Fig. 4.1
Proportion of dichorionic twins (upper curve) and monochorionic twins (lower curve) among all pregnancies in Slovenia between 1987 and 2012 (Adapted with permission from Tul [134])
Despite this increase in incidence, twin pregnancies still represent a relatively small proportion of all deliveries. However, they make a disproportionate contribution to overall perinatal mortality and morbidity. Ten to 15 % of perinatal deaths and around 15 % of cerebral palsy cases are associated with twin gestations [7–9]. This association is mostly due to an overrepresentation of twin births among very preterm births (<32 weeks of gestation). In fact, the proportion of twins among all very preterm deliveries has risen in the last decades and now approaches 35 % (Fig. 4.2) [4]. In addition, an increased risk for cerebral palsy has also been reported in twins at >37 weeks’ gestation [9].
Fig. 4.2
Proportion of twins among all infants born at <32 weeks of gestation in Slovenia between 1987 and 2010 (From Tul et al. [4])
Twin pregnancies are, therefore, a very important clinical entity in perinatal medicine, which requires vigilant follow-up, and timely recognition plus management of several possible complications. This chapter reviews most common late pregnancy complications associated with twins, as well as current recommendations on their prevention and treatment.
4.2 Types of Twin Gestations
Twin pregnancies occur when two oocytes are fertilized to form dizygotic (nonidentical or fraternal) twins or when a single fertilized ovum divides to form monozygotic twins (often referred to as identical, although genetic as well as phenotypic differences always exist) (Fig. 4.3) [10]. In dizygotic twin pregnancies, each fetus has its own amnion and chorion. There are, on the other hand, several variations of monozygotic twins. According to the most popular (but unproven) theory, these depend on the timing of the division of monozygotic conceptus (Fig. 4.4):
Diamniotic/dichorionic: If the division of the conceptus occurs within 3 days from fertilization, each fetus will be surrounded by his own chorion and amnion (Fig. 4.5).
Diamniotic/monochorionic: If division occurs between fourth and eighth day following fertilization, the chorion has already began to develop, whereas the amnion has not. Consequently, each fetus will be surrounded by its own amnion, but a single chorion will surround both twins (Fig. 4.5).
Monoamniotic/monochorionic: In less than 1 % of all monozygotic gestations, division occurs between days 9 and 12, after development of both amnion and chorion. Fetuses will, therefore, share a common sac. Division after 12 days is incomplete and results in conjoined twins. The fetuses may be fused in a number of ways, with the most common involving the chest and/or abdomen (Fig. 4.6). This is a rare condition with an incidence of 1 in 70,000.
Fig. 4.3
Incidences of various types of twins. The more twins share (genome, chorion, amnion), the greater the risk of perinatal mortality (PM)
Fig. 4.5
(a) Monochorionic triamniotic placenta. Each fetus was surrounded by its own amnion, but a single chorion surrounded all three fetuses. Notice the thin amniotic membranes. (b) Trichorionic triamniotic placenta. Each fetus was surrounded by its own amnion and chorion. Notice the thicker membranes. (c) Dichorionic triamniotic placenta. Difference in thickness of membranes can be seen thin monochorionic membrane on the left and thick dichorionic on the right
Fig. 4.6
Conjoined twins – case of cranio-thoracopagus. Fetuses may be fused in a number of ways, with the most common involving the chest and/or abdomen
Monochorionic twins are associated with significantly higher perinatal risk compared with dichorionic ones, irrespective of zygosity [11]. In one large twin cohort study, the perinatal mortality rate was found to be more than twofold increased in monochorionic compared with dichorionic twins [12]. As a result, establishing chorionicity has become a keystone in the management of twins.
Chorionicity is most reliably ascertained sonographically early in gestation. Before 10 weeks’ gestation, evidence of two distinct gestational sacs on transvaginal ultrasound suggests dichorionicity. Determination of amnionicity is less accurate at such early gestation due to the thin amniotic membrane and should be established later to exclude the possibility of monoamniotic twins. At 10–14 weeks, visualization of the so-called lambda sign (also known as the twin peak sign), which is the triangular projection of placental tissue into the base of the intertwin membrane, is an important marker of dichorionic placentation (Fig. 4.7) [13]. The presence of either a lambda sign or two separate placentas at less than 14 weeks’ gestation indicates dichorionic placentation with a sensitivity of 97 % and a specificity of 100 % [13]. Conversely, a T sign has been used to describe the ultrasonographic visualization of the attachment of the intertwin membrane to the placenta in cases of monochorionic gestation (Figs. 4.8 and 4.9). The presence of either a single placental mass or a T sign at less than 14 weeks’ gestation has a sensitivity of 100 % and a specificity of 98 % for monochorionic placentation [13].
