21 Mark D. Kilby1 and Dick Oepkes2 1 Centre for Women’s and Newborn Health, Institute of Metabolism and Systems Research, College of Medical and Dental Sciences, University of Birmingham, Birmingham, UK 2 Department of Obstetrics, Leiden University Medical Centre, Leiden, The Netherlands Multiple pregnancy has a global impact on both maternal and perinatal risk in any pregnancy and impacts on society in terms of both social and economic effects. Improvements in the health of the population, and in particular perinatal care, have led to a reduction in overall total pregnancy complications (both maternal and perinatal). However, the proportion of these attributed to twins and higher‐order pregnancies is increasing and of significant importance. Almost every maternal and obstetric problem occurs more frequently in a multiple pregnancy and there are, in addition, a number of potential intrapartum considerations that complicate routine management. The modern management of a multiple pregnancy initially concentrates on the recognition of fetal risk, as mediated primarily by chorionicity, and then the monitoring of fetal growth and well‐being using ultrasound. Attempts to reduce the risks of preterm delivery and pre‐eclampsia in the mother are equally important and equally as frustrating (as in singleton pregnancy care), with little improvement in overall management of these conditions in the last 20 years. Recognizing the specialized nature of multiple pregnancy management has led to the publication of recommendations by two scientific study groups of the Royal College of Obstetricians and Gynaecologists (RCOG) [1,2] and the commissioning of recommendations of care for multiple pregnancies by the National Institute of Health and Care Excellence (NICE) in 2009 (and published in 2011) [3]. At the heart of care is the recommendation that such pregnancies are managed within specialist multidisciplinary teams and in designated multiple pregnancy clinics so as to organize antenatal, intrapartum and indeed postnatal care. The considerable geographical and temporal variation in the incidence of multiple pregnancy reflects factors including dizygotic twinning as the result of multiple ovulation [4]. The incidence of twinning ranges from 4 per 1000 births in Japan to as frequent as 54 per 1000 in some regions of Nigeria. In addition, this ‘complication’ is more prevalent in pregnancies with advancing maternal age (presumed to be secondary to the rise in follicle‐stimulating hormone concentrations). Familial predisposition to multiple ovulations (usually dizygous twinning) may occur and is presently best explained by an autosomal dominant inheritance pattern. In contrast, monozygous twinning, the result of early cleavage division of a single blastocyst, occurs with a relatively constant incidence of approximately 3.9 per 1000. Time trends in multiple pregnancy demonstrate a remarkable change in reproductive behaviour and consequence. Some of the first documented records from Scandinavia in the eighteenth century indicate that multiple pregnancy rates may have been higher than they are today, reaching a zenith of 17 per 1000 maternities [5]. However, during the twentieth century, twin pregnancy rates appeared to be in decline until the early 1970s, since when there has been a clear rise in prevalence [6]. Since the early 1980s, the twinning rate in the UK has risen from 9.8 to 13.6 and the triplet rate from 0.14 to 0.44 per 1000 maternities. Such an increase is reflected internationally, with the greatest rise being noted in the USA. Multiple pregnancies accounted for 1.6% of all births in the UK during 2007, with approximately 98% of these being twin births [6]. A considerable proportion of the increase is due to assisted reproductive technologies, such as superovulation (using antioestrogens or gonadotrophins) and in vitro fertilization (IVF) with embryo transfer. There is evidence that the number of multiple pregnancies is influenced by the number of embryo transfers and, as such, the number of multiple pregnancies associated with IVF has reduced since the recommendation that embryo transfer numbers be reduced. In addition, epidemiological evidence suggests that both types of assisted reproduction techniques increase the incidence of monozygous twinning by up to eightfold [7]. This is particularly associated with the techniques of ‘blastocyst hatching’. Monochorionic twins comprise 20% of spontaneous and 5% of iatrogenic twin gestations. This is important as monochorionic twins have the greatest pregnancy‐related complication rates. However, it is also important to recognize other influences. The association between increasing maternal age (strongest at 30–39 years) and spontaneous dizygous twinning is worthy of note. The combined effects of delayed childbearing and high uptakes of assisted reproductive technologies at advanced maternal age have been responsible for this rise [6]. Cumulative fetal loss rate in twins is up to five times higher (and in triplets 10 times higher) than in corresponding singleton pregnancies. Rates of stillbirth and neonatal mortality for a multiple pregnancy are 14.9 and 19.8 per 1000 live births, respectively. This high perinatal fetal loss and morbidity is largely attributed to the increased risk of prematurity and also intrauterine growth restriction (IUGR) with associated iatrogenic prematurity (irrespective of chorionicity, see following section). Cerebral palsy is approximately three times more common in twins and 10 times more common in triplets compared with singletons. These figures are per fetus, whereas the more relevant figure when counselling parents is the chance of their multiple pregnancies producing any one baby with these complications. Approximately two‐thirds of twins are dizygous and one‐third monozygous. However, it is chorionicity rather than zygosity that mediates the degree of perinatal risk in any individual multiple pregnancy. This is most important as it is clinically identifiable. Cumulative fetal loss rates and perinatal mortality are up to five times higher in monochorionic twins compared with dichorionic twins [8]. This study, and a more contemporary one (during a period when modern management was possible including more widespread recognition of feto‐fetal transfusion syndrome and its fetoscopically directed treatment [9]), noted that in monochorionic diamniotic twin pregnancies, 85% resulted in double survivors, 7.5% in a single survivor and 7.5% in no surviving baby. These deaths occurred spontaneously or iatrogenically. Perinatal morbidity appears similarly related, with prenatally acquired cerebral lesions evident in the early neonatal period on ultrasound in up to one‐third of monochorionic twins compared with 3% of dichorionic twins delivered preterm [10]. Such excess morbidity and mortality is mediated predominantly (but not exclusively) through the inter‐twin placental vascular anastomoses that connect the two fetal circulations. Monozygotic pregnancies assume one of three