The frequency of multiple gestations has increased dramatically. Twins and higher order multiple gestations have pregnancies with increased risks for almost every complication of pregnancy, especially preterm labour, preterm delivery, and congenital anomalies. Monochorionic twins, by virtue of the unique placental angioarchitecture, are at risk for additional complications, such as severe discordant malformations, twin reversed arterial perfusion sequence, twin-to-twin transfusion syndrome or severe selective intrauterine growth restriction. These complications create unique challenges to those who manage multiple pregnancies. Reduction of higher order multiple pregnancies is on option to reduce pregnancy related risks and improve overall outcomes. Selective termination in complex monochorionic pregnancies can be lifesaving for the co-twin by preventing intrauterine demise or extreme prematurity. It is critical, however, to determine chorionicity before considering any approach to selective reduction. Techniques applied to dichorionic twins cannot be directly translated to cases involving monochorionic twins.
Introduction
The number of multiple gestations has substantially increased, primarily related to a more liberal use of assisted reproduction techniques. As many as two-thirds of the increase in multiple gestations have been attributed to the use of in-vitro fertilization (IVF) and ovulation induction . Of the 1645 deliveries from IVF pre-embryo transfer in Canada in 2001, 31.5% were multiple births . Similar rates occur in the USA , although the rate in Europe is lower at 24.5% .
It is well established that multiple gestations are associated with a higher rate of complications compared with singletons, much of which are related to increased frequency of prematurity and low birth weight . Perinatal mortality and morbidity increase with the addition of each fetus but, as twins are the most common outcome from IVF, they are the source for any excess in morbidity and mortality attributed to IVF. Maternal morbidity from hypertensive disease, caesarean section deliveries, and haemorrhage also increase in multiple gestations . Children born from multiple gestations also have increased rates of congenital malformations, longer after delivery hospital stays and increased risk for some degree of cognitive delay when born prematurely . All of these factors place increase stress on the family and societal resources.
Changes in patterns of practice in many parts of the world to limit the number of embryos transferred have decreased the rate of triplets and higher order multiples, although the number of twin deliveries has not decreased. In an attempt to further decrease these rates, both the American Society for Reproductive Medicine and the Society for Assisted Reproductive Technology have revised guidelines for the number of embryos to be transferred in IVF cycles . The Society of Obstetricians and Gynecologists of Canada and the Canadian Fertility and Andrology Society have issued similar guidelines . Tighter restrictions exist in the UK and Australia, where a maximum of two embryos can be transferred.
Although prevention of unnecessary multiple gestations is a goal, management of multiple gestations is now part of the routine practice of obstetricians and gynecologists. Higher risks for both the mother and the fetuses, discordance for fetal anomalies, issues unique to monochorionic multiple gestations (i.e. selective intrauterine growth restriction), necessitate a discussion of multi-fetal pregnancy reduction or selective fetal termination as management options. Pregnancy loss is not the only poor outcome in multiple gestations. The other main issue with which to be concerned is very early premature delivery. Not every woman will want to consider selective reduction. It is a highly emotionally charged issue for many families, especially if the pregnancy occurs after a long period of infertility. Women typically underestimate the risk of a multiple pregnancy, and think that complications happen to other people. Other women may not choose a reduction procedure for ethical or religious reasons. Additional barriers result if procedural expertise is not readily available or if healthcare insurance plans do not provide coverage for any procedure that is interpreted to be an abortion.
Determination of chorionicity is an explicit component on any ultrasound evaluation where a multiple gestation is diagnosed. The determination of the number of placental masses and the presence of a ‘Lamba sign’ or a ‘T sign’ at the base of the intertwine membrane are the two most useful ultrasound parameters to determine chorionicity. This determination is best established in the first trimester, and is supported by major medical organisations in many countries. Indeed, it has been recommended that no diagnosis of ‘twins’ takes place. There are only ‘monochorionic’ twins or ‘dichorionic’ twins . Although determination of chorionicity is critical for many aspects of the care of multiple pregnancies, it is essential for determining what therapeutic options may be appropriate for a woman considering selective termination. In addition, it has been advocated that a systematic process of identifying each twin be undertaken that goes beyond labelling, using a fetus’s proximity to cervix to designate Twin A . This is also best initiated in the first trimester. The fetus closest to the cervix is designated as Twin 1 and the relative orientation of the fetuses to each other is determined to be either lateral (left/right) or vertical (top/bottom). The advantage of this type of system is in its simplicity and applicability even in the presence of discordant sex multiples. Because relative orientation is determined by how the inter-twin membrane is positioned, it does not change during gestation. This process allows for serial biometry to be consistently applied to the correct fetus, for results of antenatal screening for aneuploidy to be correctly allocated and, in cases where selective reduction is undertaken, it assists in ensuring that the procedure is carried out on the ‘at risk’ or affected fetus.