Fig. 4.7
Ultrasonographic picture of dichorionic twins at 12 weeks. On the left, 2D picture with visible thick intertwin membrane with triangular projection of placental tissue into the base of the intertwin membrane (lambda sign or twin peak sign). On the right same dichorionic twins in 3D
Fig. 4.8
Ultrasonographic picture of monochorionic twins at 12 weeks. On the left, 2D picture with visible thin intertwin membrane and no bulging of placental tissue into the base of the membrane (T sign). On the right, same monochorionic twins in 3D
Fig. 4.9
Dichorionic triamniotic pregnancy at 11 weeks. Between fetuses B and C, the membrane is thin and T sign can be seen (monochorionic pair); between the fetus A and the other two, the membrane is thicker and lambda sign can be seen
Determination of chorionicity in later gestation is much more difficult and less reliable. The lambda sign tends to disappear with advancing gestational age due to regression of the chorion frondosum to form the chorion laeve [13]. Discordance of fetal gender indicates dizygocity and, therefore, dichorionicity. However, only 55 % of all twins have been reported to be of different sex making gender discordance an unreliable sign of dichorionicity [14]. Thickness of the intertwin membrane may be another helpful indicator in determining chorionicity (Fig. 4.5). Membrane thickness of less than 2 mm has been reported to have 90 % sensitivity and 76 % specificity for diagnosing monochorionic/diamniotic twins using standard 2D sonography, and sensitivity can be further improved using 3D sonography [15]. Visualization of two separate placental masses can also be used to confirm dichorionicity. However, this finding is usually present in only about one-third of twin gestations. Moreover, both the presence of a thin bridge of placental tissue between two dominant placental masses and the presence of a succenturiate placental lobe can be seen in a monochorionic gestation, thereby limiting this parameter as a useful diagnostic tool.
As establishing chorionicity becomes increasingly difficult and less reliable as gestation progresses, all women with a twin pregnancy should be offered ultrasound examination in the first trimester to assess chorionicity in addition to viability, crown-rump length, and nuchal translucency [16]. Early establishment of chorionicity allows risk assessment which greatly determines further surveillance strategies in twin pregnancies.
4.3 Preterm Birth
Preterm birth is one of the main complications associated with twin pregnancies and the most important risk factor for perinatal morbidity and mortality in these gestations [17]. Approximately 50–60 % of twins are born preterm (before completed 37 weeks’ gestation), and 10 % are born very preterm (before completed 32 weeks’ gestation) [4, 18] (Fig. 4.10). Major complications associated with prematurity include respiratory distress syndrome, necrotizing enterocolitis, intraventricular hemorrhage, and sepsis. These complications drive perinatal mortality rate of twins to seven times that of singletons and cerebral palsy rate to between four and seven times that of singletons [9, 17].
Fig. 4.10
Number of twin births in regard to gestational age in Slovenia between 2002 and 2010. Most births occurred between 34 and 39 weeks (From Bricelj et al. [18])
Prevention of preterm delivery is one of the most important goals in management of twin pregnancies. There is now substantial evidence that measuring cervical length by transvaginal ultrasound in singleton pregnancies helps to detect patients at increased risk of spontaneous preterm birth [19, 20]. This has been demonstrated to be true in twin pregnancies as well [21, 22]. Souka et al. examined the accuracy of cervical length at 23 weeks to predict spontaneous preterm delivery at less than 32 weeks in twins. Cervical length of less than 25 mm was associated with increased risk for very preterm birth. In addition, 70–80 % of women with cervical length of less than 10 mm delivered at less than 32 weeks [22] (Fig. 4.11).