placental configurations. Division within 3 days of fertilization results in separate dichorionic placentas, which in up to 50% of cases may have the appearance on ultrasound of being adjacent to each other and ‘fused’. Splitting after formation of the inner cell mass at 4 days after fertilization results in a single monochorionic diamniotic placenta, whereas splitting after 7 days results in monochorionic monoamniotic twins. Approximately one in five of all twins are monochorionic. Chorionicity may be clinically determined during pregnancy using ultrasound, with up to 90–100% accuracy in the first trimester. Ultrasound allows the following features to be ascertained. Chorionicity should be determined on ultrasound in all multiple pregnancies, as a screening test, ideally within the first trimester between 11 and 14 weeks of gestation (when specificity and sensitivity are greatest). Digital or hardcopy images should be stored demonstrating the signs of chorionicity. This is because chorionicity is relevant to: Such an examination should be routine and allow the stratification of prospective pregnancy care. In addition, ultrasonic visualization of the inter‐twin membrane/septum and genitalia may be difficult later on in pregnancy or when there is significant oligohydramnios complicating the pregnancy. The first ultrasound allocation of chorionicity should be performed in the first trimester between 11 and 14 weeks at a time when increasing numbers of pregnancies are undergoing nuchal translucency screening [11]. Ultrasonic sexing is performed for medical rather than social reasons in multiple pregnancies and, as such, achieves a high degree of accuracy, usually between 16 and 20 weeks. Monochorionic twins by definition are monozygotic, while discordant‐sex twins are dizygous. In the remaining 50%, zygosity cannot be determined without DNA fingerprinting, such as the polymerase chain reaction technique that compares parental inheritance patterns of a number of dinucleotide and trinucleotide short tandem repeats which are highly polymorphic in copy number. Such determination is rarely performed prospectively in clinical practice. Just as placental chorionicity is rechecked at birth (usually by clinical or histopathological examination), cord blood zygosity studies may be offered to parents of twins where there is indeterminate zygosity. Not only are parents curious but knowledge of zygosity influences the twins’ rearing, their sense of identity, their genetic risks and their transplantation compatibility. However, it is not routine to offer this in current practice within the NHS in the UK. Rarely there may be indications for zygosity testing in utero on invasive collection of fetal tissue, such as excluding contamination, deducing genetic risk or demonstrating dichorionicity in the presence of fetal compromise. Twins have a high incidence of spontaneous early pregnancy loss. Estimates suggest that approximately 12% of human conceptions start as twins [12]. Studies of ultrasound or miscarriage pathology indicate that twins are found at least twice as commonly in the first trimester as at birth. First‐trimester early pregnancy loss and resorption of one previously indefinable twin on ultrasound is known as the ‘vanishing twin’ phenomenon and is estimated to occur in up to 20% of twin pregnancies [13]. Spontaneous first‐trimester loss of one or more fetuses in high‐order pregnancies is estimated to occur approximately 50% of the time. When one twin dies in utero in mid‐trimester, a papyraceous fetus (the squashed paper‐like remains of the baby) may be found alongside the placenta after delivery. In some cases, this is only identifiable histopathologically. The widespread use of ultrasound in the first trimester and for routine mid‐trimester anomaly scanning to detect structural congenital malformations and Down’s syndrome is relevant to all multiple pregnancies. Zygosity determines the risk of congenital abnormality and chorionicity what can be done if it is found to be present. Zygosity may be deduced definitively in cases of monochorionicity or discordant external genitalia (dizygous), while in dichorionic concordant‐sex twins the chance of dizygosity is 75–80%. Monozygous twins have a 50% increase in structural abnormalities per baby. In particular, they have twice the frequency of congenital heart disease (a fourfold increase per pregnancy). Women with dizygous twins can be counselled that the chance of their pregnancy producing a child with Down’s syndrome is theoretically double their age‐related risk, whereas women with monozygous twins simply have their age‐related risk that both twins will be aneuploid. Serum screening is, in general, inapplicable to multiple pregnancies. In contrast, nuchal translucency and first’trimester ultrasound scanning as a fetal‐specific screening test is readily applicable and recommended by the National Screening Committee in the UK. At 18–24 weeks, women with multiple pregnancies should be offered a mid‐trimester fetal anomaly scan (which includes visualization of the four chambers of the heart and the great vessel outflow tracts) irrespective of chorionicity (as in singleton pregnancies). In countries with a gestational age limit for termination of pregnancy of 24 weeks, the anomaly scan should preferably take place before 22 weeks. In addition, in the first trimester between 11 and 13+6 weeks, all women with multiple pregnancies should be offered nuchal translucency screening for the detection of chromosomal anomalies (as well as the formal documentation of chorionicity). In dichorionic twins, the risk of aneuploidy is that of each of the individual fetuses. In monochorionic twin pregnancies, the risk of aneuploidy is the average between the twins. The use of first‐trimester serum screening as an adjunct to nuchal translucency, taking chorionicity into account, may slightly improve detection rates for Down’s syndrome in twins, but the results of large prospective studies are awaited [14]. Screening for Down’s syndrome using a maternal plasma sample to perform cell‐free DNA sequencing, also known as non‐invasive prenatal testing (NIPT), in monochorionic twins should be equally reliable as in singletons. However, reliable scientific data are still lacking. NIPT in dichorionic twins has been studied in small series, and appears promising, although with a higher failure rate and lower accuracy compared with singletons [15]. Invasive procedures in twins and other higher‐order multiple pregnancies are potentially complex techniques and should only be performed in fetal medicine referral centres [15]. The in utero topography (placental and membranes) is mapped using ultrasound. The location of the fetuses, the placental site(s) and the plane of the dividing septum in three dimensions should be noted and recorded. Such is a prerequisite for interpretation of discordant results and for selective termination of pregnancy. The operator performing the diagnostic procedure should also undertake any selective