Dichorionic multiple gestations
In the presence of a multiple gestation where there are only dichorionic configurations to the membranes, selective reduction can be easily carried out. Typically, this is a transabdominal percutaneous procedure using a 20 or 22-gauge needle. Under ultrasound guidance, the needle is positioned within the heart or thorax of the target fetus, and potassium chloride (20 Meq KCl solution in 10 cc) is injected. Ideally, the procedure is carried out between 11 and 14 weeks of gestation. Evans et al.m reporting on selective feticide for congenital malformations in 402 multiple pregnancies, showed a progressive rise in fetal loss in relation to gestational age at which the procedure was carried out: 5.4%, 8.7%, 6.8% and 9.1% for procedures carried out between 9–12, 13–18, 19–24, and over 25 weeks, respectively. These differences were not statistically significant. Procedures carried out between 11 and 14 weeks allows for information to be obtained that can assist in selecting which fetus to terminate. Invasive prenatal diagnosis by chorionic villus sampling can carried out before a reduction procedure. Nuchal translucency screening can also be carried out in this gestational age range. In addition, spontaneous loss of a non-viable pregnancy has likely already occurred, and a limited anatomic survey can be carried out to exclude any obvious anomalies. Non-invasive prenatal testing is not helpful in selecting which fetus to reduce, and its effectiveness in the context of multiple gestations is unknown . In the absence of additional factors, technical aspects usually govern the choice of which fetus to select, and the fetus furthest away from the cervix is chosen.
Risks for selective reduction are dependent on many factors. Overall, pregnancy loss rates from a large international study were reported at 9.6%, with a rate of early premature deliveries between 25 and 28 weeks at 3.7%. Increased experience was felt to be partly responsible for a decrease in complication rates . Strong correlations were observed among the starting number of fetuses, the finishing number of fetuses, and the likelihood of poor pregnancy outcome, including both pregnancy losses and prematurity. Stone et al. reported an overall loss rate of 4.7%, with a significant trend towards a lower loss rate with decreasing starting numbers of fetuses ( Table 1 ).
| Starting number | Stone et al. | Evans et al. |
|---|---|---|
| 2 | 2.1 | 5.8 |
| 3 | 5.1 | 6.5 |
| 4 | 5.5 | 12.3 |
| 5+ | 11.0 | 16.1 |
| 6 | No data | 22.8 |
Although it is clear that reduction of higher order multiples is of significant benefit, it is unclear whether the optimal number of remaining embryos should be three, two, or one. In a recently published meta-analysis comparing 1283 triplets reduced to twins with 1103 unreduced triplets, a significant decrease in delivery was reported less than 28 weeks (OR 0.3, 95% CI 0.18 to 0.5) and delivery less than 32 weeks (OR 0.36, 95% CI 0.27 to 0.46) in the group undergoing selective reduction. The rates of miscarriage (7.0% v 7.4%) and the take home baby rate (88% v 87%) were not significantly different. Although long-term outcome information is not available, intuitively it is reasonable to expect that less prematurity would result in healthier newborns. Reduction of triplets to singletons would not seem to be associated with a significant increased risk of pregnancy loss.
Although no consensus on the appropriateness of routine two to one reductions is ever likely to emerge, indications for carrying out selective termination of multi-fetal pregnancies have been described, which can be used to support this practice. These indications are (1) achieving a pregnancy that results in a live birth of one or more infants, with minimal neonatal morbidity and mortality; (2) achieving a pregnancy resulting in a live birth of one or more infants without antenatally detected anomalies; and (3) achieving a pregnancy that results in a singleton live birth . Selective termination in dichorionic twins has been carried out in the context of discordant structural or genetic anomalies. Studies would suggest that this is associated with a decreased risk for pregnancy loss or preterm delivery compared with unreduced twins, but the risk may still be above that for a natural singleton pregnancy . Moral and ethical discussions aside, it would seem that elective reduction to a singleton pregnancy results in a higher likelihood of taking home a baby compared with pregnancies that remain as twins .
