divides into two embryos within two to 14 days after fertilization. These twins have the same genetic makeup and are contemporaneous “clones” of one another. The rate of monozygotic twinning is constant throughout the world at approximately 1/250 pregnancies.2 The timing of cleavage in monozygotic twinning determines the type of placentation and the likelihood of complications. Figure 8-4 is based on the classic report by Benirshke and Kim.2 It depicts the first 17 days after ovulation in a monozygotic twin gestation and demonstrates three types of placentation: dichorionic-diamniotic (about 1/3, occurring on day 0–3), monochorionic-diamniotic (about 2/3, occurring on day 4–8), and monochorionic-monoamniotic (< 1%, occurring on day 9–12) that may result from this type of twinning (Table 8-1). Although the incidence of monozygotic twinning has been relatively constant throughout the world in spontaneous twin gestations (Table 8-2), it may be increased by assisted reproductive technologies, particularly those methods that alter the zona pellucida around the time of fertilization, or where there is delayed blastocyst implantation.18,19 Dizygotic twinning varies by race, geographic area, maternal age, and the availability of assisted reproductive technologies.20,21 In the 1980s, the highest rate of dizygotic twinning was seen in an area of Nigeria, where it was 49/1000 live births. The lowest rate was found in Japan at 1.3/1000 live births (see Table 8-2).22 This increased rate in Nigerian women is due to a higher baseline level of follicle-stimulating hormone.20 The maternal age peak for spontaneous multiples is 37 years of age.21

Table 8-2 Geographic Twinning Rates per 1000 Births

Geographic Area Monozygotic Dizygotic
Nigeria 5.0 49.0
United States    
Black 4.7 11.1
White 4.2 7.1
England and Wales 3.5 8.8
Japan 3.0 1.3

Adapted from MacGillivray I: Epidemiology of twin pregnancy. Semin Perinatol 10:4, 1986.

With the greater access to assisted reproductive technologies, the geographic differences in dizygotic twinning have blurred in developed countries. To assess the impact of assisted reproductive technologies on multiple births, Bardis et al23 reported on a survey of all deliveries for 1 week of 2003 in the United Kingdom. Of the 6913 deliveries, there were 100 twins (1.4%) and one triplet gestation (0.01%). Assisted reproductive technology was used to achieve 1.9% of all pregnancies delivered. The multiple pregnancy rate in the assisted reproductive technology group was 13.5% compared with 1.2% in spontaneous conceptions. Because of the greater morbidity and mortality seen with multiple gestations, fertility clinics have reduced the number of embryos transferred for each cycle. This has decreased the rate of increase in the number of high order multiples. Figures 8-1 and 8-2 show a flattening of the curve for triplets in the United States from 1998 to 2003.4 Despite the lower number of embryos transferred in each cycle, the likelihood of a successful pregnancy has not decreased. Single embryo transfer is becoming more generally accepted. Crintini et al24 reported on 107 infertility patients, of whom 41 had single embryo transfers and 66 had two blastocysts transferred. There was no difference in the implantation rate in 238 cycles in which one or two embryos were transferred (76% versus 66%, respectively) or in the pregnancy rate (76% versus 79%). Twinning rates, however, (3% versus 62%, respectively) differed dramatically.

Dizygotic twins, often called “fraternal” by nonmedical personnel, result from the fertilization of two eggs by two sperm (see Figs. 8-3 and 8-5). The genetic makeup of these twins differs from one another. Dizygotic twinning varies by race, advancing maternal age, geography, season, ethnic origin, and increasing parity.21,25 The most dramatic increases have been seen with the use of assisted reproductive technologies.19 With very rare exceptions, dizygotic twins always have a dichorionic-diamniotic placenta, but monozygotic twins may have either a dichorionic-diamniotic or monochorionic-diamniotic placenta depending on the timing of the twinning event (see Fig. 8-4).26 The type of placentation is the most important predictor of the pregnancy-related complications in twin gestations. Chorionicity, not zygosity, determines the likelihood of many twin-related complications.27,28 A primary goal when performing an ultrasound in a patient with a multiple pregnancy, is to determine chorionicity.


