Multiple pregnancy: diagnosis

Women pregnant after the use of assisted reproductive techniques (ART), and their doctors, are usually well aware of the increased chance of a multiple pregnancy. Although fertility clinics and ART guidelines increasingly aim to prevent the occurrence of twins and especially higher order multiple pregnancies, the incidence of multiple pregnancy in this group is at least 5 to 10 times higher than that in spontaneous pregnancies . Because the women and the clinics are always eager to confirm an intact pregnancy as soon as possible, the multiple pregnancies in this group are practically always detected early in the first trimester by (transvaginal) ultrasound. At present, there should be sufficient awareness regarding the importance of chorionicity in fertility clinics and the fact that ART increases the risk of monochorionic splitting so that correct determination can be expected. However, in spontaneous multiple pregnancies, the diagnosis is occasionally delayed when only a quick dating scan is performed without systematically examining the whole uterus. Similarly, women with triplets quite often report that at their first ultrasound, only a twin was observed.

Dating of twins

In multiple pregnancies after ART, gestational age can accurately be determined using the dates of the ovum pick-up or insemination. In spontaneous pregnancies, singletons or multiple pregnancies, the standard of care is to determine gestational age by crown–rump length (CRL) in the first trimester . Most recent guidelines state that the CRL should preferably be measured between 11+0 and 13+6 weeks of gestation; however, this is mostly based on the large quantity of data available from this period because this is when the nuchal translucency for trisomy prediction is performed. With the increasing use of more reliable cell-free DNA tests using maternal plasma to predict fetal trisomies, this strict range may become less important, and good quality measurements at 9 or 10 weeks could prove to be equally reliable. After 14 weeks, the fetal head circumference should be used. In all types of multiple pregnancies and in monozygotic pregnancies, the size and growth of the fetus can be different. Unequal placental sharing and discordancy for anomalies occur frequently and can be associated with size differences . Even monozygotic twins can be discordant for karyotype . This means that dating of twins with a discordant CRL is best performed using the size of the largest twin. Clinical problems associated with abnormal size more often result in growth restriction; thus, it seems safe to assume that the larger twin is normal and the smaller twin is too small, instead of the opposite explanation .

Dating of twins

In multiple pregnancies after ART, gestational age can accurately be determined using the dates of the ovum pick-up or insemination. In spontaneous pregnancies, singletons or multiple pregnancies, the standard of care is to determine gestational age by crown–rump length (CRL) in the first trimester . Most recent guidelines state that the CRL should preferably be measured between 11+0 and 13+6 weeks of gestation; however, this is mostly based on the large quantity of data available from this period because this is when the nuchal translucency for trisomy prediction is performed. With the increasing use of more reliable cell-free DNA tests using maternal plasma to predict fetal trisomies, this strict range may become less important, and good quality measurements at 9 or 10 weeks could prove to be equally reliable. After 14 weeks, the fetal head circumference should be used. In all types of multiple pregnancies and in monozygotic pregnancies, the size and growth of the fetus can be different. Unequal placental sharing and discordancy for anomalies occur frequently and can be associated with size differences . Even monozygotic twins can be discordant for karyotype . This means that dating of twins with a discordant CRL is best performed using the size of the largest twin. Clinical problems associated with abnormal size more often result in growth restriction; thus, it seems safe to assume that the larger twin is normal and the smaller twin is too small, instead of the opposite explanation .

Chorionicity

Chorionicity is of vital importance in the management of multiple pregnancies. It should be determined using ultrasound in the first trimester and documented with a picture in the patient file . Incorrect labeling of a twin pregnancy as dichorionic may have grave consequences due to underestimating a number of risks and insufficient frequency of monitoring. In case of any doubt, seeking assistance from a more experienced sonographer or referral to a specialized clinic is strongly advised. Dichorionic twinning is diagnosed by observing a thick membrane separating the fetuses, consisting of two layers of chorionic membrane that appear fused, with two thin layers of amnion on either side. At the site of insertion to the placenta, the chorion tissue is visualized on ultrasound between the two layers of chorionic membrane where they form a triangle or the Greek letter lambda, commonly known as the lambda sign. In monochorionic twinning, the two layers of amnion may, in the first trimester, also separate close to the placenta in a lambda-like form; however, no chorionic tissue is then observed in the triangle. This is sometimes referred to as the empty lambda sign. The so-called T-sign, describing the ultrasound appearance of the two layers of amnion where they meet the placental surface, is a second trimester phenomenon. Because dichorionic placentas can be located adjacent to each other, appearing fused, and monochorionic placentas may appear bilobed, only connected by a few blood vessels that may not be visible, the use of “one or two placentas” to aid in the determination of chorionicity is unreliable . The same holds true for first trimester visualization of the fetal sex.