Fig. 4.11
Transvaginal ultrasound image of uterine cervix. The upper picture shows a long and closed cervix associated with low risk of preterm birth, while the lower picture shows a shortened cervix with “funneling” associated with increased risk of preterm delivery
Progesterone has been known to be important in maintaining pregnancy for more than 80 years, since the classic work of Corner, Allen, and Csapo [23, 24]. A large body of experimental data demonstrates that progesterone exerts overall control on both cervical ripening and myometrial contractility [25]. In addition to basic science support for the use of progestins in pregnancy, there is also substantial empirical evidence of their potential benefit from large clinical trials. Use of progestin supplementation (either as natural (bioidentical, micronized) progesterone administered vaginally or as 17 alpha-hydroxyprogesterone acetate (17-OH P) administered intramuscularly) has been shown to reduce the risk of preterm delivery in two subgroups of pregnant women: women with history of previous spontaneous preterm birth (trials shown benefit from 17-OH P, but not progesterone) and women with a short cervix measured by transvaginal ultrasound at 19–24 weeks’ gestation (trials shown benefit from vaginal progesterone, but not 17-OH P) [26–30]. Rouse et al. showed that treatment with 17-OH P did not reduce the rate of preterm births when administered to all women pregnant with twins, which is in agreement with the results of an earlier study published in 1980 [31, 32]. Two large studies in twin gestation, randomized to receive either vaginal progesterone or placebo, also showed no benefit of vaginal progesterone treatment in twin pregnancies [33, 34]. Increased doses of vaginal progesterone also did not reduce preterm birth rate in twin gestations [35]. There is no single randomized controlled trial looking specifically at effectiveness of vaginal progesterone in women pregnant with twins who are found to have a shortened cervix. However, Romero et al. published an individual patient data meta-analysis of five randomized controlled trials comparing vaginal progesterone treatment to placebo in patients with a sonographic short cervix (≤25 mm), which also included twin gestations [36]. They found no statistical sign of different responses to progesterone among twins compared to singleton pregnancies. One could, therefore, assume that if progesterone reduces the risk of preterm birth in singletons, it should do the same in twins. In fact, there was a 30 % reduction associated with vaginal progesterone treatment in preterm birth rates at less than 33 weeks in twin gestations included in this meta-analysis, but this reduction was not statistically significant. Importantly, however, they also found an almost 50 % statistically significant reduction in composite morbidity in twins whose mothers were treated with progesterone. Based on these data, we recommend treatment of patients pregnant with twins with a cervical length of ≤25 mm with vaginal progesterone.
Similarly to progesterone, cerclage placement in all twin pregnancies does not reduce the risk of spontaneous preterm birth [37]. However, data on efficacy and even safety of ultrasound-indicated cerclage, i.e., asymptomatic twin pregnancies with a cervical length of less than 25 mm, are conflicting. A Cochrane meta-analysis of randomized controlled trials of cervical cerclage in multiple pregnancies showed that cerclage placement in twin pregnancies with a cervical length of less than 25 mm could increase the risk of preterm delivery [36]. On the other hand, a recent multicenter retrospective cohort study showed that ultrasound indicated cerclage was not associated with significant effects on perinatal outcomes compared to controls when the 25 mm cervical length was used. In addition, in women with a cervical length of ≤15 mm before 24 weeks, cerclage was associated with a significant prolongation of pregnancy by almost 4 more weeks [38]. There are even less data on efficacy of emergency or physical exam-indicated cerclage in twins, i.e., cerclage placed in a patient with a dilated cervix on examination or membranes visible at the external cervical os on speculum examination. In the absence of better therapeutic options and in view of extremely increased risk of unfavorable outcomes, many centers will offer a cerclage to such patients prior to viability (24 weeks). A small, single institution retrospective study showed that emergency/physical exam-indicated cerclage in twin pregnancies could be associated with favorable outcomes, including a higher likelihood of delivery at >32 weeks [39]. There is currently an ongoing registered randomized trial of physical exam-indicated cerclage in twin gestations (NCT02490384, https://clinicaltrials.gov/).