In 1986, Landy et al33 reported on 1000 first trimester pregnancies with an incidence of twinning of just over 3%. They found approximately 20% of these twin gestations resulted in singleton live births, with or without vaginal bleeding. They called this the “vanishing twin” phenomenon. In a prospective series of 68 twin pregnancies diagnosed with two fetal heartbeats, dichorionic and monochorionic twins diagnosed earlier than 8 weeks were significantly more likely to result in a singleton pregnancy than those diagnosed after 8 weeks.34 Dichorionic twins were more likely to result in one or two viable newborns when compared with similarly aged monochorionic-diamniotic twins (Fig. 8-6).

The prognosis for singletons resulting from a vanishing twin depends on the number of sacs seen initially and the timing of the loss of the co-twin. Dickey et al35 found that 15% of singleton in vitro fertilization (IVF) births began as a higher order gestation. Schieve et al36 noted that the greater the number of fetal hearts seen on an early twin ultrasound examination the greater the subsequent risk of low birth weight for both singletons (such as “vanishing twins”) and twins. Pinborg et al37 reported that of all IVF singletons born, 10.4% resulted from a twin gestation early in pregnancy. In their 624 IVF singletons, who were survivors of a vanishing co-twin identified by sonography in week 8, there was a higher incidence of preterm delivery and infant mortality when compared with 5237 singletons. These complications, however, were gestational age dependent. When they stratified the neonatal outcomes by the time of vanishing as early (< 8 weeks), intermediate (> 8 and < 22 weeks), and late (> 22 weeks), delivery at less than 32 weeks was 1.9%, 5.3%, and 21.4% respectively, and a neonatal intensive care unit stay of greater than 28 days was 8.7%, 15.7%, and 43.8% for early, intermediate, and late losses of a co-twin respectively. Neurodevelopmental disorders were also higher with a loss at a later gestational age with 3.3% early, 8.0% intermediate and 9.7% in the late “vanishing” group. In general, if a twin “vanished” at less than 8 weeks, the pregnancy outcomes were comparable with singletons from early singleton gestations. This highlights the importance of careful and repeated ultrasound examinations in the early stages of pregnancy to correctly diagnose and follow twins. We discuss the “appearing twin” later in this chapter.


Perinatal losses are greater in twin gestations than in singletons. Infant mortality rates in 1999 were more than five times higher at 32.9/1000 for live-born twins compared with 6.2/1000 for live-born singletons in the United States.1 Survival in twins depends on chorionicity, which determines the level of risk for anomalies, growth problems, and prematurity.32 Sebire et al32 reported that cumulative loss rate from 12 weeks to term was approximately 3% for dichorionic and almost 15% for monochorionic twins (Fig. 8-7). The excess losses for monochorionic twins relate to TTTS, prematurity, congenital abnormalities and monoamniotic-monochorionic placentation. Because these losses are more likely to occur between 16 and 22 weeks (see Fig. 8-7), we suggest that ultrasound examinations be performed every 1 to 2 weeks during that time period for known monochorionic-diamnoitic twin gestations to screen for TTTS and, if necessary, to develop a management plan. Figure 8-8 shows an ultrasound of a 21-week twin gestation with a fetal demise of one twin. The head circumference and abdominal circumference of the twin with the demise are much smaller than those of the living twin. Figure 8-9 is a photograph of a fetus papyraceus, or mummified remains of the lost twin found in the folds of the membranes of the placenta after delivery of the surviving twin at 38 weeks.

Although monochorionic twins are at greater risk of complications than dichorionic twins, they represent only 20% of all twins (see Fig. 8-5). Most perinatal losses in twins are due to problems with prematurity, anomalies, or growth. Figure 8-10 compares the gestational age at delivery of the 121,346 twin versus 3,902,691 singleton live-born infants in the United States in 2002.4 At 28, 32, and 35 weeks’ gestation, 0.8%, 2%, and 6.3% of singletons had delivered compared with 5.8%, 15.5 %, and 41.9% of twins. This demonstrates a 6.7- to 8-fold higher risk of a preterm delivery at every week from 22 to 35 weeks for twins. Maternal and fetal complications including preeclampsia, gestational diabetes, and cerebral palsy are also increased.11 Cerebral palsy rates increase with higher order multiples. They are 2.3/1000 in singleton live births, 12.6/1000 in twins, and 44.8/1000 in triplet gestations.38


