The incidence of multiple pregnancies has increased over the past two decades. One-third of the increase is due to the use of ovulation induction agents, one-third due to assisted reproduction techniques, and the remaining third due to delayed child bearing. The management and the prenatal diagnosis of a multiple pregnancy is not only related to the number of fetuses, but it is strongly influ­enced by chorionicity. More than 30% are monozygotic twins and nearly 70% are dizygotic.

Monozygotic twins are the result of a division of a single fertilized ovum. It occurs in about 2.3-4/100 preg­nancies. The rate is constant and is not influenced by he­redity, age of the mother or other factors. Monozygotic twins are always of the same sex, have the same physical characteristics and the same genetic features.

In one-third of mono-chorionic twins, embryonic split­ting occurs within 3 days of fertilization resulting in sepa­rate fetuses with independent placental circulations. Split­ting after the third day is associated with vascular commu­nications between the placentas. In monozygotic twins chorionicity is subdivided into three categories. Around 20-30% are dichorionic-diamniotic with separate or fused placentas. The most common type of monozygotic twins is mono-chorionic-di-amniotic (two amniotic cavities with a single placenta) accounting for 70%. Around 1% of mono­zygotic twins are mono-chorionic-mono-amniotic [1]. Assisted conception also increases the incidence of mono­zygotic twins between 2- and 8-fold.

Monozygotic twins are at high risk of functional and structural abnormalities, of which approximately 10-15% would be affected [2]. Abnormalities unique to mono­zygotic multiple conceptions include conjoined twinning, fetus in fetus, acardia and fetus papyraceous. Explanation for the increased incidence of abnormalities in mono­zygotic twins involve the role of hemodynamic imbalance between mono-chorionic twins through placental vascular anastomoses.

Dizygotic twin are produced from separately fertilized ova released from separate follicles (very rarely from the same follicle) at approximately the same time. Dizygotic twins may be of the same or different sex. Factors influ­encing dizygotic twinning include race (most common in Blacks, least common in Asians) and nutrition. Parity does not influence the incidence of dizygotic twinning but ma­ternal age does, with the rate of dizygotic twinning peak­ing between 35-40 years of age and then declining sharply. Dizygotic multiple pregnancies trend to be recur­rent. Women who have borne dizygotic have a 10-fold increased chance of a subsequent multiple pregnancy. Height and weight have a positive influence on twinning.

Induction of ovulation with human pituitary gonado­tropin infertile women increase the incidence of multiple pregnancies, while clomiphene citrate increases the occur­rence rate of dizygotic pregnancies to about 5-10% [3].

The overall incidence of abnormalities appears to be higher in monozygotic twins compared with dizygotic including neural tube defects. Particularly, anencephaly, holoprocencephaly, sirenomelia complex, cloacal ex­strophy and abnormalities which fit into the expanded VATER/VACTERAL associations are more common in monozygotic twins. Assessing the risk of fetal abnormality in twins can be biased in many ways. Multiple pregnancies are intensively monitored by repeated ultrasound scan­ning, so the chances of detecting abnormalities are in­creased. Advanced maternal age increases the chances of twinning and this group of patients is more likely to have prenatal diagnosis.

The Risk of Aneuploidy in Multiple Gestations. Since the frequency of monozygotic twinning remains relatively constant with increased age, dizygotic twins become relatively more frequent as maternal age in­creases. In dizygotic twins, each embryo has an independ­ent risk for aneuploidy and therefore, the risk that at least one fetus is aneuploid will be twice the maternal age risk for a singleton. The probability of both fetuses being in­volved is minimal [4].

In cases in which zygocity is uncertain, calculation of the probability of an aneuplod fetus requires an estimation of the most likely zygosity of the pregnancy, which may vary based on maternal age and race. Using these calcula­tions, a 33-year-old woman carrying twins has a risk of at least one aneuploid child comparable to the risk in a 35-year-old woman carrying a singleton. On this basis, such women should be offered prenatal testing [5].

In monozygotic conception, both fetuses come from a single gamete, share the same karyotype, and hence, the risk of an aneuploid fetus is the same as the maternal age risk for a singleton. However, it has been shown that there is a lower risk for fetal aneuploidy in twin pregnancies compared to the delivery of live newborns.

Indications for Prenatal Diagnosis. The indication for prenatal diagnosis in multiple pregnancies is the same as in singleton pregnancies. These include advanced mater­nal age, a previous conceptus with chromosomal abnor­mality, a parent with a structural chromosome rearrange­ment, and the presence of gene(s) associated with inborn errors of metabolism. In case of an autosomal recessive disorder, the risk is increased and has a 3-in-8 chance of at least one affected fetus, and a 1-in-8 chance that both will be affected.