termination so as to minimize uncertainty and obviate any need for confirmatory invasive testing. In monochorionic twins, it is acceptable practice to sample only one fetus by either amniocentesis or chorionic villous sampling (CVS). However, rare cases of heterokaryotypic monochorionic twins may be missed (occurring in <1%). For this reason, amniocentesis on both amniotic sacs is worthy of consideration if monochorionic twins are discordant for structural anomalies, nuchal translucencies or growth. In dichorionic twins, there has been controversy about whether CVS is less desirable than amniocentesis for performing karyotyping. Because of problems with contamination, some investigators suggest restricting CVS to high‐risk cases such as monogenic disease or where there is an aneuploidy risk of greater than 1 in 50. The risk of contamination is likely to be higher than the published figures (2%) since the literature is confined to discordant‐sex twins. Any benefits of CVS are outweighed by the potentially disastrous consequences of misdiagnosis due to contamination, with subsequent termination of a diploid fetus or the wrongful birth of a fetus with a chromosome abnormality. For these reasons, the RCOG guidelines [15] discuss potential advantages to amniocentesis as the preferred option for karyotyping in dichorionic twins. Such a decision has to be weighed against the increased risks of selective reduction at increased gestational ages. When performing fetal blood sampling, the intrahepatic vein may be sampled to avoid confusing the cord origins in twins. There are no randomized controlled trials to indicate procedure‐related loss rates in twins. However, background loss rates are appreciably higher. A recent systematic review of the literature relating to the risks of CVS and amniocentesis in twin pregnancies was published in 2012. This study noted that for CVS, nine studies fulfilled the inclusion criteria. The overall pregnancy‐loss rate was 3.84% (95% CI 2.48–5.47; N = 4). The rate of pregnancy loss before 20 weeks was 2.75% (95% CI 1.28–4.75; N = 3) and before 28 weeks was 3.44% (95% CI 1.67–5.81; N = 3). For amniocentesis, the overall pregnancy‐loss rate was 3.07% (95% CI 1.83–4.61; N = 4). The rate of pregnancy loss before 20 weeks was 2.25% (95% CI 1.23–3.57; N = 2), before 24 weeks was 2.54% (95% CI 1.43–3.96; N = 9) and before 28 weeks was 1.70% (95% CI 0.37–3.97; N = 5). Pooled data from four case–control studies showed a higher risk (2.59% vs. 1.53%) of pregnancy loss before 24 weeks following amniocentesis (relative risk, RR 1.81; 95% CI 1.02–3.19) [16]. In dichorionic twins discordant for fetal anomaly, selective termination of pregnancy by the induction of asystole using an abortifacient is associated with an 8% loss rate in the international registry, with lower rates if the procedure is performed before 16 weeks’ gestation [17]. Selective termination of monochorionic twins cannot be performed using injection of an abortifacient as this would lead to death of the healthy twin due to sharing of the circulation along vascular anastomoses. However, a variety of cord occlusion techniques has been developed to render selective termination feasible. However, evidence shows that there is an associated increased risk of co‐twin demise and co‐twin morbidity when these procedures are performed [18]. Survival rates of the co‐twin vary between 70 and 80% in reported single‐centre series. All the normal physiological adaptations, such as increased cardiac output, glomerular filtration rate and renal blood flow, are further increased in a multiple pregnancy. Women with twins increase their plasma volume by one‐third more than women with singletons. Red cell mass increases approximately 300 mL more than in singleton pregnancies but because this is disproportionately less than the increase in plasma volume, haemoglobin and haematocrit values fall. Maternal iron stores are diminished in 40% of women with twins so routine haematinic supplementation is recommended (usually as combined iron sulfate and folic acid supplementation). Hyperemesis gravidarum is more common in multiple pregnancies and is managed as in singleton pregnancies. Severe cases may respond to maternal steroid therapy and require pyroxidine (B6) supplementation. The majority of minor pregnancy complications such as backache, symphysis pubis dysfunction, oedema, varicose veins, haemorrhoids and stria are all increased as a result of both the physical effects of greater uterine size and greater placental hormone production [19]. Hypertensive disease of pregnancy and pre‐eclampsia are up to 10 times more common in multiple compared with singleton pregnancies but are managed once diagnosed on standard principles (as in singletons). Consideration should be given to low‐dose aspirin prophylaxis but there are no national/international recommendations to this effect. Maternal pregnancy‐related hypertension remains a significant cause of maternal morbidity (and mortality) in multiple pregnancies and a significant cause of iatrogenic preterm delivery, increasing perinatal morbidity and mortality. This occurs in 15–20% of twin pregnancies, 25% of triplets and up to 60% of higher‐order multiple pregnancies [20]. Postnatally, the physical difficulties and socioeconomic impact of coping with the demands of two or more babies are considerable. Postnatal depression is more common in women nursing twins than singletons [21]. With the high perinatal loss rates there are often associated problems of postnatal grieving and bereavement. Families of women who give birth to babies after a multiple pregnancy may require additional social support. Ultrasound is the primary tool for monitoring growth in multiple pregnancies for several reasons. The risk of IUGR (~25%) is higher than in singleton pregnancy, and in two‐thirds of cases growth will be discordant (affecting one twin only). In addition, abdominal palpation and symphysis–fundal height measurement are unreliable as indices of growth in individual fetuses as they reflect total intrauterine growth. There is no proven agreement on the ideal frequency of ultrasound examination. However, it is common policy to scan dichorionic twins at up to 4‐weekly intervals from 24 weeks’ gestation with or without Doppler measurements as indicated. Monochorionic twins are often scanned more frequently, at 2‐weekly intervals, from 16 weeks onwards, as recommended by the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) practice guideline. At this gestational age, there is significant overlap in diagnosis between early twin–twin transfusion syndrome and selective IUGR. There is controversy as to whether singleton or twin biometric charts should be used. The former appears more sensible, as twins have a higher risk of IUGR with potential morbidity and the use of twin charts thus seems akin to using separate charts for other high‐risk groups. Furthermore, increasing emphasis is placed on growth profile and fetal condition (i.e. liquor volume