First Trimester

First trimester evaluation is in the best time to determine chorionicity and amnionicity in multiple gestations. The diagnosis becomes more difficult as pregnancy progresses. First trimester diagnosis is based on the number of gestational sacs, amnions, and yolk sacs (Table 8-3 and Fig. 8-11). Transvaginal sonography is often used in the first trimester because these structures may be difficult to visualize transabdominally. Sonographers do almost as well as perinatologists in determining chorionicity. Weisz et al39 compared the accuracy of sonographers and perinatologists in determining chorionicity in 172 twin cases before 14 weeks. The overall rate of agreement on the diagnosis of chorionicity was 90.1%. With dichorionic twins, they agreed on 118 of 119 (99.2%), whereas in monochorionic-diamniotic twins, there was agreement in 94.5%. Sonographers diagnosed 10 monochorionic-monoamniotic twins that were later diagnosed as monochorionic-diamniotic. This is a difficult diagnosis, however, that is often overcalled by both perinatologists and sonographers on the first or second scan. Because of this, we sometimes bring a woman back for two or three scans before diagnosing monochorionic-monoamniotic twins. Monteagudo40 showed the use of first trimester vaginal sonography in correctly determining chorionicity and amnionicity in 43 twin pregnancies. The number of gestational sacs correlates with amnionicity; however, as discussed later, underestimation is possible.

Another diagnostic pitfall in the first trimester is the “appearing” twin. Doubilet and Benson41 have reported their experience with transvaginal ultrasound at 5.0 to 5.9 weeks’ gestation and concluded that 30/220 (13.6%) twin pregnancies were diagnosed as singletons on initial evaluation. Monochorionic twins were more likely to be “appearing” twins. Of those initially diagnosed as singletons, 24/213 (11.3%) were dichorionic twin gestations, whereas 6/7 (85.7%) were monochorionic gestations. They reported that the pregnancy outcomes for the “appearing” twin gestations were comparable with other twin gestations.

The timing of the visualization of membranes and yolk sac depends on chorionicity and gestational age (see Fig. 8-4). Bromley and Benacerraf42 retrospectively reviewed the sonographic images of monochorionic twin pregnancies scanned between 6 and 9.5 weeks. This study included 20 diamniotic and two monoamniotic pregnancies. In diamniotic pregnancies less than 8 weeks, only yolk sacs and no dividing membranes were visualized. Two yolk sacs were noted in all but one case on scanning at a later gestational age (Fig. 8-12). In the monoamniotic cases, a single yolk sac was noted at 9 weeks with a single amnion around both embryos. A similar study correlating yolk sacs with amnionicity was performed by Shen et al43 in 20 monochorionic-diamniotic pregnancies under 11 weeks’ gestation. In 85% (17/20) of monochorionic diamniotic pregnancies, two yolk sacs were noted. In three, only one yolk sac was noted. To summarize, seeing two yolk sacs is confirmation of diamniotic twins; however, a single yolk sac requires further confirmation before the diagnosis of monoamniotic twins should be made. The only criteria diagnostic of monochorionic-monamniotic twins is the finding of a single amniotic cavity (see Table 8-3).

Second and Third Trimesters

The steps in determining placentation in the second and third trimesters include documentation of gender, placental masses, and characterization of the dividing membrane(s). If differing genders or two distinct placentas are noted, then dichorionic placentation has occurred.2 If the same gender is present and there is one placenta, the type of placentation may be dichorionic-diamniotic, monochorionic-diamniotic, monochorionic-monoamniotic, or monochorionic with conjoined twins.44 The intertwin dividing membrane ultrasound characteristics must then be evaluated to determine placentation.45,46

