KH Nicolaides, NJ Sebire, RJM Snijders
above is an extract from the paper ‘Observations on an ethnic classification
of idiots’ by Langdon Down, published in 18661. Down, who was
a physician at the London Hospital, coined the phrase Mongolian idiots
because he felt that a subgroup of his patients had a resemblance to the
Mongolian peoples and this fitted in with his theory of ‘retrogression’
of ethnic type. Down’s theory of ethnic regression was in keeping with
Darwin’s contemporary scientific reasoning for evolution. In 1924, Crookshank
suggested that the regression was not merely to a primitive Oriental human
type but also to the orangutan2. Even though the theory
of ethnic regression was proven to be inaccurate, Down’s description
of the appearance of the skin was the basis for the observation, made
more than one century later, that affected individuals during the 3rd
month of intrauterine life, have a subcutaneous collection of fluid behind
the neck (Figure 1), which can be visualized by ultrasound
as nuchal translucency (Figure 2).
Langdon Down in 1866 and Fraser and Mitchell in 1876 recognized that the condition was congenital, dating from intrauterine life, and in 1914 Goddard found that there was no increased incidence within families1,3,4. A number of conditions were advocated as potential causes of Down’s syndrome, including syphilis, tuberculosis, parental alcoholism, epilepsy, insanity, nervous instability and mental retardation in a close relative, thyroid deficiency, hypoplasia of the fetal adrenal glands, dysfunction of the fetal pituitary and abnormality of the fetal thymus1,6–13.
The association between Down’s syndrome and increased maternal age was noted in 1909 by Shuttleworth6, who examined 350 cases and reported that:
As a result of the above observation, hypotheses were based upon theoretical degeneration of the ovum14–16. However, advanced maternal age could not be the only factor, because, in some cases, there appeared to be a hereditary factor as well. For instance, dizygotic twins were unequally affected whereas monozygotic twins were equally affected17. It was also noticed that the condition could be transmitted from mother to baby, and, when more than one member of a family was affected, the dependence on the mother’s age was weakened18–21. The concept of non-dysjunction in Down’s syndrome was suggested by Waardenburg in 193222. In 1934, Bleyer proposed that an unequal migration of chromosomes during cell division may result in trisomy16.
In 1956, Tjio and Levan, working with improved techniques on cultures of lung fibroblasts, established that the normal diploid chromosome number is 4623. In the same year, Ford and Hamerton found that the haploid number was 23 in human spermatocytes24. These discoveries led a number of laboratories to study the karyotype in various pathological conditions and in 1959 Lejeune et al. and Jacobs et al. demonstrated that an extra acrocentric chromosome was present in persons with Down’s syndrome, resulting in an aneuploid chromosome number of 4725,26.
There were familial cases of Down’s syndrome which were not the result of trisomy. In 1960 Polani et al. examined the chromosomes of a child with Down’s syndrome from a 21-year-old mother, there were 46 chromosomes with a centric fusion of two chromosomes (15:21)27. Familial transmission of this type of translocation was demonstrated by Penrose et al. in 1960 in a family with two Down’s syndrome sibs28. In 1961, Clarke et al. reported on a 2-year-old girl with normal intelligence but some physical features suggestive of Down’s syndrome; she was discovered to be a mosaic for normal and trisomic cells29.
Today we know that Down’s syndrome occurs when either the whole or a segment of the long arm of chromosome 21 is present in three copies instead of two. This can occur as a result of three separate mechanisms: non-dysjunction, found in about 95% of cases, translocation and mosaicism. In 1991, Antonarakis et al. examined DNA polymorphisms in Down’s syndrome infants and demonstrated that 95% of non-dysjunction trisomy 21 is maternal in origin30. The region that codes for most of the Down’s syndrome phenotype is the distal portion of band q21.1 and bands q22.2 and q22.3. This region determines the facial features, heart defects, mental retardation and probably the dermatoglyphic changes in affected individuals31.
In 1966, 100 years after the original essay of Langdon Down, it became possible to diagnose trisomy 21 prenatally by karyotyping of cultured amniotic fluid cells32,33.
The first method of screening for trisomy 21, introduced in the early 1970s, was based on the observation of Shuttleworth on the association with advanced maternal age6. It was apparent that amniocentesis carried a risk of miscarriage and this in conjunction with the cost implications, meant that prenatal diagnosis could not be offered to the entire pregnant population. Consequently, amniocentesis was initially offered only to women with a minimum age of 40 years. Gradually, as the application of amniocentesis became more widespread and it appeared to be ‘safe’, the ‘high-risk’ group was redefined to include women with a minimum age of 35 years; this ‘high-risk’ group constituted 5% of the pregnant population.
In the last 20 years, two dogmatic policies have emerged in terms of screening. The first, mainly observed in countries with private healthcare systems, adhered to the dogma of the 35 years of age or equivalent risk; since the maternal age of pregnant women has increased in most developed countries, the screen-positive group now constitute about 10% of pregnancies. The second policy, instituted in countries with national health systems, has adhered to the dogma of offering invasive testing to the 5% group of women with the highest risk; in the last 20 years, the cut-off age for invasive testing has therefore increased from 35 to 37 years. In screening by maternal age with a cut-off age of 37 years, 5% of the population are classified as ‘high risk’ and this group contains about 30% of trisomy 21 babies.
In the late 1980s, a new method of screening was introduced that takes into account not only maternal age but also the concentration of various fetoplacental products in the maternal circulation. At 16 weeks of gestation the median maternal serum concentrations of a-fetoprotein, estriol and human chorionic gonadotropin (hCG) (total and free-b) in trisomy 21 pregnancies are sufficiently different from normal to allow the use of combinations of some or all of these substances to select a ‘high-risk’ group. This method of screening is more effective than maternal age alone and, for the same rate of invasive testing (about 5%), it can identify about 60% of the fetuses with trisomy 21.
In the 1990s, screening by a combination of maternal age and fetal nuchal translucency thickness at 11–14 weeks of gestation was introduced. This method has now been shown to identify about 75% of affected fetuses for a screen-positive rate of about 5%.
evidence suggests that maternal age can be combined with fetal nuchal
translucency and maternal serum biochemistry (free b-hCG and pregnancy-associated
plasma protein (PAPP-A)) at 11–14 weeks to identify about 90% of affected
fetuses. Furthermore, the development of new methods of biochemical testing,
within 30min of taking a blood sample, has now made it possible to introduce
One-Stop Clinics for Assessment of Risk (Figure 3).
|CALCULATION OF RISK FOR CHROMOSOMAL DEFECTS|
Every woman has a risk that her fetus/baby has a chromosomal defect. In order to calculate the individual risk, it is necessary to take into account the background risk (which depends on maternal age, gestation and previous history of chromosomal defects) and multiply this by a series of factors, which depend on the results of a series of screening tests carried out during the course of the pregnancy. Every time a test is carried out the background risk is multiplied by the test factor to calculate a new risk, which then becomes the background risk for the next test34. This process is called sequential screening. With the introduction of OSCAR, this can all be achieved in one session at about 12 weeks of pregnancy (Figure 3).
The risk for many of the chromosomal defects increases with maternal age (Figure 4). Additionally, because fetuses with chromosomal defects are more likely to die in utero than normal fetuses, the risk decreases with gestational age (Figure 5).