estimation and umbilical artery Doppler velocimetry). Many centres use the measurement of discordancy in estimated fetal weight (EFW, which can be estimated using varied ultrasound parameters) as an index of discordant IUGR: This parameter has some predictive value in monochorionic twins for bad outcome in feto‐fetal transfusion syndrome and stillbirth [22], but in dichorionic twins it is a relatively poor predictor of perinatal death [23]. The standard principle of management of IUGR (i.e. delivery when the risks of continued intrauterine life outweigh those of extrauterine existence) needs modification in twin pregnancy to account for the risks to both twins. The latency between absent end‐diastolic flow velocity in IUGR in twins is longer before ‘pre‐terminal’ factors precipitate delivery than in singleton pregnancies. In addition, this latency is longest in monochorionic twins. However, careful and specialist surveillance of such pregnancies using cerebral, peripheral and intracardiac arterial and venous Doppler velocimetry is required. For example, cessation of fetal growth with pre‐terminal arterial and venous Doppler studies may warrant delivery at 26 weeks in a singleton fetus. However, discordant IUGR at such an early gestation in dichorionic twins might better be managed by allowing the severely affected IUGR fetus to die in utero, sparing the healthy fetus the risks of iatrogenic prematurity. Such risks and the balancing of decision‐making are always difficult and should be individualized. Decisions should be made in concert with parents and multidisciplinary teams (including neonatologists). In monochorionic twins, such decisions are even more complex. There is some evidence that the presence or absence of umbilical artery Doppler flow during diastole is indicative of prognosis. Positive end‐diastolic velocities indicate the best prognostic group when there is significant discordancy between monochorionic twins in terms of growth. Absent end‐diastolic velocity indicates an intermediate risk group and so‐called intermittent absent end‐diastolic velocity indicates the worst prognostic group and, in particular, the worst outcomes in terms of perinatal morbidity [24]. Indeed, in monochorionic twin pregnancies complicated by IUGR in one fetus, there is evidence that the ‘larger’ twin may have the highest morbidity in terms of neurodevelopmental sequelae. In some cases (of early onset) it may be necessary to consider selective cord occlusion rather than delivery of the whole pregnancy (depending on the gestational age of diagnosis). However, these again are difficult decisions evoking clinical complexity and parental anxieties. Management in a tertiary centre is therefore vital and individual discussions relating to procedure‐related morbidity and mortality are essential. Multiple pregnancies contribute disproportionately to preterm deliveries. Recent data indicate that, overall, 52.2% of multiple births deliver before 37 weeks and 10.7% prior to 32 weeks [25]. This is the major cause of neonatal death in multiple pregnancies: mortality rates are up to seven times higher in twins than singleton pregnancies; in triplets and higher‐order pregnancies, rates of nearly 40 per 1000 live births have been recorded [26]. The median gestational ages at delivery in twins and triplets is 37 and 34 weeks, respectively. However, the proportion of these pregnancies delivering before 30 weeks (twins 7%, triplets 15%) is much more concerning because of the associated long‐term morbidity. Parents should be informed of the symptoms and signs of threatened preterm labour and the advisability of early presentation. The stimulus to this increased risk is not clearly defined. Certainly (as in polyhydramnios) there is increased uterine distension (i.e. stretch), which may influence autocrine and paracrine intramyometrial processes. There is also focus on the potential maternal–fetal endocrine interaction, which may predispose to this increased risk. There is little evidence that the screening techniques available are highly predictive of preterm delivery (although some demonstrate promise); however, more reliable identification of twin pregnancies at risk of preterm birth may improve outcome if effective interventions are used. Management of preterm labour in multiple pregnancies differs little from that of singletons, except that the consequences of prematurity affect a greater number of babies. The following discussion concentrates on those aspects which appertain especially to multiple pregnancies. The prediction of preterm labour in twins and multiple pregnancies is as problematic as it is in singletons. One of the most promising methods of prediction of spontaneous labour in twins is the measurement of maternal cervical length using transvaginal ultrasound. A systematic review of 11 studies in the published literature (1436 pregnancies) has indicated the potential efficacy of cervical length in predicting risk of spontaneous preterm delivery in twins. Between 23 and 24 weeks’ gestation, the mean maternal cervical length is similar to that of singleton pregnancies (38 mm). At this gestation, a cervical length of 25 mm or less will have a positive summary likelihood ratio of 5.02 (95% CI 3.21–7.61) and a negative summary likelihood ratio of 0.75 (95% CI 0.54–106) for delivery prior to 34 weeks. This correlates with a change from pretest probability of preterm birth of 18.5% to a post‐test probability of 14.2% (12.9–15.9) with a negative test and 47.6% (38.9–56.4) with a positive test [27]. The use of home uterine activity monitoring or fetal fibronectin estimation [28] has not been demonstrated conclusively to be useful in prediction and therefore cannot be advocated. Preterm labour in a multiple pregnancy (as in polyhydramnios) is attributed to over‐distension (‘stretch’) of the uterus. Accordingly, there is no specific preventive measure (aside from fetal reduction in higher‐order multiple pregnancies as discussed below). Although hospitalization for bed rest has been widely practised in the past, there is little evidence to support its use. Critical appraisal of the literature and meta‐analysis of four randomized controlled trials indicate that bed rest in twins significantly increases the chances of preterm delivery, with a trend to greater perinatal mortality [29]. In contrast, a single randomized controlled trial in triplets demonstrated a non‐significant trend to less premature delivery and fewer neonatal deaths but was based on a very small number of cases [30]. Meta‐analysis of seven randomized controlled trials demonstrated that prophylactic β2‐sympathomimetic therapy was of no benefit in preventing preterm labour in twin pregnancies [31]. This was not surprising given its lack of efficacy in singleton pregnancies and is presumably due to tachyphylaxis. As in singleton pregnancies, this therapy is no longer used. In addition, cervical cerclage, and most recently vaginal progesterone therapy, has not been shown to be helpful and indeed