Membrane Insertion, “Twin-Peak” Sign

Perhaps the most useful ultrasonographic marker for determining chorionicity is the appearance of the membrane at its insertion onto the chorionic plate or fetal surface of the placenta. Finberg46 first reported the ultrasonographic appearance of the dividing membrane as it inserts onto the chorionic plate of the placenta. When the groove between the membranes at the insertion into the placenta appears thick, it is called the lambda or “twin-peak” sign, and a fused dichorionic-diamniotic placentation is likely (Figs. 8-13 and 8-14). If the insertion cleanly joins the chorionic plate of the placenta as a thin, wispy membrane, it is called “T” sign, and a monochorionic-diamniotic placentation is likely (Fig. 8-15). Wood et al47 reported the diagnostic utility of the twin-peak or lambda sign to determine chorionicity in a prospective series, with chorionicity confirmed by pathology postdelivery. The twin-peak sign correctly predicted 34/36 dichorionic pregnancies, whereas its absence correctly predicted seven of eight monochorionic pregnancies. This yielded an overall positive-predictive value of 97%. Sepulveda et al48 studied the impact of advancing gestational age on the twin-peak sign, comparing its sonographic presence at a 10- to 14-week scan with that at 16 and 20 weeks. Of the 101 twin pregnancies with a twin-peak sign at 10 to 14 weeks, 98 still had it at the 16-week scan. At the 20-week scan, however, only 87/101 pregnancies had the twin-peak sign. Therefore, advancing gestational age negatively affects the diagnosis of the twin-peak sign. Sepulveda also used a variation of the twin-peak sign to assess chorionicity retrospectively and prospectively in triplet pregnancies. He called the junction of the three interfetal membranes the “ipsilon zone” (Fig. 8-16).49 In 19/20 triplet pregnancies prospectively studied, chorionicity was correctly diagnosed in the first trimester. There were 16 trichorionic pregnancies and four dichorionic pregnancies in this study.

The correct diagnosis of chorionicity and amnioticity in twin pregnancies is often straightforward early in gestation, when the division between the two gestational sacs are thick in dichorionic gestations (see Fig. 8-11B). If one is attempting to determine the chorionicity of twins in the second or third trimester and the diagnosis is equivocal, an attempt should be made to obtain and evaluate, early, first trimester scans if they were done previously, when the diagnosis can be more certain.

Membrane Thickness

A thicker dividing membrane is more likely to be seen in a dichorionic-diamniotic placenta. (see Figs. 8-13, 8-14, and 8-15). In a retrospective study of twin pregnancies, Hertzberg et al50 determined that a thick dividing membrane (greater than 1 mm) indicated dichorionic-diamniotic twins in 38 of 42 cases. Townsend, Simpson, and Filly retrospectively reviewed 75 twin pregnancies to determine chorionicity based on membrane thickness. They found that a thick membrane predicted dichorionic-diamniotic placentation 83% of the time and was seen 89% of the time at the first sonogram.51 A thin membrane predicted 83% of monochorionic placentations but was seen only 54% of the time. A similar study from Winn et al52 used a membrane thickness of 2 mm or more as a cutoff. This accurately predicted monochoronic and dichorionic twins 82% and 95% of the time. Stagiannis et al53 studied the reproducibility of membrane thickness to predict chorionicity. In their study, 27 twin pregnancies were scanned by two observers blinded to chorionicity on 52 occasions at five areas of the membranes. Membrane thickness measurements closest to the placenta were the most reproducible, but significant variation in membrane thickness does occur, particularly in the second and third trimesters.

Membrane Layers

D’Alton et al45 prospectively studied 69 consecutive twin pregnancies to assess chorionicity by counting the number of layers in the dividing membrane visualized by ultrasound. If two layers were seen, a monochorionic pregnancy was diagnosed. If more than two were seen, a dichorionic pregnancy was predicted. Confirmation was obtained after delivery by histopathology (see Fig. 8-13). Ultrasound counting of the membrane layers correctly predicted chorionicity in 68/69 twin pregnancies. There was 100% accuracy in predicting the 51 dichorionic pregnancies and 94.4% accuracy in predicting the 18 monochorionic pregnancies. Vayssiere et al54 performed a similar prospective study counting membrane layers and correctly predicted chorionicity in 60/63 twin pregnancies.