Identification

To monitor growth as well as in antenatal diagnosis of structural or chromosomal anomalies, measurements or other findings obviously need to be assigned to the correct fetus every time. Because during the course of the pregnancy fetuses may change position repeatedly, systematic labeling by detailed description and, if feasible in the now common electronic files, drawings, is important . In twins, one is usually located more to the right and the other to the left in the uterine cavity, and this should then be noted. The position of the placenta and cord insertion for each twin should also be recorded. This is also helpful when both fetuses are exactly in the middle, one close to the cervix and one near the fundus. Only in the third trimester, labeling by position relative to the cervix is useful, where traditionally the first twin, closest to the cervix, is called A and the other is called B. In the first and second trimester, identification by labeling them right (1 or A) and left (2 or B) is preferable over lower or higher in the uterus, although left–right switch may occur. When both are in the middle, the one with its cord insertion most to the right can be called 1 or A. Of course, in different-sex dichorionic twins, the sex can also be used and recorded for identification at second and third trimester ultrasound. In triplets, the most right-sided fetus is called 1 or A, the most left-sided fetus is 2 or B, and the one in between who is commonly in fundal position is 3 or C. In triplets, adding drawings to the file can be particularly helpful.

Testing for chromosomal anomalies

To estimate the risk for each fetus in multiple pregnancy for trisomy 21, the most common chromosomal anomaly, several specific issues need to be considered. One important risk factor, maternal age, is obviously the same for both fetuses. It should be remembered that for a pregnant woman with a dizygotic twin, the risk of carrying a fetus with a trisomy is doubled. In addition, before embarking on fetal trisomy testing, counseling should include discussing the scenario of detecting trisomy in one twin only, which complicates decision-making.

Markers in maternal serum such as pregnancy-associated plasma protein A and free beta-human chorionic gonadotropin, produced by the placenta or placentas, cannot be distinguished as originating from one or the other fetus. With the current technology, the same holds true for the cell-free DNA fragments that are increasingly used for noninvasive screening for chromosomal anomalies. This could, however, in dizygotic twins, change in the future. Fetus-specific testing can be performed by ultrasound examination (nuchal translucency, markers for trisomy, and structural anomalies) and by invasive sampling of chorionic villi or amniotic fluid . In case of performing chorionic villus sampling in twins, great care has to be taken to sample at placental sites that are sufficiently far apart, preferably near each fetus’ cord insertion, with obviously meticulous labeling. When amniocentesis is indicated, sampling both sacs even in monochorionic twins (who can be discordant for chromosomal anomalies) should be performed using a new needle, syringe, and insertion for the second sample. A single needle entry traversing the membrane to enter the second sac may appear safer; however, septostomy may result in further tearing of the membrane, creating an iatrogenic monoamniotic situation with potential for cord entanglement.

In trisomy screening in otherwise uncomplicated monochorionic twins, cell-free DNA testing from maternal plasma should be as reliable as in singletons, although at present (2016), the evidence from published studies is limited . In dichorionic twins, still limited data suggest that reliability of trisomy 21 detection could be similar to that in singleton pregnancies, provided that sufficient fetal fraction is present . When methods for estimation of fetal fraction are used, which evaluate fetal fractions of each dichorionic twin, the lowest fetal fraction is taken to reduce the risk of a false-negative result . Likely because of this choice, most studies report a lower fetal fraction in dizygotic twins and a higher failure rate when a cut-off of 4% is taken compared to singletons. Test failure is associated with high BMI and in vitro fertilization . This field is rapidly changing, with likely improved screening in twins in the future. Obviously, counseling pregnant women with multiple gestations on antenatal screening and diagnosis should include discussing the possibility of finding discordant outcomes. Selective termination of an affected fetus is certainly possible, but carries some risk of losing the whole pregnancy, including the healthy co-twin. This should be discussed before performing invasive diagnostic testing. In particular, in the absence of discordant visible structural anomalies, careful mapping and labeling of the samples is essential to enable correct performance of selective termination in case of discordant results. Ideally, the operator performing the selective termination should have performed the invasive test him/herself.