Prenatal diagnosis of chromosomal abnormalities in twins is complicated because effective methods of screen­ing such as maternal serum biochemistry are not applica­ble. Maternal serum alpha-fetoprotein (MSAFP) level in a twin pregnancy is twice as high as that in a singleton pregnancy. Most multiple gestations have high levels of MSAFP and 40% of twins are associated with MSAFP levels of more than a 2.5 MOM at 16 weeks [6]. The pres­ence of more than one fetus explains the elevation of MSAFP, but the possibility of an open neural tube defect in one or more fetuses cannot be excluded. In twins where the MSAFP exceeds 5 MOM, amniocentesis should be performed [7].

Nuchal translucency thickness measurement is a valid technique in multiple pregnancies for assessing the risk for aneuploidy [8]. In 448 twin pregnancies, the nuchal trans­lucency thickness was above the 95^th centile of the normal range in 65 of the 896 fetuses (7.3%), including 7/8 (88%) with trisomy 21. The minimum estimated risk for trisomy 21, based on maternal age and nuchal translu­cency thickness, was 1/300 in 19.5% (75/896) of the twins including all eight of those with trisomy 21. The detection of trisomy 21 in twins has a similar sensitivity to that in singletons [9]. Pandya et al. [10] reviewed the usefulness of nuchal translucency in 20,543 singleton and 392 twin pregnancies.

The main problem with nuchal translucency screening in twins is that the false-positive rate of the test is higher than that in singleton pregnancies. It has been suggested that this increase in mono-chorionic twins may be an early manifestation of complications arising from the shared placental circulation such as TTTS (twin-to-twin transfu­sion syndrome), but there is currently little evidence to support this [11].

O’Brien et al. [12] suggested that “current attempts to generate risk figures for aneuploidy in multiple gestations are fraught with imprecision and should be evaluated with great caution”. Prenatal diagnosis in multiple pregnancies may differ in several ways compared to singleton gesta­tion, and is strongly influenced by chorionicity.

In a di-chorionic pregnancy there are no direct conse­quences for the co-twin if a fetal demise occurs. In mono-chorionic twins, intrauterine death of one twin occurs can cause serious complication to the surviving fetus.

In fused placentas, chorionicity must be determined by the ultrasonographic appearance of the dividing mem­brane. The difference of thickness between the thick di-chorionic and the thin mono-chorionic membranes is much more obvious during the first trimester than it is later in pregnancy. Furthermore, the presence of an echogenic chorionic tissue projection into the base of the inter-twin membrane (“twin peak” or “lambda sign”) in the first trimester, has been shown to be one of the most specific ultrasound landmarks of di-chorionic placentation. In the second trimester the “lambda sign” is progressively less prominent, making prediction at chrorionicity more diffi­cult. Its absence after 20 weeks’ gestation should be viewed with caution [12].

Invasive Procedures for Prenatal Diagnosis. All invasive prenatal diagnostic procedures must be preceded by a detailed ultrasounds examination, and each fetus must be carefully examined by an expert in ultrasonog­raphy in order to evaluate the relative position, size, gesta­tional age, anatomy of each fetus, the location of the pla­centa(s), the extra-embryonic membranes and determine the chorionicity.

Amniocentesis for prenatal diagnosis is routinely per­formed from 15 weeks’ gestation, and there is only limited information on either the efficacy or safety of twin amnio­centesis performed prior to 15 weeks’ gestation. A 22 gauge needle is introduced transabdominally under ultra­sound guidance and 20 cc of amniotic fluid is retrieved. The technique of tapping multiple sacs requires two or more needle insertions and many maneuvers have been described to facilitate taps in twins, the most common probably being the use of a dye to identify the two sacs. Use of methylene blue is now contraindicated since it is associated with the formation of fetal small bowel intesti­nal atresias and may possibly lead to fetal death [13,14]. The pathophysiology of methylene blue induced bowel atresias is uncertain. However, it seems most likely that a vascular disruptive mechanism is responsible.

Indigo carmine is the dye most commonly utilized since it appears safe, but continued surveillance is war­ranted, since although this agent has minimal toxic effects, it does have a mild vaso-pressure effect when injected intravenously. The disadvantage with the instillation of indigo carmine is that the dye has a tendency to settle at the bottom of the sac, and even when it mixes well with the fluid, it takes some time before the stained fluid sur­rounds the fetus.

As ultrasound technology has improved, it may now be possible to forego the use of dye entirely, and three such approaches have recently been proposed. Jeanty et al. [15] described a single needle insertion technique. The needle entry is made into the proximal sac near the insertion of the dividing membrane and 20 cc of amniotic fluid re­moved. The stylet is then replaced, and under ultrasound guidance, the needle is advanced through the membrane into the second sac. After discarding the first 1 mL, to avoid contamination, 20 mL of fluid from the second sac is removed. The advantages of this alternative technique are that it requires only one insertion and this reduces discomfort. This technique is a swifter procedure, it re­duces the length of the procedure, it does not require the injection of a dye, it offers positive proof of tapping the two sacs and may reduce the incidence of complications.