Estimates of the maternal age-related risk for trisomy 21 at birth are based on two surveys with almost complete ascertainment of the affected patients; in a survey in South Belgium, every neonate was examined for features of trisomy 21 and, in a survey in Sweden, information was verified using several sources such as hospital notes, cytogenetic laboratories, genetic clinics and schools for the mentally handicapped35,36. The data from these surveys were used to calculate maternal age-specific incidences of trisomy 21 at birth37.
During the last decade, with the introduction of maternal serum biochemistry and ultrasound screening for chromosomal defects at different stages of pregnancy, it has become necessary to establish maternal age and gestational age-specific risks for chromosomal defects. Such estimates were derived by comparing the birth prevalence of trisomy 2137 to the prevalence in women undergoing second-trimester amniocentesis or first-trimester chorionic villus sampling. Rates of spontaneous fetal death between different gestations and delivery at 40 weeks were estimated on the basis of both the observed prevalence in pregnancies that had antenatal fetal karyotyping and the reported prevalence in live births.
et al. examined the prevalence of trisomy 21 in 57614 women who
had fetal karyotyping at 9–16 weeks of gestation for the sole indication
of maternal age of 35 years or more; this group included 538 pregnancies
with trisomy 2138–40. They found that the prevalence of trisomy
21 was higher in early pregnancy than in live births and the estimated
rates of fetal loss were 36% from 10 weeks, 30% from 12 weeks, and 21%
from 16 weeks38. The estimated maternal age and gestational
age-related risks for trisomy 21 are given in Table 1.
In a similar study, Halliday et al. compared the prevalence of trisomy 21 in 10545 women having chorionic villus sampling or amniocentesis to the prevalence in live births from 12921 women of similar age who did not have fetal karyotyping41. Their estimated fetal loss rate between 10 weeks and term was 31% and between 16 weeks and term was 18%. Morris et al. examined outcome data from 4148 trisomy 21 pregnancies reported to the National Down Syndrome Cytogenetic Register in the UK with correction for elective terminations. Their study population included 441 cases diagnosed at 11–13 weeks of gestation and 2035 cases diagnosed at 16–18 weeks; they estimated that the loss rates between 12 and 16 weeks and term were 31% and 24%, respectively42. These estimates for spontaneous loss between the first trimester and term are lower than the 48% reported by Mackintosh et al. who compared the prevalence of trisomy 21 at chorionic villus sampling and birth; the most likely explanation for this high rate (48%), compared to rates derived in the other reports (31%), is that the study included a substantial proportion of cases in which chorionic villus sampling was performed before 10 weeks of gestation43.
Similar methods were used to produce estimates of risks for other chromosomal abnormalities40. The risk for trisomy 18 andtrisomy 13 increases with maternal age and decreases with gestation; the rate of intrauterine lethality between 12 weeks and 40 weeks is about 80% (Table 2 and Table 3). Turner syndrome is usually due to loss of the paternal X chromosome and, consequently, the frequency of conception of 45,X embryos, unlike that of trisomies, is unrelated to maternal age. The prevalence is about 1 per 1500 at 12 weeks, 1 per 3000 at 20 weeks and 1 per 4000 at 40 weeks. For the other sex chromosome abnormalities (47,XXX, 47,XXY and 47,XYY), there is no significant change with maternal age and since the rate of intrauterine lethality is not higher than in chromosomally normal fetuses the overall prevalence (about 1 per 500) does not decrease with gestation. Polyploidy affects about 2% of recognized conceptions but it is highly lethal and thus very rarely observed in live births; the prevalences at 12 and 20 weeks are about 1 per 2000 and 1 per 250000, respectively40.
Creating the model for calculation of the maternal and gestational age-specific risks made it possible to counsel patients presenting at different stages of pregnancy about the risk for their fetus having a chromosomal defect and the chance that the pregnancy will result in a live birth with a specific condition. Furthermore, these data can be applied in the evaluation of new ultrasonographic or biochemical methods of screening by calculating the expected prevalence of chromosomal defects in any study group.
The risk for trisomies in women who have had a previous fetus or child with a trisomy is higher than the one expected on the basis of their age alone. In a study of 2054 women who had a previous pregnancy with trisomy 21, the risk of recurrence in the subsequent pregnancy was 0.75% higher than the maternal and gestational age-related risk for trisomy 21 at the time of testing. In 750 women who had a previous pregnancy with trisomy 18, the risk of recurrence in the subsequent pregnancy was also about 0.75% higher than the maternal and gestational age-related risk for trisomy 18; the risk for trisomy 21 was not increased44. Thus, for a woman aged 35 years who has had a previous baby with trisomy 21, the risk at 12 weeks of gestation increases from 1 in 249 (0.40%) to 1 in 87 (1.15%), and, for a woman aged 25 years, it increases from 1 in 946 (0.106%) to 1 in 117 (0.856%).
The possible mechanism for this increased risk is that a small proportion (less than 5%) of couples with a previously affected pregnancy have parental mosaicism or a genetic defect that interferes with the normal process of dysjunction, so in this group the risk of recurrence is increased substantially. In the majority of couples (more than 95%), the risk of recurrence is not actually increased. Currently available evidence suggests that recurrence is chromosome-specific and, therefore, in the majority of cases, the likely mechanism is parental mosaicism.
The nuchal translucency normally increases with gestation (crown–rump length). In a fetus with a given crown–rump length, every nuchal translucency measurement represents a factor which is multiplied by the background risk to calculate a new risk. The larger the nuchal translucency, the higher the multiplying factor becomes and therefore the higher the new risk. In contrast, the smaller the nuchal translucency measurement, the smaller the multiplying factor becomes and therefore the lower the new risk (Figure 6).
calculate the multiplying factor (likelihood ratio), it is first necessary
to determine the distributions of nuchal translucency thickness in the
chromosomally normal and trisomy 21 groups. For a given nuchal translucency,
a likelihood ratio is calculated by dividing the percentage of trisomy
21 fetuses by the percentage of normal fetuses with that translucency.
The combined risk is then calculated by multiplying the background maternal
and gestational age-related risk by the likelihood ratio.
level of free b-hCG in maternal blood normally decreases with gestation.
The higher the b-hCG, the higher the risk for trisomy 21. Again, for a
given gestation, each hCG level represents a factor that is multiplied
by the background risk to calculate the new risk (Figure 7). The level of PAPP-A in maternal blood normally
increases with gestation. The lower the PAPP-A, the higher the risk for
trisomy 21. Again, for a given gestation, each PAPP-A level represents
a factor that is multiplied by the background risk to calculate the new
risk (Figure 7).
During the second and third trimesters of pregnancy, abnormal accumulation of fluid behind the fetal neck can be classified as nuchal cystic hygroma or nuchal edema. In about 75% of fetuses with cystic hygromas, there is a chromosomal abnormality and, in about 95% of cases, the abnormality is Turner syndrome45. Nuchal edema has a diverse etiology; chromosomal abnormalities are found in about one-third of the fetuses and, in about 75% of cases, the abnormality is trisomy 21 or 18. Edema is also associated with fetal cardiovascular and pulmonary defects, skeletal dysplasias, congenital infection and metabolic and hematological disorders; consequently, the prognosis for chromosomally normal fetuses with nuchal edema is poor46.
