In Depth

Twinning and Intrauterine Death

Richard M. Pauli, M.D., Ph.D.

Twinning and other plural births are recognized in about one in every 70 pregnancies. All except for twins are extraordinarily rare: triplets arise in only about 1 per 8000 pregnancies and quadruplets in 1 per 50,000. These estimates, of course, do not take into consideration changes which may result from medical interventions (e.g. use of agents which stimulate ovulation) nor do they include pregnancies in which there has been unrecognized early intrauterine death of a co-twin. Tragically both unrecognized and recognized intrauterine death is extraordinarily common in multiple gestation pregnancies. This article will review the process of twinning, how zygosity can be determined, the epidemiology of intrauterine death of twins, the mechanisms of intrauterine death in twins and the effects that the death of one twin may have on the surviving co-twin.

Types of Twins. Twins may be either monozygotic (=monovular, 'identical') or dizygotic (=polyovular, 'fraternal'). Dizygotic twins arise through the simultaneous fertilization of two eggs and their independent development from that point on. About two-thirds of all twins are dizygotic, although this varies considerably from one ethnic group to another. That this is true suggests that there is a genetic predisposition to this kind of twinning. And, indeed, genetic investigations support the idea that there are genes which influence the probability of polyovulation. Fraternal twins, therefore, tend to run in families. That tendency can be transmitted through both male and female relatives of an individual who has had twins, but, of course, can only be expressed through female relatives.

In contrast, the prevalence of monozygotic twinning is virtually constant across races and ethnic groups, arising once in each 250 pregnancies in all circumstances. Monovular twinning usually is non-familial. Indeed, many dysmorphologists consider monozygotic twinning to be an accidental abnormality of early development. In this view it would not be surprising if monozygotic twins demonstrated a higher frequency of birth defects in general, and, indeed, that is the case.

Monovular twinning may arise at any time from the two cell stage through the earliest stages of ectodermal plate formation, or from shortly after conception until around 15 days post conception. That timing is critical in determining the characteristics of the chorioamnionic membranes and with regard to what kinds of intrauterine complications may arise. If division into two embryos occurs within the first four days post conception then the embryos will develop in completely independent membrane units (i.e. they will be diamnionic and dichorionic); division during the blastocyst stage (4-8 days post conception) will result in the twin embryos sharing the outer chorionic cavity but they will be diamnionic; and if the splitting into two embryos arises from days 8 to 15 then the chorion and amnion already will have been determined and the resulting embryos will share both membrane enclosed cavities—they will be monochorionic and monoamnionic.

About one-third of identical twins are dichorionic and diamnionic, nearly two-thirds are monochorionic and diamnionic, and monoamnionic twin pregnancies (which include conjoined twins) are quite rare.

Determination of Zygosity. The definitive way of proving zygosity is through molecular testing (similar to the kinds of tests used to establish paternity). Monozygous twins are genetically identical and molecular genetic evaluations easily prove whether such identity is or is not present. However, logistics and expense preclude use of such molecular genetic zygosity determination in most settings. Physical features may be suggestive but are usually unhelpful in circumstances in which one or both twins have died in utero.

While not always determinative, evaluation of the fetal membranes is a crucially important way of assessing zygosity of twins. All monochorionic placentae (whether diamnionic or monoamnionic) arise only in identical twinning. So, two-thirds of identical twins can be identified as such simply by examining the character of the fetal membranes. Usually this determination can be done grossly (i.e. without microscopic assessment). If there are two separate placentae, then by definition the membranes will be dichorionic and diamnionic. If the placenta is single or fused, then the fetal surface should be examined. First, one determines if there are two amnions, or only one. If two, the amnion is then stripped away. A thick dividing chorion between two portions of the placenta indicates that it is dichorionic, while absence of such a chorionic division indicates that the twins are monochorionic.

Determination of zygosity is important in twinning associated with intrauterine death. Mechanisms of death that should be searched for differ in the two groups. Risks for a surviving co-twin also differ markedly. Implications for next pregnancies will also depend on determination of zygosity (since, for example, monozygous twinning is nongenetic and so will not tend to recur).

Twinning and Intrauterine Death: Epidemiology. Twinning carries substantial risk for intrauterine and neonatal death. About 15% of all twins die in the perinatal period. Those risks are considerably greater in monozygous twins and profoundly higher in monoamnionic pregnancies. Dizygous twins do have greater risk of intrauterine (and other perinatal) death, but those risks are quite modest when compared with the risks for monozygous twins.

In the WiSSP series, 8.3% of all stillborns were derived from plural births (about four times the expected rate). Most of those deaths were the result of monozygotic twin pairs: we can estimate that 85% of all of the twin deaths arose from monozygotic pairs. That, in turn, implies that monozygotic twins have about a ten to fifteen fold greater risk to be stillborn than do dizygous twins.

Twinning and Intrauterine Death: Mechanisms.
Most of what we understand concerning the mechanisms and timing of death in twins relates to monozygous twin pairs. Those mechanisms are inextricably linked to the timing of death and its relationship to the time of delivery.