Transvaginal placement of a silicon (Arabin) pessary around the cervix has been proposed as an alternative to progesterone and cerclage for preterm birth prevention (Fig. 4.12). It is thought to support the cervix and change its direction toward the sacrum, thereby reducing the direct pressure from the uterine contents on the cervical canal. Studies in singleton pregnancies with short cervix yielded conflicting results regarding its efficacy [40, 41]. Three randomized controlled trials examined the effects of cervical pessary on preterm birth rates in twins. A Dutch trial published in 2013 found no reduction in preterm birth at < 32 weeks in multiple pregnancies overall, but less very preterm births in women with cervical length <38 mm who had the pessary placed [42]. A recently published international randomized multicenter trial did not confirm these findings. Use of Arabin pessary in 1,180 women with twins at 20–24 weeks did not reduce preterm delivery before 34 weeks if inserted to randomly selected women (irrespectively of cervical length), neither did in subgroup of women with cervical length of <25 mm [43]. This is in contrast with the results of a Spanish trial also published in 2016, which found a significant reduction in preterm births at <34 weeks associated with pessary insertion in twin pregnancies with a cervical length of ≤25 mm [44].
Fig. 4.12
Arabin cervical pessary (in the circumference), placed around the uterine cervix in order to prevent preterm birth
4.4 Fetal Anomalies
Congenital anomalies are two to four times more common in twins compared with singletons [44–55]. Most frequent fetal anomalies in twins are cardiovascular defects [41]. These are also the most common congenital anomalies in singletons, but the relative risk is higher for twin gestations [46, 48–54]. Anomalies of the central nervous system, such as hydrocephaly and neural tube defects; of the gastrointestinal system, in particular intestinal atresia; of the genitourinary system; and of the musculoskeletal system were also reported to occur more frequently in twins compared with singletons [46, 47, 49, 50, 54]. In contrast, similar rates of chromosomal abnormalities have been reported in twins and singletons [46, 47, 49–51]. These studies have, however, failed to appreciate the different maternal age distribution between twin and singleton pregnancies [5]. Moreover, although in monozygotic twins the risk of chromosomal abnormalities is the same as maternal age-related risk, in dizygotic pregnancies, each fetus carries an independent risk, and the risk per pregnancy is doubled.
Studies that were able to examine congenital anomalies in twins by chorionicity or zygosity found that most of increased risk for anomalies is attributable to excess risk in monozygotic pregnancies [46, 47, 55, 56]. In dizygotic twins, the prevalence of structural anomalies in each twin is the same as in singleton pregnancies. Consequently, chances of at least one of dizygotic twins to have a structural anomaly are approximately twice as high as in singletons, with the addition of higher rates of compression or crowding anomalies, such as club feet [46, 47]. In monozygotic twins, the risk of congenital anomalies, such as brain, cardiac, renal, intestinal, and other abnormalities, is up to four times higher for each twin than in singletons [45–49, 55, 56]. A number of mechanisms have been proposed for the higher rate of anomalies observed in monochorionic and monozygotic twins. Monozygotic twinning itself can be regarded as an abnormality of morphogenesis since it involves zygotic splitting. This process has been associated with some specific malformations that have a predilection for midline structures, e.g., sirenomelia, cloacal anomalies, and holoprosencephaly, that are much more common in monozygotic twins [57, 58].
Higher rates of fetal congenital anomalies warrant a first trimester ultrasound examination with nuchal translucency thickness measurement and early anatomy scan in all twin pregnancies. At this time, chorionicity must be clearly diagnosed in all cases. In doubts, the patient must be sent to a referring center. We recommend a follow-up for all monochorionic twins in referring centers. Regardless of chorionicity, a detailed anatomical assessment should then be performed at 18–24 weeks’ gestation. In a single-center series of 245 consecutive twin gestations, 21 of 24 anomalous fetuses were detected by detailed ultrasound examination at 18–20 weeks (88 % sensitivity, 100 % specificity, 100 % positive predictive value, negative predictive value 99 %) [59].
In addition to structural fetal anomalies, there are also other anomalies specifically related to monochorionicity that result from vascular anastomoses between the circulations supplying monochorionic twins; two most common are twin-to-twin transfusion syndrome (TTTS) and twin anemia-polycythemia sequence (TAPS).