In the first trimester, the term translucency is used, because this is the ultrasonographic feature that is observed; during the second trimester, the translucency usually resolves and, in a few cases, it evolves into either nuchal edema or cystic hygromas with or without generalized hydrops.
Nuchal translucency can be measured successfully by transabdominal ultrasound examination in about 95% of cases; in the others, it is necessary to perform transvaginal sonography. The equipment must be of good quality (about £30000–50000), it should have a video-loop function and thecalipers should be able to provide measurements to one decimal point. The average time allocated for each fetal scan should be at least 10 minutes. All sonographers performing fetal scans should be capable of reliably measuring the crown–rump length and obtaining a proper sagittal view of the fetal spine. For such sonographers, it is easy to acquire, within a few hours, the skill to measure nuchal translucency thickness. Furthermore, it is essential that the same criteria are used to achieve uniformity of results among different operators (Figure 8):
Care must be taken to distinguish between fetal skin and amnion because, at this gestation, both structures appear as thin membranes. This is achieved by waiting for spontaneous fetal movement away from the amniotic membrane; alternatively, the fetus is bounced off the amnion by asking the mother to cough and/or by tapping the maternal abdomen.
The maximum thickness of the subcutaneous translucency between the skin and the soft tissue overlying the cervical spine should be measured by placing the
The nuchal translucency should be measured with the fetus in the neutral position. When the fetal neck is hyperextended the measurement can be increased by 0.6mm and when the neck is flexed, the measurement can be decreased by 0.4mm50.
umbilical cord may be round the fetal neck in 5–10% of cases and this
finding may produce a falsely increased nuchal translucency, adding
about 0.8mm to the measurement51. In such cases, the measurements
of nuchal translucency above and below the cord are different and,
in the calculation of risk, it is more appropriate to use the smaller
The distribution of nuchal translucency measurements as well as the quality of the images in terms of magnification, section (sagittal or oblique),caliper placement, skin line (nuchal only or nuchal and back) and visualization of the amnion separate from the nuchal membrane are taken into account in the audit of results52.
The ability to measure nuchal translucency and obtain reproducible results improves with training; good results are achieved after 80 and 100 scans for the transabdominal and the transvaginal routes, respectively53.
The ability to achieve a reliable measurement of nuchal translucency is dependent on the motivation of the sonographer. A study comparing the results obtained from hospitals where nuchal translucency was used in clinical practice (interventional) compared to those from hospitals where they merely recorded the measurements but did not act on the results (observational), reported that, in the interventional group, successful measurement of nuchal translucency was achieved in 100% of cases and the measurement was > 2.5mm in 2.3% of cases; the respective percentages in the observational group were 85% and 12%54,55.
Appropriate training, high motivation and adherence to a standard technique for the measurement of nuchal translucency are essential prerequisites for good clinical practice. Monni et al. reported that, after modifying their technique of measuring nuchal translucency thickness, by following the guidelines established by The Fetal Medicine Foundation, their detection rate of trisomy 21 improved from 30% to 84%56.
A potential criticism of screening by ultrasound is that scanning not only requires highly skilled operators but it is also prone to operator variability. This issue was addressed by a prospective study at 10–14 weeks of gestation in which the nuchal translucency was measured by two of four operators in 200 pregnant women57. This study demonstrated that, after an initial measurement, the second one made by the same (intra-) observer or another (inter-) observer varies from the first by less than 0.54mm and 0.62mm, respectively in 95% of the cases. Additionally, the study demonstrated that thecaliper placement repeatability was similar to the intra-observer and inter-observer repeatability, suggesting that a large part of the variation in measurements can be accounted for by the placement of the calipers rather than the generation of the image57. Subsequent studies have reported that the intra-observer and inter-observer differences in measurements were less than 0.5mm in 95% of cases58,59.
Digital image processing and automation ofcaliper placement may reduce the variation of measurements60. In the meantime, it is best to rely on the mean of two good measurements for the calculation of risk, rather than on a single one.
Fetal nuchal translucency thickness increases with crown–rump length49,61, and therefore it is essential to take gestation into account when determining whether a given translucency thickness is increased. In a study involving more than 100000 pregnancies, the median increased from 1.2mm at 11 weeks to 1.9mm at 13+6 weeks62. Figure 9 illustrates the increases in the 5th, 25th, 50th, 75th and 95thcentile of nuchal translucency with crown–rump length; the 99th centile is about 3.5mm throughout this gestational range.
In the early 1990s, several reports of small series in high-risk pregnancies demonstrated a possible association between increased nuchal translucency and chromosomal defects in the first trimester of pregnancy (Table 4)63–80. Although the mean prevalence of chromosomal defects in 20 series involving a total of 1698 patients was 29%, there were large differences between the studies with the prevalence ranging from 11% to 88%. This variation in results presumably reflects differences in the maternal age distributions of the populations examined as well as in the definition of the minimum abnormal translucency thickness, which ranged from 2mm to 10mm.
a series of screening studies in high-risk pregnancies were carried out;
these involved measurement of nuchal translucency thickness immediately
before fetal karyotyping, mainly for advanced maternal age. Pandya et al.
examined a total of 1273 pregnancies and reported that the nuchal translucency
thickness was above the 95th centile of the normal range in about 80%
of trisomy 21 fetuses81. Similar findings were obtained in
an additional four studies of pregnancies undergoing first-trimester fetal
karyotyping 73,74,76,78. However, in another study involving
1819 pregnancies, nuchal translucency thickness of equal to or greater
than 3mm identified only 30% of the chromosomally abnormal fetuses (no
data were provided specifically for trisomy 21) and the false-positive
rate was 3.2%72.
An important finding of the screening studies in high-risk pregnancies was that the prevalence of chromosomal defects is dependent on both fetal nuchal translucency thickness and maternal age. For example, in a study of 1015 pregnancies with increased fetal nuchal translucency thickness at 10–14 weeks of gestation, the observed numbers of trisomies 21, 18 and 13 in fetuses with translucencies of 3mm, 4mm, 5mm and > 6mm were approximately 3 times, 18 times, 28 times and 36 times higher than the respective number expected on the basis of maternal age67. The incidences of Turner syndrome and triploidy were 9 times and 8 times higher, whilst the incidence of other sex chromosome aneuploidies was similar to that expected67.
There are nine studies that have examined the implementation of nuchal translucency screening in routine practice and the results are summarized in Table 554,74,82–88. The number of trisomy 21 pregnancies in all but one86 of these studies is too small to allow assessment of the sensitivity of the test. However, these studies demonstrate a series of important points:
The Frimley Park Hospital, Camberley and St. Peter’s Hospital, Chertsey, UK82
Frimley Park and St. Peter’s are general hospitals within the NHS offering routine antenatal care, and their combined annual number of deliveries is approximately 6000. Prior to the introduction of nuchal translucency scanning, the policy of these hospitals was to offer amniocentesis to women aged 35 years or older. During 1993 there were 11 fetuses with Down’s syndrome and only two of these were detected prenatally. Subsequently, nuchal translucency screening at 10–14 weeks of gestation was introduced and the implementation of this policy was achieved without the need for increasing the number of staff or the equipment. Women with fetal translucency of 2.5mm or more were offered fetal karyotyping. In addition women aged 35 years or older were offered amniocentesis at 16 weeks’ gestation. The data of the first 5 months after the introduction of the new policy were analyzed following completion of the pregnancies. During this period, 74% of women delivering in the two hospitals attended for first-trimester scanning and the nuchal translucency was successfully measured in all pregnancies. The nuchal translucency was raised in 3.6% of cases and the total percentage of invasive procedures was 5.1%. All four cases of Down’s syndrome that occurred in this period were diagnosed prenatally82.
