Epidemiology

Estimates of the incidence of twin conceptions are hampered by the unknown number of abortions and early fetal deaths that occur in multiple pregnancies. Ultrasound studies in early pregnancy sometimes reveal the death and later reabsorption of one fetus in the first trimester – the vanishing twin syndrome ( ; ). Thus, the prevalence of twins at a certain time is all that can be accurately estimated. Furthermore, some multiple births may not be recorded if a pair, one stillborn and one liveborn, are delivered before 24 weeks’ gestation. The stillborn would be registered as an abortion and the liveborn baby appears in the records as a single birth.

The incidence of multiple births has steadily increased in the developing countries since the early 1980s ( ; ). In England and Wales it rose from 9.0/1000 births in 1980 to 14.8/1000 in 2001. The incidence has been stable since then, although it has markedly increased among women aged 45 and above ( Fig. 23.1 ). The incidence of triplets rose much faster still until 1998, quadrupling in 15 years. It subsequently declined and in 2008 the rate in England and Wales was 2.4 per 10 000 maternities ( ).

Trends between 1998 and 2008 in twin births (per 10 000 maternities), with illustration of rates within some age categories.

(Data from .)

The increase in multiple births is partly due to assisted reproductive technology (ART), such as ovulation induction and multiple embryo transfer in vitro fertilisation (IVF), in the treatment of subfertility ( ). However, there is also a contribution from the rise in average maternal age ( ), which is strongly associated with twinning ( Fig. 23.1 ). The rate of twinning associated with ART depends on many aspects of practice, including ultrasonic monitoring of the response to ovulation induction, the number of embryos transferred and the use of selective reduction. There is significant international variation in practice which is mirrored by variation in multiple rates associated with ART, and rates of multiple births are higher in the USA ( ).

Most of the international and temporal variation in twinning is explained by variation in dizygotic (DZ) twinning (see below). This is partly explained by varying use of ART methods, but is also explained by racial variation in the frequency of DZ twinning. In general, black Africans have the highest rates; the Far Eastern, mongolian, races the lowest; the rates for Asian Indians and Caucasians lie between these. The prevalence of monozygotic (MZ) twin births had been constant worldwide at 3.5/1000 maternities until recently. A small increase in the MZ twinning rate has been noted since the 1980s ( ) and this may also reflect increased use of ART as separation of an embryo into MZ twins is more common with IVF conceptions.

Perinatal mortality

The perinatal mortality rate is over four times higher in twins and four to nine times higher in triplets. The increased rate of loss in multiples is explained by an increased risk of both stillbirth and neonatal death. In an annual review of all births and perinatal deaths in Scotland in 2007, the risk of stillbirth in multiple births was 16.1 per 1000, compared with the risk in singletons of 5.3 per 1000 (Information Services Division, NHS ). The risk of neonatal death in multiples was 18.0 per 1000, compared with the risk in singletons of 2.8 per 1000. The majority of the stillbirths lacked a clear explanation for the death, whereas the majority of the neonatal deaths were related to prematurity. The risk of stillbirth is strongly associated with chorionicity and this is discussed below. There is also an increased risk of death of the second twin as a result of anoxia, arising from complications following vaginal birth of the second twin, which is apparent in term births following vaginal delivery of twin one but is not apparent among infants delivered by elective caesarean section ( ). Twins conceived as a result of infertility treatment tend to be delivered earlier and to be of lower birthweight than spontaneously conceived twins ( ), and babies from these pregnancies have a higher perinatal mortality rate than those spontaneously conceived ( ). The explanation is uncertain but the cause of the parental subfertility may be a contributory factor.

Zygosity

DZ twins arise when two ova are released and fertilised in one menstrual cycle, whereas MZ twins arise when one ovum is fertilised and the resulting zygote divides into two. The ratio of MZ to DZ twins varies in different populations. In the UK approximately two-thirds are DZ, so, in all, about one-third of the pairs will be of unlike sex, one-third both girls and one-third both boys. MZ splitting appears to be 6–12 times more common following ovulation stimulation, whether or not fertilisation took place in vitro ( ; ). There also appear to be some extremely rare examples of MZ twinning occurring as an autosomal-dominant trait ( ). DZ twinning is known to be affected by a number of factors in addition to race, many of which appear to be related to differing maternal gonadotrophin levels. The rates are known to increase with maternal age, height, parity and frequency of intercourse ( ).