Maternal surveillance in multiple pregnancy

In multiple pregnancy, all pregnancy complications are more common and can be more severe than in singletons. In the first trimester, nausea and vomiting is generally worse, likely because of increased hormone production by the placentas. Rapid growth of the uterus may be associated with more pain, and from the second trimester onward, mechanical pressure on abdominal organs is increased. The standard intervals used in singletons for clinic visits to evaluate complications such as pregnancy-induced hypertension, preeclampsia, and diabetes gravidarum are often shortened in multiple pregnancies, although exact recommendations on frequency cannot easily be given. Even in singletons, such guidelines vary enormously in different countries and cultures. The same holds true for advice on whether and when to reduce physical activity, stop working, or other often-used but poorly supported measures .

Surveillance to reduce the risk of preterm birth

Gestational age at birth is reduced in multiple pregnancy, with the majority delivering before 37 weeks. Prematurity is one of the most important health problems in obstetrics, associated with neonatal and infant mortality and irreversible handicaps and huge costs of prolonged stay of newborns in neonatal intensive care units (Hack 2008). Twin pregnancy is a major contributor to this problem. Primary prevention, by more careful treatment of infertility, is increasingly recognized as highly valuable, although due in part to financial issues and health insurance policies; still ART results in too many multiple pregnancies .

Secondary prevention of prematurity in multiple pregnancy is not yet very successful. Studies aiming to prevent preterm birth in twins, e.g., by administration of progesterone, insertion of a cervical pessary, or cerclage, with or without first selecting a high-risk group using cervical length measurements, have thus far been disappointing . The current best practice advice would be to only provide medical or mechanical interventions in the context of appropriately designed and adequately powered randomized controlled trials. Because proven useful interventions are lacking, the routine (screening) performance of single or serial cervical length measurements by transvaginal ultrasound in multiple pregnancies is best only when performed within trials. Obviously, observing a short (e.g., <25mm) cervix in a symptomatic patient and the increased risk for preterm birth warrant specific management depending on gestational age.

Fetal surveillance in dichorionic twins

The main clinical fetal problem in the antenatal period for dichorionic twins, besides threatened preterm birth, is growth restriction. Manual palpation of the uterus cannot be used because it cannot assess each twin’s size separately. Serial ultrasound to be performed at least every 4 weeks in uncomplicated twins is the only useful tool. Plotting the measurements of head circumference, abdominal circumference (AC), and femur length in appropriate graphs and evaluating inter-twin differences as well as the trend of serial assessment is the current standard of care. Because the only reason for reduced growth in multiple pregnancies is placental function, any deviation from growing as a normal singleton is likely due to placental insufficiency. Therefore, the use of twin-specific growth charts is not advised. Although practically all twins show reduced growth after 30–32 weeks of gestation , this should not be regarded as normal but as a sign of placental insufficiency, warranting close surveillance similar to that in singletons.

Like uncomplicated singletons, there is no evidence that routine screening using Doppler evaluation of the umbilical artery blood flow or other vessels has any benefit. In dichorionic twins with (suspected) growth restriction, the same recommendations as clinicians would adhere to in singletons apply . Increasing surveillance; evaluation of amniotic fluid; Doppler blood flow evaluation of the umbilical artery, middle cerebral artery (MCA), and ductus venosus and evaluation of fetal movements are advised, identical to that in fetal growth restriction (FGR) in singletons. This is also true for decisions on even more frequent surveillance by admission to hospital and fetal heart rate (cardiotocography) evaluation. The only difference with singletons lies in the management decisions in case only one twin is growth restricted, and possibly, beneficial preterm birth for the smaller twin would negatively affect the normally growing twin (the feto-fetal conflict). Because this makes the management of these pregnancies complex, referral to specialized (often tertiary care) centers could be beneficial. The National Institute for Health and Care Excellence guidelines and the ISUOG Practice Guidelines recommend estimated fetal weight (EFW) discordance assessment at every ultrasound examination in twins from 20 weeks onward, with referral in case of a discordance of 25% or more .

In the last weeks of the pregnancy, determination of the fetal position, in particular of the presenting fetus, is important as a major factor in decisions, in which patient preferences must play a role on the mode of delivery. Discussing timing and mode of delivery in multiple pregnancies is beyond the scope of this chapter.