University College Hospital, London, UK83
In a screening study of 1704 women with singleton pregnancies attending University College Hospital, London, for routine antenatal care at 8–14 weeks of gestation, transabdominal ultrasound examination was performed. In 20% of cases, the sonographers forgot to measure the nuchal translucency thickness. In a further 18% of those women in whom a measurement was attempted, this was unsuccessful. In 28% of the 1127 cases in whom measurements were made, the scans were carried out before 10 weeks of gestation. The nuchal translucency was equal to or greater than 3mm in 6% of the cases. The population contained three fetuses with trisomy 21, all in women aged equal to or greater than 39 years, and increased nuchal translucency was found in one83.
Queen Charlotte’s and Guy’s Hospitals, London, UK54
This report combined the data from two centers; in one the study was observational and in the other it was interventional. The nuchal translucency was equal to or greater than 3mm in four (50%) of the eight trisomy 21 pregnancies. In the interventional center, 969 pregnancies were examined, the nuchal translucency was successfully measured in all cases and the translucency was equal to or greater than 3mm in 20 (2.0%) of the 966 chromosomally normal pregnancies. In contrast, in the observational center, 512 pregnancies were examined, the nuchal translucency was successfully measured in 470 (92%) of cases and the translucency was equal to or greater than 3mm in 73 (14.5%) of the 505 chromosomally normal pregnancies. These results suggest that the accuracy of measurements depends on the motivation of the sonographers54.
University Hospital, Groningen, The Netherlands84
This was a screening study of an apparently low-risk population, but in 54% of the cases the mothers were equal to or greater than 36 years old or had a history of a previous chromosomally abnormal fetus/child. In total, 923 fetuses were scanned transabdominally at equal to or less than 13 weeks of gestation by four ultrasonographers who were instructed not to take more than 3 minutes in making a nuchal translucency measurement. In 54% of cases, the fetal crown–rump length was <33mm. Furthermore, in 42% of cases, the nuchal translucency could not be measured. In this population, there were seven cases of trisomy 21 and the authors suggested that the sensitivity of nuchal translucency screening is low because only two of the fetuses had increased translucency. However, in reality, only three of the fetuses with trisomy 21 had a crown–rump length of > 38mm and a nuchal translucency measurement, and in two of these the translucency was increased84.
Helsinki University Hospital and Jorvi Hospital, Finland86
In this study, transvaginal sonography was performed in 10010 singleton pregnancies at 10–16 weeks of gestation. Scans were performed by one of six sonographers who were successful in obtaining an ultrasound measurement in 98.6% of cases. Increased nuchal translucency (equal to or greater than 3mm) was observed in 76 (0.8%) of the fetuses and this group included seven (54%) of the 13 fetuses with trisomy 21; the sensitivity for pregnancies at 10–14 weeks was 66% (four of six), for a screen-positive rate of only 0.9%86.
Danube Hospital, Vienna, Austria87
In a screening study of 4371 women with singleton pregnancies attending a government hospital in Vienna for routine antenatal care at 10–14 weeks of gestation, transabdominal ultrasound examination was performed and the fetal nuchal translucency thickness was successfully measured in all cases. The nuchal translucency thickness was equal to or greater than 2.5mm in 1.7% of the cases and this group included three (43%) of seven with trisomy 2187.
Academic Medical Center, Amsterdam, The Netherlands88
This study examined 1547 pregnancies, including 24% aged > 36 years old, at 10–14 weeks. Scans were performed by one of six sonographers who were successful in obtaining an ultrasound measurement in 96% of cases. Nuchal translucency was equal to or greater than 3mm in 33 (2.2%) cases and this group included six (67%) of the nine fetuses with trisomy 2188.
Albert Szent-Gyorgyi Medical University Hospital, Szeged, Hungary74
In this study involving 3380 women at 9–12 weeks of gestation, nuchal translucency was successfully measured transvaginally in all cases. Increased translucency (equal to or greater than 3mm) was observed in 81 (2.4%) of fetuses and this group included 28 (90%) of 31 fetuses with trisomy 2174.
University Hospital, Zurich, Switzerland85
In this study, nuchal translucency was measured in 1131 pregnancies at 10–13 weeks of gestation. Increased translucency (equal to or greater than 3mm) was observed in 24 (2.1%) of fetuses and this group included two (67%) of three fetuses with trisomy 2185.
The Multicenter Screening Study
In a multicenter study in the UK, involving the Harris Birthright Centre and four District General Hospitals (St. Peters, Chertsey; Frimley Park, Camberly; Queen Mary’s, Sidcup; Heatherwood, Ascot), nuchal translucency screening at 10–14 weeks of gestation was carried out in 20804 pregnancies, including 164 cases of chromosomal abnormalities61. This study demonstrated that:
The Fetal Medicine Foundation Ongoing Multicenter Project
There are now 43 countries with centers approved by The Fetal Medicine Foundation for carrying out nuchal translucency screening. In the audit of results from the first 100000 pregnancies examined in the UK, the nuchal translucency was above the 95th centile in more than 70% of fetuses with trisomy 2162. The scans were carried out by 306 appropriately trained sonographers in 22 centers. In each pregnancy, the fetal crown–rump length and nuchal translucency were measured and the risk of trisomy 21 was calculated from the maternal age and gestational age-related prevalence, multiplied by a likelihood ratio depending on the deviation in nuchal translucency from normal (Figures 10–12). The distribution of risks was determined and the sensitivity of a cut-off risk of 1 in 300 was calculated62.
In total, 100311 singleton pregnancies were examined and follow-up was obtained from 96127 cases, including 326 with trisomy 21 and 325 with other chromosomal abnormalities. The median gestation at the time of screening was 12 weeks (range 10–14 weeks) and the median maternal age was 31 years (range 14–49 years); in 13315 (13.3%) cases, the maternal age was at least 37 years. The fetal nuchal translucency was above the 95th centile for crown–rump length in 4210 (4.4%) of the normal pregnancies and in 234 (71.8%) of those with trisomy 21 (Figure 10). The estimated risk for trisomy 21 based on maternal age and fetal nuchal translucency was above 1 in 300 in 7907 (8.3%) of the normal pregnancies and in 268 (82.2%) of those with trisomy 21. For a screen-positive rate of 5%, the sensitivity was 77% (95% confidence interval (CI) 72–82%)62.