may actually be harmful [32]. Indeed, an individual patient‐level meta‐analysis of randomized trials of cervical cerclage in women with a ‘short’ cervix (on ultrasound at 20–23 weeks) indicated an increased risk of preterm delivery prior to 35 weeks [33]. Most contemporary focus has fallen on the role of maternal administration of progesterone in potentially reducing the risk of preterm delivery. The Study Of Progesterone for the Prevention of Preterm Birth In Twins (STOPPIT) was a randomized, double‐blind, controlled trial to assess the role of daily vaginal progesterone (90 mg) for 10 weeks from 24 weeks’ gestation [32]. This study indicated that progesterone did not reduce the composite outcome risk of delivery or intrauterine death before 34 weeks in women with twin pregnancy. This effect was independent of chorionicity (although there was a trend towards worsening of outcome in monochorionic twins). Such findings are in concordance with other studies demonstrating no efficacy [34,35], in which progesterone was administered as intramuscular 17‐hydroxyprogesterone caproate (250 mg). Despite these findings, individual patient data analysis has indicated that vaginal progesterone may be effective in the reduction of adverse perinatal outcome in women with a cervical length of 25 mm or less; however, further research is warranted to confirm this finding [36]. Recently, there has been interest as to whether physical interventions, such as the fitting of an Arabin cervical pessary, reduces the risk of preterm birth in twin pregnancies with a short cervix (<25 mm). A recent study, performed under the collaborative umbrella of the Fetal Medicine Foundation found no significant improvement in the risks of preterm birth in the intervention group (compared to the conservative group). As such, there is no evidence that prophylactic measures, either physical or pharmacological, prevent spontaneous preterm labour in multiple pregnancies. The use of β2‐sympathomimetic infusions in multiple pregnancy, along with steroids and fluid overload, are known risk factors for the rare but potentially fatal complication of pulmonary oedema. As in singleton pregnancies, such active tocolysis has all but been abandoned. Equally, tocolysis with maternally administered oral nifedipine or intravenous atosiban (an oxytocin receptor antagonist) only leads to a relatively modest prolongation of gestation and studies informing efficacy of use in multiple pregnancies are sparse. Certainly, as in singleton pregnancies, the use of such therapy is usually only advocated to allow prophylaxis with corticosteroids. This is exemplified in a retrospective cohort study of 432 twin pregnancies (1982–1986), which noted that 54% of twins were born after spontaneous preterm delivery; of these, 23% were associated with preterm premature rupture of membranes and a further 23% were iatrogenic [37]. In these iatrogenic indications, 44% were secondary to maternal hypertension, 33% secondary to fetal compromise and/or IUGR, 9% secondary to antepartum haemorrhage and 7% associated with one or more fetal deaths. Maternal glucocorticoids have been clearly demonstrated to reduce the incidence of respiratory distress syndrome and its perinatal consequences in numerous randomized controlled trials [38]. However, only one uncontrolled study with separate data from multiple pregnancies suggested reduced benefits from antenatal corticosteroid administration in multiple pregnancies compared with singleton pregnancies [39]. The relative ‘resistance’ of multiple pregnancies to pulmonary surfactant maturation compared with singleton pregnancies has been postulated and has raised the possibility that a larger dose is required in multiple pregnancies but this remains to be objectively tested [40]. A report by Holmes et al. [41] has indicated that among 18% of 325 twin pregnancies delivering before 34 weeks’ gestation, 70% did so within 24 hours of presentation, the usual interval required for maximal corticosteroid efficacy. It has therefore been proposed that corticosteroids be administered prophylactically. Such a proposal is controversial. It is certainly theoretically possible that this could do more harm than good, as such therapy would potentially need to be administered on a weekly basis and there is no evidence that repeat doses of steroids are of increased value. Furthermore, and not withstanding the safety of steroids in follow‐up studies, there remain concerns about the potential adverse fetal effects of repeated steroid courses on glial formation and hippocampal development in children exposed in utero. Monochorionic twinning, which complicates 20% of all twin pregnancies, may be said to be a congenital anomaly of the placenta where the inter‐twin circulations communicate through placental vascular anastomoses. These occur in almost all monochorionic twin pregnancies. Bidirectional superficial arterial–arterial or venous–venous anastomoses may potentially compensate for any haemodynamic imbalance created by deep unidirectional arterial–venous anastomoses. Relatively small inter‐twin ‘transfusions’ are thus likely to be a normal physiological event in monochorionic twins. However, the imbalance of flow between twins may well be pathological and is always a potential risk in such pregnancies (increasing perinatal mortality considerably). When one of a monochorionic twin dies in utero, there is a significant risk of ischaemic damage predominantly to the fetal brain (18% in monochorionic twins), although there are reports of damage to pulmonary and hepatic systems, intestinal atresia, limb reduction and renal necrosis. For twin pregnancies overall, a recent systematic review derived a risk of 9% (95% CI 6–13) for neurological abnormality [42]. Death of one of a monochorionic twin substantially increases the risk of co‐twin demise in utero. Again, a recent systematic review has reported that surviving monochorionic twin fetuses have a six times greater risk of intrauterine fetal death (12%) following single fetal demise after 20 weeks’ gestation than initially surviving dichorionic twins (4%; OR 6, 95% CI 1.8–19.8) [42]. Gestational age at the time of intrauterine fetal death influences the extent and type of fetal brain injury. Single intrauterine fetal death after the second trimester can lead to periventricular leucomalacia, multicystic encephalomalacia or germinal matrix haemorrhage. In the third trimester of pregnancy, subcortical leucomalacia, basal ganglion damage or lenticulostriate vasculopathy may develop. There is debate about whether development of such central nervous system (CNS) anomalies are gestationally dependent. However, the long‐term morbidity of such an event prior to 14 weeks is controversial and the association less pronounced. Unlike dichorionic twins, discordant fetal compromise with risks of intrauterine demise in one twin have to be balanced against the potentially adverse effects of iatrogenic prematurity in the co‐twin if delivery needs to be expedited