First, about 1% of all monozygous twin pairs are conjoined (so-called 'Siamese' twins). conjoined twinning is associated with an extremely high risk of intrauterine or neonatal death. Secondly, there is a three fold excess of other significant primary malformations in identical twins. Many of these (sirenomelia, VATER association, exstrophy of the bladder, holoprosencephaly) are of sufficient severity that intrauterine or neonatal death may result from the malformations per se. Indeed, there is at least a three fold excess rate of miscarriage in monozygous twins and, when those embryos are carefully assessed, many end up having major malformations.

Vanishing twins can arise in both monozygous and dizygous twin pregnancies. This term refers to pregnancies in which early on a twin sac or other unequivocal evidence of twin pregnancy is found, but subsequently all evidence of one of the twins has disappeared. It arises when one twin dies in the first eight weeks of pregnancy. The phenomenon of the vanishing twin was once thought to be rare. With increasing use of ultrasound it is now recognized to be astonishingly common. In assisted ovulation pregnancies (where careful sequential ultrasound is performed) almost 50% of multiple pregnancies vanish. Some studies suggest that the vanishing twin phenomenon is the most common cause of first trimester bleeding.

Also associated with early twin death is a fetus papyraceous. Death of one twin at around 15-20 weeks gestation may not cause loss of the pregnancy or harm to the surviving co-twin. The dead fetus will be partially resorbed and retained. If retained for more than about 8 weeks prior to delivery, most of the fluid will be lost from the fetus and a paper-like, flattened and tiny fetus may be discovered within the membranes and placenta after delivery. Searching for a fetus papyraceous should be a routine part of placental exam after stillbirth since it can provide clues to mechanism of death of the other co-twin if found.

Midtrimester death of monoamnionic twins often results from cord entanglement. Sharing the same amnionic cavity implies high risk for such lethal accidents. Obviously, this can't occur with other forms of twinning.

Monochorionic twins share placental structures. Not infrequently this shared placenta will include accidental communications between blood vessels of the two twins. (It is very rare for such communications to arise in dizygous twins.) The effects of such vascular connections depend upon their size (small ones with low flow or equal pressures are almost always found and are of no consequence). Arteriovenous communications within the placenta means that blood from a high pressure system (artery of the 'donor' twin) can be shunted into a low pressure one (the vein of the 'recipient' twin). This then can result in the twin-twin transfusion syndrome. Both twins can be harmed by such a process. The donor will show anemia, growth failure and, often, oligohydramnios, while the recipient will be bigger, plethoric and volume overloaded. Both twins are at risk for intrauterine or neonatal death associated with such a process. About 20% of monochorionic twin pairs show the effects of significant twin-twin transfusion. In contrast, arterioartery communications are considerably less common. They will have an effect only if one high pressure system 'overwhelms' the other. If that occurs then there will be reversal of flow in the 'defeated' co-twin, which reversal of flow has devastating consequences. The defeated twin is oxygen deprived and has particularly poor perfusion of the upper part of the body. This then results in a profoundly anomalous and lethal disruption in this twin, called an amorphous-acephalus (and sometimes acardiac) twin.

Asynchronous death of one twin will often result in the death of the second, otherwise healthy, monozygous twin as well [Figure]. Two mechanisms, both dependent on vascular communications within the placenta, have been proposed to explain this phenomenon. When one twin dies it may release sufficient amounts of thromboplastin from dead tissues to cause generalized clotting problems in the survivor. Alternatively, death may simply result in a sudden drop in blood pressure and resistance which then results in massive, acute shunting of blood to the dead co-twin. The survivor then suffers profound volume depletion and hypotension. Such mechanisms can not only cause intrauterine death but can result in non-lethal sequelae in the co-twin. The abnormalities sometimes demonstrable are all explicable on the basis of disruptional effects of decreased blood supply (either secondary to thrombi/emboli or hypotension). They include porencephaly, limb reduction defects, intestinal atresias, aplasia cutis, and so forth. Tragically, nearly one-third of twins who have had a co-twin die after around 20 weeks gestation are shown to have major (usually central nervous system) problems.

Conclusions. Twinning results in marked increased risk of intrauterine death. A variety of mechanisms can cause these losses, most of which only affect monozygous twin pairs. Careful assessment following stillbirth of one or both of a twin pair can help answer the question, 'Why did this happen?' and, in most instances, provide reassurance that death arose from an accidental process inextricably linked to twinning itself and thus usually implying little or no risk for recurrence in next pregnancies.

Further reading*
Benirschke K (1993): Intrauterine death of a twin: mechanisms, implications for surviving twin, and placental pathology. Semin Diagnost Pathol 10:222-231.

Benirschke K, Driscoll SG (1967): The placenta in multiple pregnancy. Handb Pathol Histol 7:187-226.

Benirschke K, Harper VDR (1977): The acardiac anomaly. Teratology 15:311-316.

Jones KL, Benirschke K (1983): The developmental pathogenesis of structural defects: The contribution of monozygotic twins. Semin Perinatol 7:239-243.

Levi S (1976): Ultrasonic assessment of the high rate of human multiple pregnancy in the first trimester. J Clin Ultrasound 4:3-5.

Livingston JE, Poland BJ (1980): A study of spontaneously aborted twins. Teratology 21:139-148.

*Copies of these and other relevant articles are available for personal use by request from WiSSP.