Table 6 illustrates the observed prevalence of trisomy
21 according to the predicted risk based on maternal age and fetal nuchal
translucency thickness. These results demonstrate the high degree of accuracy
of the model.
|Lethality of trisomy 21 fetuses with increased nuchal translucency|
Screening for chromosomal defects in the first rather than the second trimester has the advantage of earlier prenatal diagnosis and consequently less traumatic termination of pregnancy for those couples who choose this option. A potential disadvantage is that earlier screening preferentially identifies those chromosomally abnormal pregnancies that are destined to miscarry. Approximately 30% of affected fetuses die between 12 weeks of gestation and term38,41,42. This issue of preferential intrauterine lethality of chromosomal defects is, of course, a potential criticism of all methods of antenatal screening, including second-trimester maternal serum biochemistry; the estimated rate of intrauterine lethality between 16 weeks and term is about 20%38,41,42. This section examines the interrelation between increased nuchal translucency in trisomy 21 and fetal lethality.
In a study of 108 fetuses with trisomy 21 diagnosed in the first trimester because of increased nuchal translucency, the parents chose to continue with the pregnancy in five cases, whereas in 103 cases they opted for termination94. Trisomy 21 was also diagnosed in one of the fetuses in a twin pregnancy where the parents elected to avoid invasive prenatal diagnosis or selective fetocide94. In five of the six fetuses, the translucency resolved, and at the second-trimester scan the nuchal-fold thickness was normal (less than 7mm). All six trisomy 21 babies were born alive. One had a major atrioventricular septal defect and died at the age of 6 months. Another two of the babies had small ventricular septal defects and these were managed conservatively, awaiting spontaneous closure. These data suggest that increased nuchal translucency does not necessarily identify those trisomic fetuses that are destined to die in utero.
In a study of 70 pregnancies with trisomy 21 diagnosed at 12 (range 11–14) weeks of gestation, the parents opted for elective termination which was carried out at 14 (12–20) weeks. Ultrasound examination to determine if the fetus was alive was carried out at the time of chorionic villus sampling as well as just before termination95. Eight fetuses died in the interval between chorionic villus sampling and termination and the rate of lethality increased with nuchal translucency thickness from 5.3% for those with nuchal translucency of 0–3mm to 23.5% for nuchal translucency of equal to or greater than 7mm. Assuming that the relative rate of intrauterine lethality of trisomy 21 fetuses according to the nuchal translucency thickness remains the same throughout pregnancy, it was estimated that a policy of screening by maternal age and fetal nuchal translucency followed by selective termination of affected fetuses would be associated with at least a 70% reduction in the live birth incidence of trisomy 21.
Among the 100000 pregnancies that were screened within the multicenter project, trisomy 21 was diagnosed, prenatally or at birth, in 326 cases62. On the basis of the maternal age distribution in this population and the maternal age-related prevalence of trisomy 21 in live births, it was estimated that 266 babies with trisomy 21 would have been born alive had there not been any antenatal testing and selective termination of affected pregnancies.
In the screen-negative group (estimated risk of less than 1 in 300), there were 35 live births with trisomy 21 and 23 other cases where the pregnancies were terminated following prenatal diagnosis. On the extreme assumption that all 23 of these pregnancies would have resulted in live births, then the number of trisomy 21 live births in the screen-negative group would have been 58, or 22% of the total 266 potential live births with trisomy 21. Consequently, assessment of risk by a combination of maternal age and fetal nuchal translucency, followed by invasive diagnostic testing for those with a risk of equal to or greater than 1 in 300, and selective termination of affected fetuses would have reduced the potential live birth prevalence of trisomy 21 by at least 78% (208 of 266)62.
|INCREASED NUCHAL TRANSLUCENCY AND OTHER CHROMOSOMAL DEFECTS|
The Fetal Medicine Foundation Multicenter Project of screening for trisomy
21 by a combination of maternal age and fetal nuchal translucency thickness
at 10–14 weeks, 325 with chromosomal abnormalities other than trisomy
21 were identified62. In 229 (70.5%) of these, the fetal nuchal
translucency was above the 95th centile of the normal range for crown–rump
length (Table 7). Furthermore, in 253 (77.9%) of the pregnancies,
the estimated risk for trisomy 21, based on maternal age and fetal nuchal
translucency, was more than 1 in 300.
In trisomy 21, the median nuchal translucency thickness is about 2.0mm above the normal median for crown–rump length. The corresponding values for trisomies 18 and 13, triploidy and Turner syndrome are 4.0mm, 2.5mm, 1.5mm and 7.0mm, respectively.
addition to increased nuchal translucency, there are other characteristic
sonographic findings in these fetuses (Table 8). In trisomy 18, there is early onset intrauterine
growth restriction, relative bradycardia and, in about 30% of the cases,
there is an associated exomphalos (Figure 13)96.
Trisomy 13 is characterized by fetal tachycardia, observed in about two-thirds
of the cases, early onset intrauterine growth restriction, and holoprosencephaly
or exomphalos in about 30% of the cases97. In triploidy, there
is early onset asymmetrical intrauterine growth restriction (Figure 14), relative bradycardia, holoprosencephaly,
exomphalos or posterior fossa cyst in about 40% of cases and molar changes
in the placenta in about one-third of cases99.
Turner syndrome is characterized by fetal tachycardia, observed in about 50% of the cases, and early onset intrauterine growth restriction98.
|CROWN–RUMP LENGTH IN CHROMOSOMALLY ABNORMAL FETUSES|
birth weight is a common feature of many chromosomal abnormalities100,101.
Furthermore, prenatal studies during the second and third trimesters of
pregnancy have reported a high prevalence of aneuploidies in severe
intrauterine growth restriction102.
examining first-trimester growth in chromosomally abnormal fetuses have
demonstrated that trisomy 18 and triploidy are associated with moderately
severe growth restriction, trisomy 13 and Turner syndrome with mild growth
restriction, whereas in trisomy 21 growth is essentially normal (Table
9 and Table 10, Figure 15).
In 10–45% of pregnancies, women are uncertain of their last menstrual period, they have irregular menstrual cycles or they became pregnant soon after stopping the oral contraceptive pill109,110. Additionally, because of considerable variations in the day of ovulation, in approximately 10% of women with certain dates and regular 28-day cycles, there is a discrepancy of more than 7 days in gestation calculated from the menstrual history and by ultrasound111. For these reasons, accurate dating of pregnancy necessitates ultrasonographic examination. A policy of routine pregnancy dating by measurement of crown–rump length will not affect the interpretation of results in screening by nuchal translucency thickness for trisomy 21. In the case of the other chromosomal defects, dating by crown–rump length will actually improve their detection since nuchal translucency normally increases with gestation.
|FETAL HEART RATE IN CHROMOSOMALLY ABNORMAL FETUSES|
examining first-trimester fetal heart rate in chromosomally abnormal fetuses have
demonstrated that trisomy 13 and Turner syndrome are associated with tachycardia,
whereas in trisomy 18 and triploidy there is a tendency for bradycardia.
In trisomy 21, there is a mild increase in fetal heart rate (Table 11, Figure 16).
In a study of 10083 normal pregnancies at the Harris Birthright Research Centre for Fetal Medicine, the mean fetal heart rate decreased with gestation from 169bpm at a fetal crown–rump length of 38mm to 154bpm at a crown–rump length of 84mm. The data were normally distributed and the 95th and 5thcentile were 10bpm above and below the appropriate normal mean for crown–rump length, respectively. In trisomy 13, Turner syndrome and trisomy 21, the respective mean fetal heart rates were 14bpm, 11.4bpm and 1.4bpm above the normal mean for crown–rump length, whereas, in trisomy 18 and triploidy, the fetal heart rates were 3.4bpm and 4.8bpm below the normal mean, respectively.