in monochorionic twins. A balanced and sometimes complex discussion may ensue, not only to prevent intrauterine death in the potentially compromised twin but also to prevent sequelae in the co‐twin. However, once single intrauterine fetal death has occurred in a monochorionic twin, delivery should not be immediate. The pregnancy should be evaluated for secondary sequelae, initially by the use of ultrasound. Such assessment should be within a tertiary fetal medicine centre. The assessment often includes the use of sequential ultrasound scanning (with increasingly the adjuvant investigation of fetal MRI) to prospectively evaluate the presence of CNS neuropathology that may develop up to 4 weeks after the sentinel event. Chronic feto‐fetal transfusion syndrome (FFTS) occurs in approximately 15% of monochorionic pregnancies (twins and triplets). It is responsible for up to 40% of deaths in monochorionic twins, where it is more commonly known as twin–twin transfusion syndrome (TTTS). The underlying pathophysiology involves chronic shunting of blood from the donor twin to the recipient twin, leading to an inter‐twin haemodynamic imbalance. Such a vicious circle of events leads to the ‘donor’ twin becoming hypoperfused, growth restricted and with associated oliguria and the development of anhydramnios. The co‐twin, the ‘recipient’, becomes polyuric with often severe polyhydramnios and a hyperdynamic circulation that may cause both diastolic and systolic cardiac dysfunction, ending in the development of hydrops fetalis and death (if no treatment ensues). FFTS/TTTS is diagnosed when there is gross discordancy in amniotic fluid volume in monochorionic twins, with polyhydramnios in the recipient and anhydramnios in the donor sac. It constitutes severe disease if the onset is prior to 26 weeks. A staging system has been described and is useful in annotating the condition in a consistent manner but does not always denote a logical order of disease progression (Table 21.1). In general terms, the prognosis is better in early‐stage disease (stages I and II) and worse in more advanced disease (stages III and IV). However, cardiac dysfunction may be present in up to 20% of fetuses with stage I disease (by Quintero staging) and as a consequence complicated monochorionic twin pregnancies may progress to the most adverse stages without warning. Table 21.1 Staging system for FFTS/TTTS. Untreated, perinatal loss rates in the mid‐trimester approach 95%. The principal clinical problem is severe polyhydramnios, which may be associated with premature rupture of membranes or preterm labour (or a combination of these) usually before 26 weeks’ gestation. In addition, differences in inter‐twin haemodynamics may be associated with single or double twin demise (with antecedent CNS morbidity). Fetoscopic laser ablation/coagulation of the inter‐twin communicating vessels has been demonstrated by critical appraisal of the literature to be the optimal treatment in FFTS/TTTS [43]. This is therefore the treatment of choice in monochorionic twin pregnancies presenting with this complication prior to 26 weeks’ gestation. Modification of the fetoscopic laser technique, using the SOLOMON method to ablate across the equatorial plane between arteriovenous anastomoses previously ablated using the selective sequential method, further reduces neonatal morbidity and, in particular, the complication of twin anaemia polycythaemia sequence (TAPS) [44]. The incidence of TAPS occurring spontaneously in monochorionic diamniotic twins is up to 5%. However, it may complicate up to 13% of cases of TTTS following laser ablation. TAPS is believed to be due to the presence of miniscule arteriovenous anastomoses (<1 mm) which allow slow transfusion of blood from the donor to the recipient, leading to highly discordant haemoglobin concentrations at birth (Fig. 21.1). The postnatal diagnosis of TAPS is based on the finding of chronic anaemia (including reticulocytosis) in the donor and polycythaemia in the recipient. The criteria for diagnosis include a difference in haemoglobin concentration between the twins of more than 80 g/L and reticulocyte count ratio greater than 1.7 or small vascular anastomoses (<1 mm in diameter) in the placenta. The prenatal diagnosis of TAPS is based on the finding of discordant middle cerebral artery (MCA) Doppler abnormalities, including an MCA peak systolic velocity (PSV) more than 1.5 multiples of the median (MoM) in the donor, suggesting fetal anaemia, and an MCA PSV less than 1.0 MoM in the recipient, suggesting polycythaemia. Additional ultrasound findings in TAPS include differences in placental echogenicity and thickness, with a bright thickened section for the donor twin and an echolucent thin recipient section. The polycythaemic twin might have a ‘starry sky’ appearance of the liver due to diminished echogenicity of the liver parenchyma and increased brightness of the portal venule walls. Fig. 21.1 TAPS is believed to be due to the presence of miniscule arteriovenous anastomoses (<1 mm). Source: Dr E. Lopriore. Reproduced with permission of Dr E. Lopriore. (See also colour plate 21.1) The outcome of twin pregnancies complicated by TAPS is variable. Severe TAPS may result in intrauterine death of both twins. At the other end of the spectrum, mild TAPS may still allow the birth of two healthy neonates (apart from having a significant haemoglobin difference between the two). It appears that the main neonatal morbidity consists of anaemia (requiring transfusion) and polycythaemia (possibly requiring partial exchange transfusion). However, cases of severe cerebral damage have been reported in neonates with TAPS. Recent evidence suggests that in monochorionic twins complicated by TAPS, the risk of neurodevelopmental delay is increased (20%) Therefore, brain imaging during the third trimester and neurodevelopmental assessment at the age of 2 years are recommended. The management options depend on the gestational age at assessment, parental choice, the severity of the disease and technical feasibility of intrauterine therapy. Therefore, the management of twin pregnancies complicated by TAPS should be individualized. The commonest options include conservative management, early delivery, laser ablation or intrauterine transfusion (IUT) for the anaemic twin, combined IUT for the anaemic twin and partial exchange transfusion to dilute the blood of the polycythaemic twin. In order to screen for TAPS, the MCA PSV should be measured from 20 weeks onwards in both fetuses, and during the follow‐up of treated TTTS cases. Prevention of TAPS by modification of the fetoscopic laser ablation technique remains the best way to prevent morbidity. This rare condition (complicating 1 in 35 000 pregnancies) arises in monochorionic twins with two umbilical cords often linked by large arterio‐arterial anastomoses. Flow from one, the ‘pump’ twin, supplies the other, the ‘perfused’ twin, in a retrograde fashion. The perfused twin almost always has associated significant congenital