International Society for Prenatal Diagnosis Position Statement: cell free (cf)<scp>DNA</scp> screening for Down syndrome in multiple pregnancies

Abstract

The aim of this Position Statement from the International Society for Prenatal Diagnosis (ISPD) is to review the relevant published literature and make evidence-based recommendations regarding screening twin and triplet pregnancies for Down syndrome via cell free (cf)DNA testing. This Position Statement should not be taken to be an endorsement that cfDNA is the optimal choice for all women with multiple pregnancies. Rather, it is one of many choices that will be faced: invasive diagnostic testing for the most comprehensive and actionable information, screening tests for common aneuploidies and/or additional disorders such as microdeletion/duplication syndromes, or to choose to have no such testing. These choices, however, are best made with access to reliable and unbiased information. Although focused on Down syndrome, trisomies 13 and 18 will also be addressed but other disorders are not, even though some cfDNA tests can identify them. The evidence will include screening performance in multifetal pregnancies for combinations of serum and ultrasound markers as well as by cfDNA testing methodologies. Current professional guidelines do not address screening multiple gestations by cfDNA testing and/or do not include the most recently published data. All compare the performance of cfDNA in twin pregnancies to that reported for cfDNA screening in singleton pregnancies. In contrast, this Position Statement will compare cfDNA testing to other screening methods available for multifetal pregnancies focusing on test characteristics such as the detection rate (sensitivity), false positive rate (1-specificity), and the test failure rate. Attention is also paid to the challenges of diagnostic procedures in multifetal pregnancies and timing of selective reduction. This review was undertaken to inform laboratories offering such testing, clinicians offering tests and receiving results, policy-makers updating their recommendations and, most importantly, informing couples pregnant with a multiple gestation. In 1895, Dr. Dionys Hellin published the natural rate of twin and triplet maternities as 1 in 89 and 1 in 89 × 89 (112 and 1.3 per 10,000 maternities, respectively). These estimates have been confirmed as being reasonable and can serve as a baseline rate.1, 2 Among dizygotic twin pregnancies, a well described positive association occurs with advancing maternal age.3 In contrast, the rate of monozygotic twinning is relatively constant at 35 per 10,000, regardless of race, geography, or maternal age.3 The natural proportions of twins are about one-third monozygotic and two-thirds dizygotic. The lowest rates of multifetal deliveries occur in Asia with 30-40 per 10,000. Much higher rates of 180 per 10,000 are reported in central Africa.4 Some suggest the rate of Down syndrome in twin term deliveries (one or both affected) may be higher than in singleton pregnancies, after accounting for zygosity.5, 6 However, observational data suggest the ratio may actually be close to one.7 Modeling in this current report assumes the rates for the common trisomies to be the same in twin (one or both affected) and singleton deliveries for a given maternal age. Figure 1 shows the change in twin and triplet liveborn delivery rates in Australia, the Netherlands and the United States from 1980 to 2018. These mirror the increasingly higher rates of multiple gestations around the world8-11 and are likely due to several factors. Both Black/African Americans and Hispanic mothers are more likely to have spontaneous multiple gestations and these groups may now represent a greater proportion of the population. Older women are also more likely to have spontaneous multiple pregnancies and, in many countries, the average maternal age at delivery is rising. Maternal age may account for about one-third of the increase seen in the US.12 The factor responsible for most of the remaining increase was the introduction of in vitro fertilization (IVF) in the 1980s.13 Perhaps half of these services are provided to women age 35 years and older. The more recent decline in twin and triplet pregnancies may reflect a temporal change in IVF practices relating to single embryo transfer guidelines.14 The conception rate for twins is greater than the birth rate as the disappearance of gestational sacs or embryos after documented heart activity is not unusual. This is known as co-twin demise and occurs in between 30% and 40% of twin sacs or embryos.15 The precise mechanisms and pathophysiology are obscure but the remaining placenta may still be functional for some time. When death occurs beyond the first trimester,16 the survivor has a higher risk for low birth weight and small for gestational age.17-19 The demised or vanished twin can explain discrepancies between cfDNA test results and the fetal karyotype.20, 21 Beginning in the 1980s, IVF involving multiple embryo transfers was routine due to low implantation rates.14, 22 In the last decade, a decrease in the rates of twins and higher-order multiple deliveries occurred, although still remains higher than would naturally occur (Figure 1). The 2016 guideline from the American Society of Reproductive Medicine and the Society of Assisted Reproductive Technology recommends decreasing the number of embryos transferred per cycle.14 Other factors also decreasing the rate of multiple pregnancies include promotion of single embryo transfers due to the technological advances of culturing embryos