Previous studies on trisomy 21 fetuses have reported conflicting results. In a longitudinal study of one trisomy 21 fetus at 6–9 weeks of gestation, the heart rate was consistently below the 3rd centile of the normal range112. In another cross-sectional study of five affected fetuses at 7–13 weeks, the heart rate was always within the normal range113. A study of 17 trisomy 21 fetuses at 10–13 weeks reported that, in 23.5% of cases, the heart rate was either above the 97th centile or below the 2.5th centile114. In another study of 85 trisomy 21 fetuses at 10–14 weeks, the heart rate was above the 95th centile in 21% of cases and the increase in heart rate was not related to fetal nuchal translucency thickness. This finding raises the possibility of including fetal heart rate in the model of risk assessment for trisomy 21 along with maternal age and fetal nuchal translucency115. In our extended series of 451 fetuses with trisomy 21 at 10–14 weeks, 13.7% had a heart rate above the 95th centile (Table 11).
In normal pregnancy, the fetal heart rate increases from about 110bpm at 5 weeks of gestation, to 170bpm at 9 weeks and then gradually decreases to 150bpm by 14 weeks115–118. The early increase in heart rate coincides with the morphological development of the heart, and the subsequent decrease may be the result of functional maturation of the parasympathetic system116,118,119.
The tachycardia in Turner syndrome and trisomy 13 fetuses may be due to a delay in the functional maturation of the parasympathetic system, resulting in a delay in the physiological decrease in heart rate with gestation after 9 weeks. Alternatively, the higher heart rate of such fetuses represents a compensatory response to the heart strain that may also be responsible for the increased nuchal translucency120. In fetal life, the heart normally performs near the peak of the Frank–Starling curve of ventricular function121 and therefore compensatory increase in cardiac output can only be achieved by relative tachycardia122. Maximum tachycardia may be reached, with early heart failure offering an explanation for the lack of a significant association between the extent of increase in nuchal translucency thickness and fetal heart rate. The same hypothesis may also be advanced for the observed mild increase in heart rate of trisomy 21 fetuses.
relative bradycardia of trisomy 18 fetuses may be related to the fact
that, in this chromosomal abnormality, there is early onset growth restriction
and the developmental delay is more severe than in trisomies 21 and 13;
in such fetuses, the maturation in heart rate would be equivalent to about
8 weeks of gestation. Triploidy is associated with a high rate of early
intrauterine lethality and the observed bradycardia in some of these fetuses
may represent a preterminal event.
The tables 11a and 11b showsthe effects of the chromosomal defects on fetal heart rate at 10-14 weeks. (Liao et col, Ultrasound Obstet Gynecol 2000; 16: 610-611)
|DOPPLER ULTRASOUND FINDINGS IN CHROMOSOMALLY ABNORMAL FETUSES|
ultrasound studies have demonstrated that impedance to flow (measured
as pulsatility index) decreases with gestation123,124.
This decrease is believed to be a consequence of the increase in the number
of vessels (and their relative volume) within the chorionic villi and
the expansion of the intervillous circulation125.
is contradictory evidence on the possible association of trisomy 21 at
11–14 weeks of gestation and increased umbilical artery pulsatility index.
Martinez et al. reported that the umbilical artery pulsatility
index was above the 95th centile in 55% of their nine cases of trisomy
21 and this was not always associated with an increased nuchal translucency;
they estimated that the measurements of both factors might allow detection
of up to 89% of cases of trisomy 21126. In contrast, Jauniaux
et al. examined 11 cases of trisomy 21 and reported that there
was no significant difference in umbilical artery pulsatility index
compared to normal fetuses and that in none of their cases was the
pulsatility index above the 95th centile127. Similarly, Brown
et al. examined 19 trisomy 21 fetuses with increased nuchal
translucency at 11–14 weeks and reported that the umbilical artery
pulsatility index was not significantly different from normal; the
pulsatility index was above the 95th centile in only two of the cases124.
In normal second- and third-trimester fetuses, pulsatile umbilical venous flow is observed only during fetal breathing. Pulsatile venous flow is also observed in fetuses with growth restriction and in non-immune hydrops and is considered to be a late and ominous sign of fetal compromise128,129. Evidence from growth-restricted human fetuses and animal models suggests that pulsatile venous flow may result from an increased reversal of flow in the inferior vena cava during atrial contraction, associated with heart failure and abnormal cardiac filling129,130.
A Doppler study at 11–14 weeks of gestation reported the presence of pulsatile flow in the umbilical vein in about 25% of 302 chromosomally normal fetuses and in 90% of 18 fetuses with trisomy 18 or 13; in 18 fetuses with trisomy 21, the prevalence of pulsatile venous flow was not significantly different from that in chromosomally normal fetuses, but in trisomies 13 and 18 the prevalence was increased131.
ductus venosus is a unique shunt that carries well-oxygenated blood from
the umbilical vein through the inferior atrial inlet on its way across
the foramen ovale. It appears to be the most useful vessel in assessing
disturbed cardiac function132. Blood flow in the ductus is
characterized by high velocity during ventricular systole (S-wave) and
diastole (D-wave) and by the presence of forward flow during atrial contraction
(A-wave). In cardiac failure, with or without cardiac defects, there is
absent or reversed A-wave (see Chapter 3, page 102)133.
is possible to assess ductus venosus blood flow at 11–14 weeks of gestation
by Doppler ultrasound, both transabdominally and transvaginally. A right
ventral mid-sagittal plane of the fetal trunk is first obtained during
fetal quiescence and the pulsed Doppler gate is placed in the distal portion
of the umbilical sinus. The inferior vena cava, left and medial hepatic
veins and the ductus venosus drain into a common sub-diaphragmatic vestibulum
and therefore, when attempting to obtain flow velocity waveforms from
the ductus, care should be taken to avoid contamination from the other
A study, examining ductal flow at 11–14 weeks in fetuses with increased nuchal translucency, reported absent or reversed flow during atrial contraction in 57 of 63 (90.5%) chromosomally abnormal fetuses and in only 13 of 423 (3.1%) chromosomally normal fetuses134. In seven of the 13 chromosomally normal fetuses with absent or reversed flow, an ultrasound scan at 14–16 weeks demonstrated a major cardiac defect134.
Examination of ductal flow is time-consuming and requires skilled operators. It is therefore unlikely that this assessment will be incorporated into the routine first-trimester scan. However, the data suggest that the assessment of ductal flow can potentially play a major role as a secondary method of screening in order to achieve a major reduction in the false-positive rate of primary screening for chromosomal abnormalities by a combination of maternal age, fetal nuchal translucency and maternal serum free b-hCG and PAPP-A at 11–14 weeks. A policy of reserving invasive testing only for those with abnormal ductal flow could reduce the overall need for chorionic villus sampling from 5% to less than 0.5%, with a small reduction (5–10%) in the estimated sensitivity of 90%134.
|NUCHAL TRANSLUCENCY AND MATERNAL SERUM BIOCHEMISTRY|
trisomy 21 during the first trimester of pregnancy, the maternal serum
concentration of free b-hCG is higher than in chromosomally normal
fetuses (Table 12), whereas PAPP-A is lower (Table 13). Pregnancy-specific b-1 glycoprotein (SP1),
a-fetoprotein and inhibin-A do not provide a useful distinction between
affected and normal pregnancies135–137.