malformations, often including a rudimentary heart and aorta. The term ‘twin reversed arterial perfusion (TRAP) sequence’ is preferred, so named because reversed deoxygenated arterial supply is associated with only rudimentary development of the upper body structures within the fetus. Thus the acardiac twin is perfused by its co‐twin (the pump twin) via the inter‐twin placental anastomoses. Perinatal mortality in the pump twin in untreated cases is approximately 50% due to associated polyhydramnios and cardiac failure that may ensue [45]. Although polyhydramnios may be elevated by amnioreduction, definitive treatments that cause occlusion of the perfused (acardiac) twin’s cord or rudimentary aorta may be achieved by a variety of fetoscopic or ultrasound‐guided techniques (intrafetal laser ablation, radiofrequency thermal ablation and fetoscopic cord occlusion). There is still controversy whether to offer intervention electively, or only when signs of cardiac compromise occur, and at what gestational age intervention is best performed. Earlier intervention appears to be associated with a more advanced gestational age at birth. With careful case selection improved outcome for the pump twin is described in up to 85% [46,47]. Most significant complications include co‐twin demise or hypoperfusion (with cerebral morbidity) and/or preterm ruptured membranes. Of monozygotic twins, 1% lie in the same sac (monoamniotic) exposing them to the risks of cord entanglement. This may prove problematic most commonly (but not exclusively) in the intrapartum period. For this reason most cases are delivered by elective caesarean section. These twins also have higher reported overall perinatal mortality rates of approximately 30%. This appears largely related to the risk of sudden unexplained intrauterine death (often before 34 weeks’ gestation). Therefore the timing of delivery of these twins is controversial. Anecdotal and cohort studies have suggested that use of prophylactic maternal sulindac (cyclooxygenase‐2 inhibitor) to reduce fetal urine output and thus amniotic fluid volume reduces the risk of cord entanglement [48]. In combination with this, others have advocated hospitalization from 26–28 weeks onwards and elective premature delivery of these twins between 32 and 34 weeks’ gestation. The most recent evidence indicates that the risk of these events are relatively low, with close outpatient surveillance being recommended rather than very early elective delivery [49–51]. However, the consensus view suggests a course of prophylactic corticosteroids followed by elective caesarean delivery by 33 weeks in these rare twin pregnancies. Whatever the chorionicity of the twin pregnancy, it is best practice for intrapartum care to be discussed and a multidisplinary plan set out in the early third trimester of pregnancy. The indications for elective caesarean section are relatively few. Congenital anomalies associated with significant risk of cephalopelvic disproportion (including conjoined twins) would be an obvious indication and potentially monoamniotic twins would be another (see below). In addition, monochorionic pregnancies, complicated by placental anomalies associated with increased perinatal mortality (i.e. TTTS or TRAP), are generally delivered at 34–36 weeks and usually by caesarean section [2]. Perinatal mortality increases slightly after 38 weeks’ gestation in twins and there are therefore many obstetricians who elect for delivery then. However, there are no data to indicate whether or not this rise in mortality applies to twins whose growth and well‐being are known to be normal on ultrasound. Induction of labour is not contraindicated in twin pregnancies. Mode of delivery is decided on standard principles based on presentation of the first twin (cephalic in 70%, breech in 30%) and the documentation of optimal fetal growth and well‐being. Those with previous caesarean section scars are probably best delivered by repeat caesarean section because of the greater risk of scar dehiscence/rupture due to both uterine distension and intrauterine manipulation of the second twin. Recent data have indicated that a planned elective caesarean section may achieve an up to 75% reduction in the risk of perinatal death compared with vaginal delivery by reducing risks of acidosis and anoxia (especially to the second twin) [52–55]. Caesarean section has been advised when the first twin is breech, which would obviate the rare risk of interlocking twins and entrapment of the head of a presenting breech above the second cephalic twin. The use of intrapartum ultrasound may allow detection of such problems with early recourse to emergency caesarean section. However, there is no evidence that vaginal delivery of a presenting breech that would otherwise satisfy criteria for vaginal delivery (EFW less than 3.5–4 kg, flexed head and not footling) is inappropriate in selected cases. The presentation of the second twin is of no importance until after birth of the first. Even in twins where the first twin is cephalic presentation (at term), there may be the need for obstetric delivery of the second twin, with potentially an increased risk of perinatal morbidity. However, a Cochrane review (on delivery of the second twin not presenting cephalically) indicates that caesarean section increases maternal febrile morbidity without improving neonatal outcome [56]. This awaits critical appraisal. For vaginal delivery, continuous cardiotocography of both twins is best achieved by a combination of internal and external monitoring on a dual‐channel recorder. An intravenous line is sited and maternal blood drawn for group and save, in view of the increased incidence of caesarean section and postpartum haemorrhage. Augmentation of labour with oxytocin may be used as in singletons. An epidural anaesthetic is strongly advised in case of the unexpected need for internal manipulation of the second twin. If one is not sited, an anaesthetist will be required at delivery with early recourse to spinal or even general anaesthesia. The place of delivery is debatable but there is an increasing trend for twins to be delivered in operating theatres so that there is immediate redress to emergency caesarean section if necessary. Delivery of the first twin proceeds as for a singleton. Its cord is clamped to prevent fetal haemorrhage (from the second twin along any placental anastomoses). An experienced obstetrician discerns the presentation of the second twin, either by abdominal or vaginal examination or, increasingly, by the use of transabdominal ultrasound. Oblique or transverse lies are then converted to a longitudinal lie by external version and held in place by an assistant. Uterine contractions should be monitored and if necessary augmented using oxytocin. The membranes should be left intact to facilitate version. External cephalic version may be used to manipulate the fetal head over the pelvic inlet. Internal cephalic version is preferred as a primary procedure by many experienced