Chorionicity

Given that there are two fetal membranes, the amnion and chorion, there are three possible combinations of sacs in twin pregnancy:

• 1
  

  both twins may have completely independent membranes: dichorionic and diamniotic (DCDA)

  
  
• 2
  

  both twins may have separate amnions but share the chorion: monochorionic, diamniotic (MCDA)

  
  
• 3
  

  both twins may share both membranes: monoamniotic (MA).

The relationship between chorionicity and zygosity is complex ( Fig. 23.2 ). For practical purposes, MCDA and MA twins are invariably MZ. However, DCDA can follow both MZ and DZ twinning and further tests are required postnatally to determine zygosity in non-sex-discordant twins. Chorionicity is best determined by ultrasound in the first trimester of pregnancy (see Ch. 9 and Fig. 9.14 ) when the presence of the lambda sign (DCDA) or T sign (MCDA) has an accuracy close to 100% ( ). However, even in the third trimester, chorionicity can be determined with reasonable accuracy. Diagnosis of MA twinning requires detailed and expert assessment as a thin dividing membrane can easily be missed with suboptimal ultrasound.

Relationship between zygosity and chorionicity.

(Reproduced with permission from .)

Monoamniotic twins

MA twins arise when the splitting of the single zygote is delayed until the end of the second week after fertilisation and account for approximately 1% of MZ twin pregnancies. Earlier studies have reported perinatal mortality rates as high as 30–70% ( ) but in cases diagnosed during the second trimester the mortality rate is probably much lower – about 10%. Nevertheless, MA twins represent a serious and shared challenge to obstetricians and neonatologists in balancing the risk of intrauterine death against the complications associated with preterm delivery. The primary cause of fetal death is cord compression as a result of cord entanglement. There have been no randomised trials on the management of MA pregnancies but protocols include intensive ultrasound surveillance, sulindac to reduce amniotic fluid volume and delivery by planned caesarean delivery at 32 weeks’ gestational age ( ). Superficial vascular anastomoses are common in MA pregnancies ( ) and twin–twin transfusion syndrome (TTTS) is rare.

Conjoined twins

Conjoined twins occur in approximately 1/50 000 pregnancies ( ). They are a form of MZ twinning in which the division of the zygote is incomplete. The increased incidence of additional unrelated malformations suggests that conjoined twinning may be associated with a fundamental disturbance of embryogenesis ( ). The factors associated with conjoined twinning are female sex ( ; ; ), triplet sets with an MA pair ( ) and it may be more common in parts of Africa ( ). Seasonal clustering has also been reported ( ). The site and extent of the fusion are highly variable. Thoracopagus is the commonest form and accounts for about 70% of cases ( ; ). Inevitably ethical dilemmas arise with conjoined twins, particularly when surgical separation means that preserving the life of one is likely to be at the expense of the other ( ).

Monochorionic, diamniotic twins

MCDA twins have an additional risk to other twin pregnancies, namely the complication of TTTS ( Ch. 9 ). All monochorionic placentas have vascular anastomoses between the two fetal circulations: intertwin transfusion is therefore a normal event. It is only when this transfusion becomes unbalanced, i.e. with net transfer of blood from one twin (donor) to the other (recipient), that TTTS occurs. This haemodynamic imbalance in the parenchymatous arteriovenous network is associated with an absence of superficial anastomoses ( ). Acute TTTS results from an acute haemodynamic imbalance across the superficial arterioarterial or venovenous anastomoses. It is less common than chronic TTTS and tends to occur during labour, when it can cause severe hypovolaemia in the donor and hypervolaemia in the recipient ( Fig. 23.4 ). However, the most common presentation of TTTS is chronic. It complicates, to a varying degree, up to a third of monochorionic pregnancies and is the cause of up to 20% of perinatal mortality in twins. In the past, the diagnosis was made on the neonatal criteria of an intrapair cord blood haemoglobin difference of 5 g/l and birthweight discrepancy. It is now recognised that, in the more severe cases, presenting in the second trimester, these differences are usually absent and the main problems for the recipient fetus are severe polyhydramnios, cardiac hypertrophy, tricuspid regurgitation, venous overload and hydrops and, for the donor, oligohydramnios, sometimes resulting in a ‘stuck’ twin ( ).