Maternal serum free b-hCG normally decreases with gestation after 10 weeks. In trisomy 21 pregnancies, the levels are increased and the difference between these and those of normal pregnancies increases with advancing gestation. This may account for the variation in the reported median MoM between the various studies (Table 12)138–158, because there was a considerable variation in the gestational age range of the populations that were examined. Consequently, population parameters derived from studies using a wide gestational age range are not appropriate if screening is to be focused on the optimal time for nuchal translucency measurement (11–14 weeks). The increase with gestation in the difference between trisomy 21 and normal pregnancies has also been shown in studies of paired samples from trisomy 21 pregnancies collected in the first and second trimesters153. In a study involving 210 trisomy 21 pregnancies that were examined at 10–14 weeks, the median free b-hCG was 2.15MoM (95% CI, 1.94–2.33); at a 5% screen-positive rate, the detection rate using free b-hCG alone is about 35% and, in combination with maternal age, the detection increases to about 45%158.
Maternal serum PAPP-A normally increases with gestation. In trisomy 21 pregnancies, the levels are lower but the difference between trisomy 21 and normal pregnancies decreases with advancing gestation. This may account for the variation in the reported median MoM between the various studies (Table 13)140,143,145,146,150,152,153,155–166. In a study involving 210 trisomy 21 pregnancies that were examined at 10–14 weeks, the median PAPP-A was 0.51MoM (95% CI, 0.44–0.56); at a 5% screen-positive rate, the detection rate using PAPP-A alone is about 40% and, in combination with maternal age, the detection increases to about 50%158.
considering to combine biochemical markers, it is necessary to take into
account the degree of correlation between the markers. In our study, involving
210 trisomy 21 and 946 chromosomally normal controls, the correlations
were 0.216 and 0.160, respectively158. Additionally, each marker
showed a small but significant negative correlation with maternal weight
(PAPP-A, r = -0.278; free b-hCG, r = -0.146). After combining
free b-hCG and PAPP-A with maternal age in mathematical models, it
has been estimated that the detection rate of trisomy 21 is about 60%
at a 5% screen-positive rate (Table 14) 150,152,153,155–158,167,168.
is no significant association between fetal nuchal translucency and maternal
serum free b-hCG or PAPP-A in either trisomy 21 or chromosomally normal
pregnancies 147,158,164. The estimated detection rate for trisomy
21 by a combination of maternal age, fetal nuchal translucency and maternal
serum PAPP-A and free b-hCG is about 90% for a screen-positive rate of
5% (Table 15)148,157,158,164,167,169. Alternatively,
at a fixed detection rate of 70%, the screen-positive rate would
be only 1%158. The performance of the combined test now requires
assessment in prospective studies.
An important development in biochemical analysis is the introduction of a new technique (random access immunoassay analyzer using time-resolved-amplified-cryptate-emission), which provides automated, precise and reproducible measurements within 30 minutes of obtaining a blood sample158. This has made it possible to combine biochemical and ultrasonographic testing as well as to counsel in one-stop clinics for early assessment of fetal risk (OSCAR).
|NUCHAL TRANSLUCENCY FOLLOWED BY SECOND-TRIMESTER BIOCHEMISTRY|
Brizot and Penelope Noble
At 16 weeks of gestation, the median maternal serum concentrations of a-fetoprotein, estriol, hCG (total and free b) and inhibin A in trisomy 21 pregnancies are different from normal. The risk for trisomy 21 can be derived by multiplying the background maternal age and gestational age-related risk by the likelihood ratios for these substances, after corrections for the interrelations between them. The risk of trisomy 21 is increased if the levels of hCG and/or inhibin A are high, and the levels of a-fetoprotein and/or estriol are low. The estimated detection rates are 50–70% for a screen-positive rate of about 5%.
In women having second-trimester biochemical testing following first-trimester nuchal translucency screening (with or without maternal serum biochemistry), the background risk needs to be adjusted to take into account the first-trimester screening results. Since first-trimester screening identifies almost 90% of trisomy 21 pregnancies, second-trimester biochemistry will identify – at best – 6% (60% of the residual 10%) of the affected pregnancies, with doubling of the overall invasive testing rate (from 5% to 10%). It is theoretically possible to use various statistical techniques to combine nuchal translucency with different components of first-trimester and second-trimester biochemical testing. One such hypothetical model has combined first-trimester nuchal and PAPP-A with second-trimester free b-hCG, estriol and inhibin A, claiming a potential sensitivity of 94% for a 5% false-positive rate170. Even if the assumptions made in this statistical technique are valid, it is unlikely that it will gain widespread clinical acceptability171.
Two studies have reported on the impact of first-trimester screening by nuchal translucency on second-trimester serum biochemical testing. In one study, the proportion of affected pregnancies in the screen-positive group (positive predictive value) of screening by the double test in the second trimester was 1 in 40; after the introduction of screening by nuchal translucency, 83% of trisomy 21 pregnancies were identified in the first trimester and the positive predictive value of biochemical screening decreased to 1 in 200172. In the second study, first-trimester screening by nuchal translucency identified 71% of trisomy 21 pregnancies for a screen-positive rate of 2%, and the positive predictive value of second-trimester screening by the quadruple test was only 1 in 150173.
In women who had first-trimester screening by a combination of fetal nuchal translucency and maternal serum PAPP-A and free b-hCG, it is clearly advisable that second-trimester biochemical testing is avoided because, first, the sensitivities of first- and second-trimester biochemical screening are similar; second, the main component of the second-trimester biochemical screening is free b-hCG, and, third, there is a good correlation between first- and second-trimester maternal serum hCG levels. If both first- and second-trimester biochemical testing have been carried out, then the likelihood ratio from the measurement of nuchal translucency can be multiplied with the results of either first- or second-trimester serum testing. This is certainly valid for second-trimester programs that are mainly based on free b-hCG because the interrelation between nuchal translucency and this metabolite has been established148.
Major chromosomal abnormalities are often associated with multiple fetal defects that can be detected by ultrasound examination. For example, trisomy 21 is associated with a tendency for brachycephaly, mild ventriculomegaly, flattening of the face, nuchal edema, atrioventricular septal defects, duodenal atresia and echogenic bowel, mild hydronephrosis, shortening of the limbs, sandal gap and clinodactyly or mid-phalanx hypoplasia of the fifth finger. Trisomy 18 is associated with strawberry-shaped head, choroid plexus cysts, absent corpus callosum, enlarged cisterna magna, facial cleft, micrognathia, nuchal edema, heart defects, diaphragmatic hernia, esophageal atresia, exomphalos, renal defects, myelomeningocele, growth restriction and shortening of the limbs, radial aplasia, overlapping fingers and talipes or rocker bottom feet.
overall risk for chromosomal abnormalities increases with the total number
of defects that are identified (Figure 17)174. It is therefore recommended
that, when a defect/marker is detected at routine ultrasound examination,
a thorough check is made for the other features of the chromosomal abnormality
known to be associated with that marker; should additional defects be
identified, the risk is dramatically increased. In the case of apparently
isolated defects, the decision of whether to carry out an invasive test
depends on the type of defect.