Dichorionic multiple gestations
In the presence of a multiple gestation where there are only dichorionic configurations to the membranes, selective reduction can be easily carried out. Typically, this is a transabdominal percutaneous procedure using a 20 or 22-gauge needle. Under ultrasound guidance, the needle is positioned within the heart or thorax of the target fetus, and potassium chloride (20 Meq KCl solution in 10 cc) is injected. Ideally, the procedure is carried out between 11 and 14 weeks of gestation. Evans et al.m reporting on selective feticide for congenital malformations in 402 multiple pregnancies, showed a progressive rise in fetal loss in relation to gestational age at which the procedure was carried out: 5.4%, 8.7%, 6.8% and 9.1% for procedures carried out between 9–12, 13–18, 19–24, and over 25 weeks, respectively. These differences were not statistically significant. Procedures carried out between 11 and 14 weeks allows for information to be obtained that can assist in selecting which fetus to terminate. Invasive prenatal diagnosis by chorionic villus sampling can carried out before a reduction procedure. Nuchal translucency screening can also be carried out in this gestational age range. In addition, spontaneous loss of a non-viable pregnancy has likely already occurred, and a limited anatomic survey can be carried out to exclude any obvious anomalies. Non-invasive prenatal testing is not helpful in selecting which fetus to reduce, and its effectiveness in the context of multiple gestations is unknown . In the absence of additional factors, technical aspects usually govern the choice of which fetus to select, and the fetus furthest away from the cervix is chosen.
Risks for selective reduction are dependent on many factors. Overall, pregnancy loss rates from a large international study were reported at 9.6%, with a rate of early premature deliveries between 25 and 28 weeks at 3.7%. Increased experience was felt to be partly responsible for a decrease in complication rates . Strong correlations were observed among the starting number of fetuses, the finishing number of fetuses, and the likelihood of poor pregnancy outcome, including both pregnancy losses and prematurity. Stone et al. reported an overall loss rate of 4.7%, with a significant trend towards a lower loss rate with decreasing starting numbers of fetuses ( Table 1 ).
| Starting number | Stone et al. | Evans et al. |
|---|---|---|
| 2 | 2.1 | 5.8 |
| 3 | 5.1 | 6.5 |
| 4 | 5.5 | 12.3 |
| 5+ | 11.0 | 16.1 |
| 6 | No data | 22.8 |
Although it is clear that reduction of higher order multiples is of significant benefit, it is unclear whether the optimal number of remaining embryos should be three, two, or one. In a recently published meta-analysis comparing 1283 triplets reduced to twins with 1103 unreduced triplets, a significant decrease in delivery was reported less than 28 weeks (OR 0.3, 95% CI 0.18 to 0.5) and delivery less than 32 weeks (OR 0.36, 95% CI 0.27 to 0.46) in the group undergoing selective reduction. The rates of miscarriage (7.0% v 7.4%) and the take home baby rate (88% v 87%) were not significantly different. Although long-term outcome information is not available, intuitively it is reasonable to expect that less prematurity would result in healthier newborns. Reduction of triplets to singletons would not seem to be associated with a significant increased risk of pregnancy loss.
Although no consensus on the appropriateness of routine two to one reductions is ever likely to emerge, indications for carrying out selective termination of multi-fetal pregnancies have been described, which can be used to support this practice. These indications are (1) achieving a pregnancy that results in a live birth of one or more infants, with minimal neonatal morbidity and mortality; (2) achieving a pregnancy resulting in a live birth of one or more infants without antenatally detected anomalies; and (3) achieving a pregnancy that results in a singleton live birth . Selective termination in dichorionic twins has been carried out in the context of discordant structural or genetic anomalies. Studies would suggest that this is associated with a decreased risk for pregnancy loss or preterm delivery compared with unreduced twins, but the risk may still be above that for a natural singleton pregnancy . Moral and ethical discussions aside, it would seem that elective reduction to a singleton pregnancy results in a higher likelihood of taking home a baby compared with pregnancies that remain as twins .