4.4.1 Twin-to-Twin Transfusion Syndrome (TTTS)
Monochorionic embryos/fetuses are connected by transplacental vascular anastomoses that lead inevitably to sharing of the circulation. There are three main types of anastomoses in such placentas: venovenous (VV), arterioarterial (AA), and arteriovenous (AV). Both AA and VV anastomoses are direct superficial connections on the surface of the placenta with the potential for bidirectional flow (Fig. 4.13). In most cases, the shared circulation between monochorionic twins is relatively balanced, but in as many as 8–10 % of cases, an unbalanced intertwin shunt may be created through AV anastomoses, leading to so-called TTTS [16]. TTTS can theoretically occur at any time during pregnancy, but usually presents in the second trimester (Fig. 4.14). It is diagnosed by polyuric polyhydramnios in the recipient twin, with a maximal vertical pocket of amniotic fluid measuring at least 8 cm, and oliguric oligohydramnios in the donor twin with a maximum vertical pocket of 2 cm or less (Fig. 4.15) [16]. The most commonly used TTTS staging system was developed by Quintero et al. in 1999 [60]. The TTTS Quintero staging system includes five stages, ranging from mild disease with isolated discordant amniotic fluid volume to severe disease with demise of one or both twins. The criteria for stage I are polyhydramnios in the recipient twin and oligohydramnios in the donor twin with a visible bladder in the donor twin. In stage II the bladder of the donor twin is not visible at any point during the evaluation, but Doppler studies are not abnormal. In stage III Doppler studies are abnormal in either twins and are characterized by absent or reversed diastolic flow in the umbilical artery, reversed flow in the ductus venosus or pulsatile umbilical venous flow. In stage IV ascites, pericardial or pleural effusion, scalp edema or overt hydrops, are present (Fig. 4.16). Stage V is characterized by the demise of one or both twins (Fig. 4.17). This system has some prognostic significance and provides a method to compare outcome data using different therapeutic interventions. Although the stages do not correlate perfectly with perinatal survival, it is relatively straightforward to apply, may improve communication between patients and providers, and identifies the subset of cases most likely to benefit from treatment. In severe cases of TTTS, perinatal mortality without treatment ranges between 70 and 100 % (Fig. 4.17). Most experts currently consider fetoscopic laser photocoagulation of placental anastomoses to be the best available approach to treat severe TTTS at less than 26 weeks (Fig. 4.18). This procedure has been associated with an overall perinatal survival of 50–70 % in those with severe disease. Other therapeutic options may be considered in selected cases and include expectant management, amnioreduction, intentional septostomy of the intervening membrane, and selective termination [16].
Fig. 4.13
Placenta of monochorionic twins after injection of umbilical blood vessels with dye. Venovenous anastomoses are visible (arrows)
Fig. 4.14
Reversed a wave in the ductus venosus. An early sign of twin-to-twin transfusion syndrome (TTTS) at 14 weeks of gestation
Fig. 4.15
Twin-to-twin transfusion syndrome (TTTS) at 20 weeks of gestation. Polyhydramnios in the recipient twin and oligohydramnios in the donor twin can be diagnosed looking at the position of the amniotic membrane (thick arrow). Notice also the dilated bladder of the recipient twin (thin arrow). In addition, there is intrauterine growth restriction in the donor twin (not part of the Quintero classification system)
Fig. 4.16
Twin-to-twin transfusion syndrome (TTTS) stage IV according to the Quintero classification system. Notice the scalp edema in one of the fetuses
Fig. 4.17
Twin-to-twin transfusion syndrome (TTTS) resulting in demise of both fetuses at 17 weeks. This could be potentially prevented with early recognition and laser therapy
Fig. 4.18
(a, b) Placenta (left image) and placental blood vessels (right image) of monochorionic quadramniotic quadruplets after laser therapy for fetofetal transfusion syndrome and injection of umbilical blood vessels with acrylate monomers. Tiny anastomoses are visible even after laser therapy (arrows) (With permission from Tul et al. [135])
4.4.2 Twin Anemia-Polycythemia Sequence
Another unique anomaly found exclusively in monochorionic gestations is TAPS. It is a form of chronic fetofetal transfusion and complicates up to 6 % of monochorionic diamniotic twin pregnancies, typically in the late second or third trimester [61]. TAPS is defined as the presence of anemia in the donor and polycythemia in the recipient twin, diagnosed antenatally by middle cerebral artery (MCA) peak systolic velocity (PSV) of 1.5 multiples of median in the donor and MCA PSV <1.0 multiples of median in the recipient, in the absence of oligohydramnios-polyhydramnios [61, 62]. Extreme cases of TAPS can progress to fetal death. Suggested treatment options include laser photocoagulation, intrauterine blood transfusion, selective termination, and early delivery, but further studies are required to determine the natural history and optimal management of TAPS.