to the blastocyst stage and ensuring higher per embryo implantation success. Estimates suggest that more multiple gestations now stem from ovulation induction/intrauterine insemination cycles than from IVF.23 Finally, another contribution to the decrease is the increase in multifetal pregnancy reductions being performed.24 Twin pregnancies conceived with IVF are more than 95% dizygotic, higher than the two-thirds expected in those naturally conceived.25 Dizygotic twins are also more common in pregnancies conceived with ovulation inducing agents alone (without IVF) since these drugs increase the likelihood of ovulation and fertilization of multiple oocytes. Prenatal diagnostic testing remains the gold standard for obtaining genetic information about the pregnancy. Either chorionic villus sampling (CVS) or amniocentesis can be offered but there are special considerations compared to singleton pregnancy. Whether the pregnancy is monochorionic or dichorionic is optimally determined by ultrasound at 9-12 completed weeks.26 When a patient with a multifetal pregnancy chooses a diagnostic procedure, each fetus should be identified and its placental position documented. This assures each fetus is sampled only once and that the results can be accurately attributed to the correct fetus following prenatal diagnosis 1-2 weeks later. Ideally, the specialist who performed the ultrasound and mapping should perform the diagnostic procedure. Monochorionicity indicates monozygosity and this implies that the two fetuses will have identical genomes. Under this assumption, a single transabdominal needle entry or single transcervical catheter entry can be performed, decreasing the risk of pregnancy loss. However, monochorionic pregnancies with discordant fetal anomalies due to post-zygotic events such as non-disjunction and twinning errors do occur.27-32 For this reason, amniocentesis may be the more appropriate procedure to assure each fetus is sampled independently. A maternal fetal medicine or other specialist provider who performs diagnostic procedures frequently in multiple pregnancies can minimize the chance of a sampling error. About 90% of dichorionic twin pregnancies are dizygotic with the remainder monozygotic. For dichorionic pregnancies, both placentas should be sampled, typically between 11 and 13 completed weeks. CVS via transabdominal or transcervical approaches can be used, including a combination of both. The risk of CVS cross-contamination (sampling the same fetus twice) is approximately 1%.33 If there is a karyotype abnormality in one fetus, this allows time for a selective reduction. In triplet pregnancies it is important to perform an 11 to 13 week ultrasound to determine chorionicity and map out the location of the fetuses to their respective placentas to allow for successful diagnostic testing. Because of the technical expertise needed to perform CVS in triplet pregnancies and the possibility of cross contamination,34 a perinatologist or other specialist experienced in such procedures is preferred. Data on zygosity in triplets conceived spontaneously vs via ART remain limited. Among ART triplets, MZ twin pairs were markedly less likely than among spontaneous triplets (6.5% vs 48%, respectively).35 Selective termination is defined as the termination of an anomalous fetus in a multifetal gestation. These procedures are generally performed in the late first or early second trimester after one (or more) fetuses have been diagnosed. There are various procedures that can be used depending on the provider's preference, the gestational age, and zygosity. A recent report36 of one tertiary referral center's experience with dichorionic diamniotic twin reductions to singleton reductions compares outcomes between procedures performed in the late first trimester (12-14 weeks) with those performed in the early second trimester (16-20 weeks). The earlier reductions (N = 172) were often due to patient's request or abnormal ultrasound findings while the latter group (N = 76) were mainly due to structural or genetic abnormalities. Fetal loss before 24 weeks was similar (0.6% early and 1.3% late, P = .52) but with later reductions, neonatal morbidity (2.9% vs 10.7%, P = .025), and delivery prior to 35 weeks gestation (1.8% vs 12%, P = .002) were higher and birthweight was lower (2800 vs 3025 g, P = .012). There are limited data regarding selective termination in monozygotic twins. One study37 did provide results of selective termination in complicated monochorionic twin (N = 73) and triplet (N = 7) pregnancies undergoing cord coagulations at two tertiary fetal medicine centers. The gestational ages at procedure ranged from 15 to 29 weeks, likely later than the gestational age range for a similar procedure after diagnosing a common aneuploidy via cfDNA screening. Loss rates of twin and triplet pregnancies were 16% and 21%, respectively. Overall, 79% of deliveries occurred after 32 weeks. One year follow-up found developmental delays in 8%, mainly in the group delivering prior to 29 weeks. The use of only maternal age to screen for Down syndrome (or other common trisomies) is easy and inexpensive but has relatively poor performance. For example, using age 37 years or older at delivery in the 2018 US population, 59% of affected twin and 66% of affected triplet pregnancies would be identified with false positive rates of 15% and 19%, respectively. As with singleton pregnancies, approaches to modifying the maternal age alone risk for Down syndrome also exist for multifetal pregnancies. One summary of published studies reviewed serum screening