Monochorionic multiple gestations
Assisted reproductive technology not only increases the likelihood of dichorionic twins, but, compared with spontaneous conception, the incidence of monozygotic and thus monochorionic twins increases three-fold after ovulation induction or conventional IVF and 13-fold after intracytoplasmic sperm injection .
The angioarchitecture of a monochorionic placenta is characterised by the presence of vascular connections between the circulations of the fetuses. The combination of number and type of vessels involved are unique to each pregnancy. The net effect is a dynamic bidirectional flow of blood between the co-twins. Abnormalities in either the nature of the connections or in the balance of blood flow create unique clinical entities. Monochorionic, diamniotic twin pregnancies can be complicated by significant clinical problems such as severe discordant malformations, twin reversed arterial perfusion (TRAP) sequence, twin-to-twin transfusion syndrome or severe selective intrauterine growth restriction. In some situations, selective termination of one fetus is necessary to optimise the chances for survival of the normal co-twin. The presence of vascular connections creates unacceptable risks for using intra-fetal injection as a method for selective termination. These vascular connections also link the outcomes of the fetuses. When intrauterine death of one fetus occurs, the risk of death or cerebral damage in the co-twin increases, likely because of acute exsanguination of the survivor into the lower pressure circulation of the dead twin through the placental anastomoses . Hypotension and hypoxia then lead to under-perfusion of the co-twin, causing tissue damage. This risk has been reported to be as high as 30–50% . In complicated monochorionic pregnancies, selective termination can be advocated as a therapeutic alternative to termination of the entire pregnancy.
A variety of occlusive techniques have been used to achieve selective termination in monochorionic twin pregnancies, including bipolar cord coagulation (BCC), laser cord coagulation, and cord ligation. These techniques require the insertion of a relatively large diameter instrument through a 3.8 mm operative sleeve into the amniotic sac of the fetus to be terminated, creating a significant risk for membrane complications, haemorrhage, and preterm labour. Radiofrequency ablation (RFA) is a technique that has been taken from the oncology field where it has been used for the local treatment of various neoplasms. Radiofrequency procedures use a high-frequency sinusoidal current (400–500 kHz) to induce local tissue ionic agitation that results in frictional heat. The energy dispersed at the active extremity of the inserted electrode produces local thermal injury. Radiofrequency energy is applied until there is ‘roll-off’. Roll-off indicates a drop in power output as tissue impedance increases as a result of tissue necrosis that prohibits the passage of electrical current. Radiofrequency ablation has been advocated as an alternative to selective termination procedures. It is potentially a less invasive option that uses a radiofrequency electrode with a 17-gauge (1.4 mm) diameter probe, and is carried out solely under ultrasound guidance. Two devices are in common usage. The first is a 17-gauge LeVeen RFA probe (Boston Scientific, Natick, MA). The other is a 17-gauge Starburst SDE™ radiofrequency needle (Angiodynamics, Inc., Queensbury, NY, USA). Each functions slightly differently, but the effect is the same. In a similar way, interstitial laser uses a 400–600 micron laser fibre passed down a 17- or 18-gauge needle. The needle is positioned under ultrasound guidance, and the laser fibre is advanced a few millimeters from the tip of the needle into the target tissue. Both NdYag and diode lasers have been used as the power source. Power is applied in short bursts until the tissue close to the fibre becomes echogenic and blood flow stops.
No perfect technique exists, and the choice is usually dictated by the technical considerations and available instrumentation ( Table 2 ). Bipolar cord coagulation is preferred when enough amniotic fluid allows for insertion of the operative sleeve and deployment of the device. Monoamniotic twin cases are best performed with bipolar cord coagulation because of the need for cord transection to prevent complications from cord entanglement once the termination is completed. If the cord segment to the demised fetus is left intact, it can act as a weight that can cause compression of the surviving twins cord. Radiofrequency ablation is preferred in cases of oligohydramnios or anhydramnios, in cases involving smaller fetal tissue volumes or when the umbilical cord leading to the twin to be terminated is short.