Despite all these advantages, a number of potential disadvantages exist. This technique establishes a commu­nication between the two sacs. Because the opening be­tween the two sacs is very small it is likely that it will heal very rapidly. The puncture of the dividing membrane may result in an enlarging hole, and pseudo-mono-amniotic twins with entrapment of fetal parts or the umbilical cord, and the amniotic band syndrome [16]. Another drawback is that contamination of the needle with cells from the first sac or the dividing membrane could lead to an incorrect diagnosis of mosaicism for the second fetus. This compli­cation can be avoided by replacing the stylet prior to membrane perforation, and by discarding the first 1 mL of fluid from the second sac. Both these unfavorable events are not considered a clinical contraindication to the proce­dure.

An alternative approach to assure sampling from each sac utilizes two needles inserted simultaneously under ultrasound guidance. After aspiration of amniotic fluid from the first sac the needle is left in place and a second insertion is made into the other sac. The technique seems to be accurate and safe but experience with this approach is limited.

The possibility of chromosomally discordant results requires that amniotic fluid samples be labeled in such a way as to assure that the specific location of each fetus will remain identifiable. It has been suggested that the precise location of each fetus and placenta, both in rela­tionship to each other or to the cervix, and a detailed dia­gram should be drawn at the time of the procedure, in order to minimize the possibility of confusing the samples.

Most series of pregnancy outcome following second trimester amniocentesis have a pregnancy loss rate up to 20 weeks’ gestation of between 1 and 2.5%, and with a loss rate at less than 28 weeks being much higher. In an European multicenter study, the pregnancy loss rate was estimated to be 2.3% up to 20 weeks’ gestation and 3.7% up to 28 weeks [17]. Ghidini et al. [18] reported on a large control study of amniocentesis in 101 twin pregnan­cies with 108 control twin pregnancies. They demon­strated no significant difference in the miscarriage rates. The total fetal loss rate in the amniocentesis group was 3.5% and the fetal loss rate was similar (3.2%) in the control group. The technique used involved two-needle insertions to sample both sacs.

Chorionic villus sampling (CVS) has been demon­strated to be safe and efficacious for sampling twin gesta­tions, and has the advantage of an earlier diagnosis than amniocentesis. Genetic results can be obtained much more rapidly than with amniotic fluid cells, either in hours by direct preparations of the cytotrophoblast layer, or in 3-7 days by tissue culture of the villus mesenchymal core. It is best performed between 10 and 13 weeks’ gestation, and sampling each sac is performed either through the trans­cervical or transabdominal route under ultrasound guid­ance. Each of the approaches has advantages and disad­vantages. Technically, transcervical CVS may be more difficult to perform and the “learning curve” for trans­cervical CVS appears to involve several hundred patients. Transabdominal CVS is technically similar to mid-trimes­ter amniocentesis and has gained a quicker acceptance and application than the transcervical technique. The amount of tissue obtained with transabdominal CVS is smaller than that obtained by transcervical. Since no marker is available to assure sampling of each frondosum, continu­ous ultrasound localization of the needle tip, or the tip of the catheter, is required. If doubt exists, a follow-up proce­dure should be performed; this can either be a repeat CVS performed immediately, or a second trimester amniocente­sis.

Contamination of one sample with villi from the sec­ond sac is possible, and poses a serious potential problem, leading to failure to diagnose a chromosomal abnormality in one or both twins. Twin-twin contamination can occur, if the needle or the catheter is dragged through one frond­osum while attempting to sample the second one. Contam­ination occurs most frequently when sampling is per­formed close to the dividing membrane which contains villi from both frondosums. Twin-twin contamination can be avoided by using a combination of the transcervical and transabdominal techniques.

Since there is the possibility of one fetus having an abnormal result, careful location of the fetuses is equally as important with CVS as it is with amniocentesis. Al­though the position of the sacs will remain certain in 2-3 weeks after sampling, it is standard practice to confirm the original diagnosis in both fetal and chorionic tissues be­fore termination or selective termination.

The risk of fetal loss after CVS before 28 weeks’ gestation did not seem to differ between twin and single­ton pregnancies (4.9 vs. 4.0%, respectively). Fetuses with chromosomal aberrations were included among these losses. When only chromosomally normal pregnancies are considered, the overall loss rate of 202 twin pregnancies from five centers experienced in performing CVS in multi­ple gestations, becomes 3.7%, which is considerably less than that of amniocentesis [19]. Wapner et al. [20] re­ported that the pregnancy loss rate before 20 weeks’ gesta­tion was 3.3% following CVS, and 2.8% in contempora­neously collected group of women undergoing twin amnio­centesis. These two reports confirm that in experienced centers, CVS is an equally safe alternative to amniocente­sis for sampling twins.