4.5 Stillbirth
Overall, twins are at an approximately fivefold increased risk of fetal death compared to singletons (Fig. 4.19) [8]. This risk is predominately influenced by the marked increase in rates of fetal demise in monochorionic twins (7.6 % monochorionic versus 1.6 % in dichorionic twins) [12]. Even in apparently uncomplicated monochorionic twins, single fetal deaths may still occur in as many as 2.6 % of pregnancies [63]. Of note, population-based studies yielded much higher incidences of stillbirth in monochorionic-diamniotic twins as compared to hospital-based surveys coming from tertiary centers with special interest in monochorionic twinning [63–68]. This supports the idea that a strict protocol of close surveillance should be implemented for all monochorionic gestations and that elective preterm delivery might be a reasonable policy to avoid unexpected intrauterine fetal deaths in these pregnancies (see Sect. 4.8) [69]. This is the main reason why we recommend follow-up of monochorionic twins in refereeing centers.
Fig. 4.19
(a, b) Following an intrauterine death of one of the twins, there may be complete reabsorption of the death fetus (when intrauterine death occurs very early in gestation) or formation of a fetus papyraceus (i.e., a “mummified” or compressed fetus) (left). Ultrasound images of fetus papyraceus are shown below (Courtesy of Andrea Tinelli)
Intrauterine death of a fetus in a twin pregnancy may be associated with poor outcome for the co-twin, but the degree of risk is again dependent on chorionicity. Vascular intraplacental connections between twins may place the surviving co-twin in a monochorionic pregnancy at a significantly higher risk. In a recent meta-analysis, death of one twin was associated with co-twin demise in 15 % of monochorionic gestations and 3 % of dichorionic gestations [70]. Similarly, the incidence of neurologic morbidity following death of a co-twin was 26 % in monochorionic gestations, compared with 2 % in dichorionic gestations [70]. Previously thought to be related to the passage of thromboplastin-like substances after the death of the twin, the more widely accepted theory today is that acute hypotension in the dying fetus results in a “sink” phenomenon [71]. Acute exsanguination of the normal co-twin results in its death or survival with neurologic sequelae. Thus, immediate or emergent delivery confers no advantage to the surviving fetus after the death of its co-twin. There is a theoretical concern of maternal complications, such as disseminated intravascular coagulation, due to retention of the death fetus when continuing pregnancy after intrauterine demise of one twin. [71–73]. The incidence of this complication, however, seems to be exceedingly low [72–74].
4.6 Intrauterine Growth Restriction
Data from numerous studies, as well as our own from the Slovenia’s National Perinatal Information System, indicate a trend of fetal growth deceleration after the 28th week in twin gestations in comparison with singletons (Fig. 4.20) [18, 75–80]. This may be the result of placental insufficiency when twins approach term [81]. Intrauterine growth restriction (IUGR) of one or both fetuses can be determined by ultrasound, which estimates fetal weight with reasonable accuracy [82]. Numerous sequential measurements must be made as IUGR is defined as inappropriate growth during a certain time period [83]. The appropriateness of fetal growth is assessed by using birth weight by gestational age – growth curves, preferably twin-specific growth curves. Alternatively, IUGR in twins is often defined as estimated fetal weight below the tenth percentile using singleton growth curves or presence of ≥15–25 % discordance in estimated fetal weight between the lighter and heavier twin [75–80].
Fig. 4.20
Comparison between growth curves for twins and singletons. Fiftieth centile and areas between 10th and 90th centile are plotted (Notice the slower growth of twins in the third trimester (From Bricelj et al. [18]))
The prevalence of IUGR has been reported to be 26 % in dichorionic twins and as high as 46 % in monochorionic twins [84]. Monochorionicity increases the overall risk of IUGR in twin pregnancies due to disproportionate placental sharing (Fig. 4.21). In one prospective series, selective IUGR, defined as a birth weight discordance of at least 25 % in the absence of TTTS, was reported to complicate about 15 % of monochorionic pregnancies and was associated with perinatal mortality of 5–10 % (Fig. 4.22) [66].