occurring in the first, second, or both trimesters in twin pregnancies (Figure 2).38 With second trimester analytes, a trade-off occurs: a similar or higher detection rate, but at lower false positive rates of 5%-12% compared with age alone. First trimester incorporation of nuchal translucency (NT) with maternal age increases detection to 93%, but with a 10% rate of offering diagnostic testing. NT alone offers 80% detection at a reasonable 5% false positive rate. Notably, adding the first trimester serum markers does not improve screening appreciably in multifetal pregnancies as they are pregnancy-specific rather than fetal-specific. The integrated test is the most complex testing methodology but does allow for up to a 93% detection rate at a 5% diagnostic testing rate in twin pregnancies and many of those affected fetuses can be detected in the first trimester. The reported performance of cfDNA screening for the common trisomies in twins is also included (Figure 2, upper left) but is discussed in a later section. Less information is available for screening twin pregnancies for trisomies 18 and 13, but both can be best identified using first trimester NT measurements, biochemical measurements, and maternal age.39, 40 The use of cfDNA screening for Down syndrome in multifetal pregnancies has received increasing attention in recent years. Table 1 provides a brief summary of published recommendations from 10 professional societies.38, 41-49 The search was limited to those published in 2015 or later. The majority of recommendations published prior to 2018 tended to not address this issue. Many (4 of 10) recommend further investigation (2016-2018).38, 42, 44, 48 Others were silent on the issue (3 of 10)43, 45, 47 or were opposed to testing in twins (1 of 10).46 One other recommends screening for Down syndrome in twins but not for trisomies 18 or 13 due to lack of data.41 The earliest recommendation, published in 2015, implied screening was acceptable for all three disorders by stating performance in twins was similar to that in singleton pregnancies. Only two recommendations directly addressed triplet pregnancies. Both recommended the use of first trimester ultrasound markers at 11-13 weeks gestation and did not address cfDNA testing.42, 49 Since this document was drafted, the pre-publication of a new ACOG Practice Bulletin was released.50 That document makes a level B recommendation that “cell-free DNA screening can be performed in twin pregnancies”. In addition, England recently released a plan to employ cfDNA testing in both singleton and twin pregnancies.51 Maternal plasma cfDNA screening tests measure the distribution of nonmembrane bound DNA fragments derived from the various human chromosomes. A fetal chromosomal aneuploidy is suspected when the amount (percentage) of cfDNA fragments from a particular chromosome differs from the amount expected for a euploid karyotype. The test rationale behind fetal trisomy detection is based on the assumption that the mother is euploid. Hence, any deviations in the chromosomal DNA amounts are first assumed to be of fetal origin and thus warrant further definitive diagnostic testing. Several cfDNA test methodologies are available for aneuploidy screening and they are generally categorized as whole-genome or targeted approaches.52 Whole-genome methods are based on massively parallel sequencing of large numbers of randomly captured cfDNA fragments to identify and then count the cfDNA contributions from each human chromosome.53 Targeted approaches selectively analyze cfDNA from the chromosomes of interest (eg, chromosomes 21, 18, 13, X, and Y). The cfDNA fragments from those chromosomes are first amplified and then identified by sequencing or microarray.54 After whole-genome or targeted DNA data are generated, statistical analyses are performed to determine if the amount of DNA from one chromosome is relatively or compared with the expected amounts based on chromosome In one of the targeted DNA on the chromosomes of interest are selectively between are determined for many single is suspected when the of one chromosome from the of the other When using this is needed for the between the fetus and pregnant may be such as in pregnancies involving or include testing by whole-genome as well as targeted The detection and false positive rates for trisomies 21, 18, and 13 among twin pregnancies between methodologies. There however, data to compare the performance between and twins and the is in those twins each fetus is test should that the of each fetus is actually and to this of fetal would be an important A review of the literature regarding the performance of cfDNA testing in twin pregnancies was published in That review included only studies and did not include additional studies that were of or were published after that review was Table 2 is a of results from both the and the additional For example, results from the shows a of twin pregnancies with at one fetus with Down Among the there was one false two false positive results also The results for the additional studies include 40 additional affected twin pregnancies with a Down syndrome The results from the two groups of were with a of twin pregnancies with were monochorionic and of affected twin pregnancies an aneuploidy in both fetuses A of twin pregnancies at one fetus with a common trisomy Down syndrome, 29 trisomy 18 and trisomy with detection rates of and respectively. The false positive rate among pregnancies was These chromosome detection and false positive rates are also in the upper of Figure The late first trimester for these three