obstetricians, as it seems to be associated with a higher success and lower complication rate than external cephalic version. One or preferably both feet are grasped and brought down into the vagina followed by an assisted breech delivery with contractions and maternal effort. Historical series suggest that the risk to the second twin is increased the greater the delay until delivery. Classically, intervals of greater than 30 min are acceptable providing the cardiotocograph is satisfactory and the presenting part is descending. Uterine inertia with a longitudinal lying second twin is corrected by oxytocin infusion. This is a not uncommon occurrence in the intrapartum management of twins. Fetal distress may be managed by ventouse delivery, even if the head is high or breech extraction. The already stretched vaginal tissues after the birth of the first twin allow these procedures in circumstances where they would normally be contraindicated. Caesarean section for second twin is occasionally indicated for disproportion, usually where the second twin is much bigger than the first. An oxytocin infusion is given prophylactically in the third stage of labour to minimize the risks of postpartum haemorrhage. There is some evidence that the risk of perinatal loss is greater at the end of the third trimester in monochorionic twins compared with dichorionic twins. However, there is insufficient evidence that elective delivery before 36 weeks improves outcome. Most current consensus‐based guidelines recommend delivery between 36 and 38 weeks’ gestation [3]. Perinatal and maternal risk increases exponentially with increasing fetal number. Most higher‐order multiple pregnancies are the result of assisted reproductive technologies and thus should be preventable with closer monitoring of follicular response and single (or, at most, two) embryo transfers in IVF therapy. Indeed, there are proven arguments for restricting the number of embryos transferred to one in order to minimize twin and triplet risk, and this course of action appears to have limited adverse effect on live birth rates when more cycles are allowed [57,58]. Every woman/couple with a higher‐order multiple pregnancy should have a discussion with a senior obstetrician relating to increased maternal and perinatal risks. This should involve the discussion and option of multiple fetal pregnancy reduction. In addition to perinatal mortality rates, parents should be counselled as to the mean gestational age at delivery (33 weeks for triplets, 31 weeks for quadruplets). In addition, 10% of triplets and 25% of quadruplets deliver before 28 weeks’ gestation, with severe neurological sequelae rates of 12% and 25% (respectively) in survivors [59]. The chief perceived disadvantage of multiple fetal pregnancy reduction, usually accomplished by administration of a percutaneous fetal intrathoracic injection of abortifacient (commonly potassium chloride), is complete miscarriage. International registry data demonstrate that this is lowest with reduction to twins, with rates for starting triplets and quadruplets of 7% and 15%, respectively [60]. There is now a consensus that multifetal pregnancy reduction between 10 and 12 weeks should be recommended for quadruplets and higher multiples so as to lower both maternal and fetal risks. The situation with triplets has been more controversial, with many considering this a social issue for parents. However, recent data indicate that in a fetal reduction group (N = 482) compared with an expectantly managed group (N = 411), the rate of miscarriage was significantly higher (8.1% vs. 4.4%, RR 1.83, 95% CI 1.08–3.16; P = 0.036) and the rate of preterm delivery lower (10.4% vs. 26.7%, RR 0.37, 95% CI 0.27–0.51; P <0.0001) [61]. Higher‐order multiple pregnancies should be managed in tertiary perinatal centres with a fetal medicine service. Care is almost always individualized. Management is along standard lines for twins but with greater emphasis on preventing preterm delivery and on monitoring fetal growth and well‐being. Although there have been successful reports of triplets and even quadruplets being delivered vaginally, most higher‐order pregnancies are now delivered by caesarean section. This alleviates difficulties with electronic fetal monitoring, avoids unrecognized hypoxaemia (especially given the high incidence of IUGR) and prevents birth trauma from manipulative delivery of non‐cephalic presenting fetuses. Given the higher incidence of preterm labour in the mid‐trimester, the option after delivery of the presenting fetus of conservative management with passive retention of residual fetus to prolong gestational age should be considered [62]. Increasingly, there is consensus opinion that the management of multiple pregnancies should be concentrated in a designated ‘multiple pregnancy clinic’ with experienced midwifery and obstetric discussion and decision‐making and with access to immediate diagnostic ultrasound and multidisciplinary opinions (i.e. anaesthetic, neonatal paediatric and psychological services). This care should be holistic in approach (in the widest sense) and could be organized regionally or in subregional centres depending on local population needs and numbers. Such clinics would allow the timely diagnosis of complications of multiple pregnancy along with an individualized plan of care for the prenatal, intrapartum and postnatal periods in women with multiple pregnancies. The rate of multiple pregnancies appears to be rising, a phenomenon elevated by increased maternal age and the use of assisted reproduction technologies. Even so, the greatest proportion of multiple pregnancies are twins. Obstetric care should be undertaken with specialist teams in a multiples clinic so that prenatal care (influenced by chorionicity), intrapartum care and postnatal well‐being may be discussed and planned prospectively. Such developments will hopefully minimize the increased maternal and perinatal risks that exist in such complex pregnancies.
Multiple Pregnancy
Incidence
Perinatal loss
Chorionicity and zygosity
Ultrasonic determination of chorionicity
Zygosity determination
Miscarriage
Prenatal diagnosis
Invasive procedures
Maternal homeostatic responses
Intrauterine growth restriction
Preterm labour
Prediction
Prevention
Management
Complications of monochorionic twinning
Acute feto‐fetal transfusion
Chronic feto‐fetal transfusion
Stage I
Polyhydramnios /oligohydramnios with bladder of the donor still visible
Stage II
Bladder of the donor not visible
Stage III
Presence of absent end‐diastolic flow velocity in the umbilical artery, reverse flow in the ductus venosus or pulsatile umbilical venous flow in either twin
Stage IV
Hydrops in either twin
Stage V
Demise of one or both twins
Twin anaemia polycythaemia sequence
Twin reversed arterial perfusion sequence
Monoamniotic twins
Labour and delivery of twins and multiple pregnancies
Higher‐order multiples
The concept of a multiple pregnancy clinic
Conclusion
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