The diagnosis of TTTS is based on amniotic fluid discrepancy measured as the maximum vertical cord free pool using ultrasound. The criteria for diagnosis are that the deepest pool in the recipient sac is greater than 8 cm and less than 2 cm in the donor sac. Thereafter, the condition is staged based on further ultrasonic criteria and the system described by is in widespread use:

• 
  

  stage I: bladder of donor visible

  
  
• 
  

  stage II: bladder of donor is not visualised for 1 hour but the Doppler is normal

  
  
• 
  

  stage III: Doppler abnormalities

  
  
• 
  

  stage IV: hydrops is present

  
  
• 
  

  stage V: demise of one or both twins.

In the newborn, the recipient twin is characteristically heavier and polycythaemic and faces the complications of high blood viscosity, such as cardiac failure, hyperbilirubinaemia and intravascular thromboses, whereas the donor shows signs of intrauterine growth retardation, anaemia and hypoproteinaemia. Renal failure and/or renal tubular dysgenesis due to chronic renal hypoperfusion in utero may also occur ( ). When anaemia is the presenting feature, care must be taken to distinguish from other causes of anaemia such as fetomaternal haemorrhage. Hypovolaemia in the donor leads to activation of the renin–angiontensin system, which has trophic effects on vascular smooth muscle. Postnatally, the treatment of the polycythaemic or anaemic twin is similar to singletons with the same problems ( Ch. 30 ). In addition, the recipient may have short- and long-term cardiovascular problems ( ). The risk of congenital heart disease is 15–23 times higher in twins affected by TTTS, typically pulmonary stenosis and atrioventricular defects. With the vascular and blood volume changes in TTTS the recipient twin is at risk of increased blood pressure and right heart disease (ventricular hypertrophy/cardiomyopathy, atrioventricular valvar regurgitation and right ventricular outflow obstruction) ( ). Right ventricular outflow obstruction may become severe enough to require valvotomy in infancy.

There is a relatively high risk of brain damage, often due to periventricular leukomalacia, in TTTS whether or not the donor twin dies, although the risk is undoubtedly increased when there is intrauterine death of one twin. The risk of necrotising enterocolitis is also increased ( ).

The obstetric management of TTTS is discussed in detail in Chapter 9 ( ).

Postnatal determination of zygosity

Like-sex DZ twins can only be distinguished reliably from the one-third of MZ twins who have dichorionic placentas by DNA analysis. The minisatellite probe test is a highly reliable tool ( ), requiring only small quantities of blood or other tissues. Tests are most practically done on placental samples at delivery or on cheek swabs taken from infants or children. Cord blood is frequently used but great care is required to avoid contamination with maternal blood. Blood groups or other genetic markers such as red cell enzymes, serum proteins and tissue enzymes (in particular from the placenta) have all been used for determining zygosity but DNA analysis has now become the standard method. The steps to determine zygosity in newborn twins are shown in Figure 23.3 ( ).

Flow chart for determining zygosity. MZ, monozygotic; DZ, dizygotic.