Multiple Sonographic Markers to Determine Chorionicity/Amnionicity

No single sonographic marker of chorionicity is completely reliable. The use of multiple markers enhances the accuracy of predicting chorionicity. Scardo et al55 prospectively studied the accuracy of multiple sonographic markers to predict chorionicity and amnionicity in 100 consecutive twin pregnancies with postdelivery confirmation. They assessed placental number, fetal gender, membrane thickness, and presence of the twin-peak sign (see Figs. 8-14 and 8-15). They performed an average of 3.6 ultrasound examinations, starting at a mean gestational age of 22.6 weeks. Use of the composite sonographic markers had at least a 91% sensitivity and specificity for correctly determining chorionicity. The timing of these examinations was relatively late in gestation. With the greater acceptance of first trimester ultrasound, the accuracy of determining chorionicity should improve. Carroll et al56 performed a similar study to predict chorionicity at 10 to 14 weeks’ gestation using the number of placental sites, presence of a twin-peak sign, and intertwin membrane thickness. This study included 150 pregnancies with postnatal confirmation of chorionicity and included 116 dichorionic and 34 monochorionic pregnancies. Prenatal ultrasound correctly predicted chorionicity in 139/150 pregnancies (93.3%). The most accurate predictors of dichorionic placentation were the twin-peak sign or separate placentas with a sensitivity and specificity of 97.4% and 100%, respectively.

Monoamniotic Twins

When the dividing membrane is not seen early in the ultrasound evaluation of a twin pregnancy, the diagnosis of monochorionic-monoamniotic twins should be considered. This occurs in approximately 1% of all monochorionic twins (see Figs. 8-3, 8-4, and 8-5). We require several examinations to search for the dividing membrane before making this diagnosis, because the membrane may not initially be apparent by ultrasound. Distinguishing monochorionic-monoamniotic twins from a “stuck” twin in an oligohydramniotic sac can be difficult. Normal amniotic fluid volume and two free-floating twins with no visualized membrane separating them should clinch the diagnosis of monochorionic-monoamniotic twins. Visualization of two cord insertions into the chorionic plate of the placenta in very close proximity to one another is also suggestive of monochorionic-monoamniotic twins (Fig. 8-17).

Color and two-dimensional ultrasound readily show the umbilical cord entanglement that leads to the strikingly increased mortality for monochorionic-monoamniotic twins (Fig. 8-18).2,5759 Nyberg et al60 first reported ultrasound evidence of cord entanglement as a sign of monochorionic-monoamniotic twins. Townsend and Filly confirmed this finding in five sets of nonconjoined, monochorionic-monoamniotic twin pregnancies (Fig. 8-19).61 Subsequent studies have supported this finding, which is characteristic of monochorionic-monoamniotic twin pregnancies. Overton et al62 reported using color flow and pulsed wave Doppler to insonate a common cord mass to confirm monochorionic-monoamniotic twins in the first trimester (Fig. 8-20). First trimester vaginal sonography using color Doppler has also demonstrated cord entanglement.63 Sonography may also be useful in the antenatal management of monochorionic-monoamniotic twins. Belfort et al64 reported extremely high blood flow velocities associated with umbilical vein compression by the entangled cords when following three monochorionic-monoamniotic twin pregnancies (Fig. 8-21). Abuhamad et al65 followed two monochorionic-monoamniotic twin pregnancies with longitudinal Doppler flow velocities of the umbilical artery. Notching of the umbilical artery waveform was associated with increased narrowing of the umbilical vessels secondary to cord entanglement. Doppler interrogation of the aggregate cord mass may also demonstrate arterial flow, at differing heart rates in both fetuses, which is called a “galloping” Doppler pattern (Fig. 8-22). As with all multiple gestations, a velamentous insertion of the umbilical cord (that is, where the cord inserts into the membranes rather than the chorionic plate) is more common in monochorionic-monoamniotic placentas (Fig. 8-23).

Several different protocols have been proposed for the management of monochorionic-monoamniotic twins. Because of the rarity of this condition, there are no randomized clinical trials. Rodis et al57

Only gold members can continue reading. Log In or Register to continue

Premium Wordpress Themes by UFO Themes