disorders in twins are to be and as for singleton and the false positive rates are to for each The positive for successful cfDNA testing in twins would then be approximately and 19%, respectively. The lower for trisomies 18 and 13 are mainly due to their lower These are expected to be similar to those found in screen positive singleton pregnancies from a population. One was identified as this review was being completed and was not included in Table This US group cfDNA testing results with results performed in the same In such a all false results would be Overall, detection rates for trisomy 21, 18, and 13 were at and with the findings in Table However, the false positive rate of is higher than those in Table The of cfDNA in the maternal by the fetal is the fetal In a low fetal is with test or false results for trisomies 18 and 13, but less for Down cfDNA screening tests the fetal and a as a When for singleton pregnancies range from to These would also to twins. For it would be to measure or the The fetal between each fetus of a are = but still by or more in 10% of The average fetal for each twin is lower than that for the fetal for twins is only The of trisomy 13 or to have even lower fetal further first of second rates of second are in first second Twin with lower to to first second first second MZ to to to first second One to Both to first second twins triplets to to low MZ first second twins triplets low low In multifetal pregnancies, there is a range of methodologies used for fetal and reported may reflect the pregnancy or chromosome cfDNA is and when both twin fetuses are the is the fetal When only one fetus is the is to that fetus of placental DNA of and cfDNA on all can provide the fetal regardless of fetal and number of Finally, the ratio between to maternal provides estimates of the fetal fetus in a will have while they will be identical for twins fetal For can provide a fetal for each fetus and can also determine one or more of the fetal methods have been used in combination with whole-genome or targeted cfDNA aneuploidy screening tests some use a fetal to determine other methods are also In multifetal pregnancies, this is on fetal or fetal is and the twins are likely or likely For twins and fetal of the same fetal as for singleton pregnancies are often that measure fetal in twins often that of singleton cfDNA at to a These results are to as or and are often due to fetal or other such as or other technical After a test a be or a second The test will a of such The no rate for cfDNA aneuploidy screening among twin pregnancies from to with a of fetal is the provided failure rates in twin pregnancies when a second was offered to of and of those women provided a second rates in the ranged from to with a of The failure rate was from an to a included with singleton pregnancies and the rates in twins were and higher than in analyses identified maternal weight and IVF as important of test failure If the pregnancy is still the 11 to 13 week obtaining an NT or may be pregnancies are Among 10,000 about twin and three triplet pregnancies would the first trimester of Down syndrome in triplet pregnancies was as for then 10 would occur among triplet pregnancies. that number of would a of about 11 However, even this not be as current professional guidelines recommend or are silent cfDNA testing in this even though numbers of were available in the literature 1). a majority of triplet pregnancies are now due to IVF and some may have genetic testing for in a lower than expected rate of studies reported cfDNA test results in more than 10 triplet but were reported to be screen positive or from a known pregnancy. detection rates for Down syndrome and other common trisomies are failure rates are likely to be higher than for twins and Table when fetal estimates are not However, there are data that cfDNA testing can identify the fetal in all triplet combinations of on current of cfDNA testing the between mother and fetus, one this from screening and twins to a performance in triplet pregnancies. each fetus in a triplet pregnancy cfDNA to in a fetal (eg, at per this can be as the fetus (or and the mother all two of chromosome 21, while the affected fetus has an additional for the screening performance should that found for twin pregnancies when testing is and were identified from laboratories and two In some relevant were identified via a search When this was not the was a and and for In one there was an of directly relevant to twin pregnancies on the while for the it was to or not in a that would be by an average Others that the test was available for both singleton and twin pregnancies with or no information. was to be by a rather than an women with a false of addressed relating to diagnostic testing or the of pregnancy in multiple pregnancies and provided estimates of screening performance or test failure screening laboratories have provided more both of laboratories to for more information. were to at an level or that that of to couples with multifetal pregnancies were being addressed in any of these Ideally, laboratories would provide of information for twin pregnancies as for singleton pregnancies including they prior and provide false and failure rates with positive should also include is known and not known about screening in twins and higher for couples to make have a to that may be as a of the screening and make couples prior to their to be via couples will naturally there to be no in performance or in follow-up testing and The is not responsible for the or of any information by the than should be to the for the

Topics & Concepts