Fetal surveillance in monochorionic twins; first trimester

In addition to what applies in dichorionic twin surveillance, the presence of vascular anastomoses on the placental surface in all monochorionic twins results in a whole range of additional risks to monitor for. Early prediction of an increased risk for these specific monochorionic twin complications has been investigated, in particular by evaluating discordancy in nuchal translucency and CRL. For trisomy risk calculation, the nuchal translucency in monochorionic twins should be averaged. A significantly discordant nuchal translucency is also suggested to be associated with an increased risk for twin–twin transfusion syndrome (TTTS) . However, the predictive values are insufficient to be used as routine screening tools. Test characteristics for the prediction of TTTS of NT discordance of ≥20% were a sensitivity of 52–64% and a specificity of 78–80% . The same study found an NT discordance of ≥20% in 25% of monochorionic twins, with a risk of early fetal death or TTTS in more than 30%. The risk of complications was less than 10% if the NT discordance is < 20%. Similar considerations concern differences in CRL. A clear discordance is associated with an increased risk of structural and chromosomal anomalies, single demise, pregnancy loss, TTTS, and selective FGR (sFGR) . However, because concordant CRL does not rule out these complications, all monochorionic twins should be closely monitored, making risk group selection based on CRL discordance less useful. First trimester detection of a significant discordant NT (≥ 20%) or CRL (≥ 10%) in monochorionic twins warrants further evaluation by a fetal medicine expert, and early referral is strongly advised .

A specific diagnosis to be made in the first trimester is twin reversed arterial perfusion (TRAP) sequence. This occurs in 1% of monochorionic twins and is thought to be caused by a particular vascular anastomosis configuration. In almost all cases, there is an arterio-arterial (AA) anastomosis directly from one of the umbilical arteries of the so-called pump twin to the (often single) umbilical artery of the other, abnormal, twin, with pulsatile blood flow toward the latter twin. This reversed arterial blood flow is also present in the fetal aorta and leads to early disruption of fetal heart development and, due to the low oxygen content of this blood, to underdevelopment of the upper part of its body. Occasionally, some cardiac activity can still be seen in early gestation. In the majority of the cases, however, the twin with the reversed arterial blood flow has no visible heartbeat and hence the traditional name “acardiac twin.” The observation of heart activity in some cases makes it, in our view, more likely that the reversed perfusion is indeed the cause of the disease. An alternative theory on the origin states that early spontaneous demise of one monochorionic twin does not, like in dichorionic twins, lead to a “vanishing twin” situation but is followed by inflow of arterial blood through a large AA anastomosis, preventing the twin with the arrested heart activity to really die. On color Doppler examination of the umbilical cord and aorta of the acardiac twin, the reversed pattern is easily identified. Poor oxygenation often leads to major structural abnormalities, especially in the upper body, and generalized edema. The unsuspecting sonographer may erroneously diagnose this situation as demise of one twin and may only be aware of another diagnosis when movements of the legs of this twin are observed. Early detection and appropriate referral are especially important because interventions may improve the prognosis for the “pump twin” who is at risk for cardiac failure and early demise. Another risk of such pregnancies is the development of polyhydramnios due to polyuria of the pump twin, for which the cause is unclear. High-output cardiac failure may contribute to the polyhydramnios, although in TRAP sequence, it occurs more often than in other fetal pathologies associated with a chronic high-output state such as tumors and anemia. Untreated, the polyhydramnios may cause rupture of membranes and immature birth .

All treatment options in TRAP sequence aim to arrest the blood flow from the pump twin to the acardiac twin. It remains controversial whether to offer intervention to all women with a TRAP pregnancy as soon as the diagnosis is made or to select a high-risk group based on the size of the acardiac twin or signs of cardiac overload (ductus venosus waveforms). In addition, when elective intervention is offered, it is yet unknown if first or second trimester treatment leads to better outcomes . The perceived advantage of early intervention is that small instruments such as ultrasound-guided interstitial laser or radiofrequency ablation (RFA) can be used, and only a small “dead” mass remains in utero, reducing the risk of preterm birth. Later intervention, e.g., after 16 weeks, is still possible using RFA; however, other techniques such as fetoscopic laser or bipolar forceps cord coagulation are often preferred. These techniques may be equally or even more successful in arresting blood flow; however, they appear to be related to an increased risk of preterm rupture of membranes and preterm birth. A randomized trial (TRAPIST trial) is currently ongoing, aiming to answer the question of best timing of intervention in TRAP sequence.