Fetal growth in multiple pregnancy

The growth and development of the twin fetus are affected not only by the same intrauterine factors as the singleton fetus but also by the interaction with the second fetus. At best one fetus must compete for nutrition, and at worst one twin may be severely, even lethally, damaged by the co-twin ( ). Intrauterine growth rates differ between singletons and twins from about 26 weeks, with the divergences then increasing with increasing gestational age ( ; ; ; ). As with singletons, primiparous pairs grow more slowly than multiparous ( ). The typical pattern of intrauterine growth of twins is similar to that of a growth-retarded singleton in that the weight falls disproportionately more than the occipitofrontal circumference. Birthweight and occipitofrontal circumference centile charts for English twins are now available ( ). Some 52% of twins and 92% of triplets are of low birthweight (<2500 g) compared with 6% of singletons, and as many as 10% of twins and 32% of triplets are of very low birthweight (VLBW; <1500 g) ( ). Indeed, twins and triplets contributed at least 35% of the VLBW infants in a population-based study ( ). The average weight of a newborn twin is about 800 g less than a singleton ( ) but if allowance is made for differences in gestational age the discrepancy is reduced to 500 g. Opposite-sex twins tend to be heavier than like-sex twins ( ), and boys are heavier than girls in both like- and opposite-sex pairs ( ). Dichorionic twins tend to be heavier than monochorionic ( ) but DZ twins are heavier than both di- and monochorionic MZ twins ( ). Discordant growth refers to interfetal differences in birthweight between the large and small infant, expressed as a percentage of the large twin’s weight: its defining value varies from 10% to 25% in different reports. Birthweight discordance may be due to the different sites of implantation of the two placentas or of the umbilical cords, but the commonest cause of large discrepancies in fetal growth is probably haemodynamic imbalance in the chronic form of TTTS ( Fig. 23.4 ). Discordant size during the first half of gestation is likely to be associated with an intrinsic factor such as a malformation ( ). However, twin pregnancies where the babies are ultimately discrepant for birth weight exhibit difference in growth even as early as 11 weeks’ gestational age ( ).

The fetofetal transfusion syndrome, showing intrauterine growth retardation and pallor of the donor (left) and plethora of the recipient.

Antenatal management

A scheme for the typical antenatal management is outlined in Figure 23.5 . When twins are identified, a scan should be requested at around 12 weeks’ gestational age by a sonographer who is certified for both determination of chorionicity and measurement of nuchal translucency. Chorionicity is most accurately determined at this gestational age. Moreover, this is within the window for measurement of nuchal translucency to estimate Down syndrome risk. This is important as the serum measurements used to assess Down syndrome risk cannot generally be applied in twin pregnancies. Routine scans are generally scheduled to assess fetal growth and well-being at 4-weekly intervals in DCDA twin pregnancies and 2-weekly intervals in MCDA and MA twin pregnancies. Moreover, many units would offer an earlier anomaly scan in these cases because of the higher frequency of congenital abnormality in MZ twins (see below).

Overview of antenatal management of twins. CS, caesarean section.

Chromosomal anomalies

These are usually discordant in DZ twins and, surprisingly, also occasionally in MZ twins ( ). The twins are then known as heterokaryotypes and it is assumed that the maldistribution of chromosomes occurred at about the same time as the twinning process. Heterokaryotypes XY/XO are the explanation for the occasional pair of MZ twins of different sexes ( ).

Klinefelter syndrome appears to be more common in twins and in their relatives ( ). There also appears to be an increased incidence of Turner syndrome in twins ( ) and of twinning among the (normal) family members of patients with Turner syndrome ( ). suggest that there may be a postzygotic mechanism common to twinning and X-chromosome loss. Genotypic discordance in MZ twins may be caused by a variety of mechanisms ( ; ). X inactivation is likely to be the cause of some cases of discordancy for genetic disorders such as Duchenne’s muscular dystrophy in MZ twins.

Some anomalies, such as oesophageal atresia ( ) and some cardiac malformations ( ), occur more commonly in twins. Cardiac malformations are usually discordant in MZ twins, suggesting that the malformation develops as a result of some process peculiar to MZ twins.

The teratogenic effects of both drugs and intrauterine infections may have intrapair differences in expression ( ). This may be because of a difference in fetal susceptibility. In other instances where the teratogenic effect of the drug is sharply limited, as in the case of thalidomide, the discordance could be due to the insult acting at the very beginning or end of the sensitive period. As DZ twins can be conceived several days apart, one embryo may be a few days retarded or accelerated in development and therefore escape unscathed.