If the mid-trimester scan demonstrates major defects, it is advisable to offer fetal karyotyping, even if these defects are apparently isolated. The prevalence of these defects is low and therefore the cost implications are small. If the defects are either lethal or they are associated with severe handicap, fetal karyotyping constitutes one of a series of investigations to determine the possible cause and thus the risk of recurrence. Examples of these defects include hydrocephalus, holoprosencephaly, multicystic renal dysplasia and severe hydrops. In the case of isolated neural tube defects, there is controversy as to whether the risk for chromosomal defects is increased. Similarly, for skeletal dysplasias where the likely diagnosis is obvious by ultrasonography, it would probably be unnecessary to perform karyotyping. If the defect is potentially correctable by intrauterine or postnatal surgery, it may be logical to exclude an underlying chromosomal abnormality – especially because, for many of these conditions, the usual abnormality is trisomy 18 or 13. Examples include facial cleft, diaphragmatic hernia, esophageal atresia, exomphalos and many of the cardiac defects. In the case of isolated gastroschisis or small bowel obstruction, there is no evidence of increased risk of trisomies.
on an individual
estimated risk for a chromosomal abnormality, rather than the arbitrary
advice that invasive testing is recommended because the risk is ‘high’.
The estimated risk can be derived by multiplying the background risk
(based on maternal age, gestational age, history of previously affected
pregnancies and, where appropriate, the results of previous screening
by nuchal translucency and/or biochemistry in the current pregnancy) by
the likelihood ratio of the specific defect175–177. For the
following conditions, there are sufficient data in the literature to estimate
the likelihood ratio for trisomy 21.
There are no data on the interrelation between these second-trimester ultrasound markers and nuchal translucency at 11–14 weeks or first- and second-trimester biochemistry. However, there is no obvious physiological reason for such an interrelation and it is therefore reasonable to assume that they are independent. Consequently, in estimating the risk in a pregnancy with a marker, it is logical to take into account the results of previous screening tests. For example, in a 20-year-old woman at 20 weeks of gestation (background risk of 1 in 1295), who had a 11–14 week assessment by nuchal translucency measurement that resulted in a 5-fold reduction in risk (to about 1 in 6475), after the diagnosis of mild hydronephrosis at the 20-week scan, the estimated risk has increased by a factor of 1.5 to 1 in 4317. In contrast, for the same ultrasound finding of fetal mild hydronephrosis in a 40-year-old woman (background risk of 1 in 82), who did not have nuchal translucency or biochemistry screening, the new estimated risk is 1 in 55.
There are some exceptions to this process of sequential screening, which assumes independence between the findings of different screening results. The findings of nuchal edema or a cardiac defect at the mid-trimester scan cannot be considered independently of nuchal translucency screening at 11–14 weeks. Similarly, hyperechogenic bowel (which may be due to intra-amniotic bleeding) and relative shortening of the femur (which may be due to placental insufficiency) may well be related to serum biochemistry (high free b-hCG and inhibin-A and low estriol may be markers of placental damage) and can therefore not be considered independently in estimating the risk for trisomy 21. For example, in a 20-year-old woman (background risk for trisomy 21 of 1 in 1295), with high free b-hCG and inhibin-A and low estriol at the 16-week serum testing resulting in a 10-fold increase in risk (to 1 in 129), the finding of hyperechogenic bowel at the 20-week scan should not lead to the erroneous conclusion of a further three-fold increase in risk (to 1 in 43). The coincidence of biochemical and sonographic features of placental insufficiency makes it very unlikely that the problem is trisomy 21 and should lead to increased monitoring for pre-eclampsia and growth restriction, rather than amniocentesis for fetal karyotyping.
During the last 30 years, extensive research has aimed at developing a non-invasive method for prenatal diagnosis based on the isolation and examination of fetal cells found in the maternal circulation. Erythroblasts have attracted most attention because they are abundant in early fetal blood; they are extremely rare in normal adult blood and their half-life in adult blood is only about 30 days. Trophoblastic cells entering the maternal circulation are cleared by the maternal lungs and are therefore not useful candidates for prenatal diagnosis. Fetal white blood cells are present in maternal blood but their number is very low and they have a very long half-life (about 5 years), which may therefore lead to contamination from previous pregnancies.
About 1 in 103–107 nucleated cells in maternal blood are fetal194–196. The proportion of fetal cells can be enriched to about 1 in 10–100 by techniques such as magnetic cell sorting (MACS) or fluorescence activated cell sorting (FACS) after attachment of magneticallylabeled or fluorescent antibodies on to specific fetal cell surface markers194,197–199. The most commonly used antibody is anti-CD71, which is directed against the transferrin receptor present on the surface of all cells actively incorporating iron198,200. Other cell types in maternal blood, such as activated lymphocytes, have this receptor but anti-CD71 provides a reasonable level of enrichment once maternal lymphocytes have been removed. Magnetic cell sorting is cheaper, quicker and requires less expertise to perform than FACS. The technique utilizes metallic beads labeled with an antibody specific for the target cell. The antibody is incubated with the sample and the cell–antibody–bead complex is isolated by placing on a magnet. Successful use of MACS involves prior separation of cells by triple density centrifugation. Essentially, the maternal blood sample is placed in a tube containing three sugar-based reagents of different density and, after centrifugation, the middle layer containing erythroblasts and neutrophil granulocytes is separated. These cells are incubated with magnetically labeled CD71 antibody and MACS is then carried out (Figure 18).
resulting sample is unsuitable for traditional cytogenetic analysis because
it is still highly contaminated with maternal cells. However, with the
use of chromosome-specific DNA probes and fluorescent in situ hybridization
(FISH), it is possible to suspect fetal trisomy by the presence of three-signal
nuclei in some of the cells of the maternal blood enriched for fetal cells.
It is now possible to identify simultaneously all major chromosomal abnormalities
by the use of multicolor probes directed against chromosomes 21, 18, 13,
Y and X in interphase nuclei (Figure 19). One of the major problems with FISH is
that 1–2% of normal diploid cells give three-signal nuclei and about 10–20%
of trisomic cells give two-signal nuclei201.
Bianchi et al. detected three-signal nuclei from a trisomy 21 pregnancy after enrichment for fetal cells in maternal blood by FACS202. Ganshirt-Ahlert et al. found three-signal nuclei in 9–17% of cells from ten trisomy 21 and six trisomy 18 pregnancies after sorting by MACS; in ten chromosomally normal pregnancies, 0–7% of cells had three-signal nuclei203. Simpson and Elias reported the presence of three-signal nuclei, after sorting by FACS, in 2.8–74% of cells from five trisomy 21 and two trisomy 18 pregnancies, but in none of 61 chromosomally normal pregnancies204.
Al-Mufti et al. took maternal peripheral blood immediately before chorionic villus sampling from 230 women with singleton pregnancies at 11–14 weeks of gestation199. These pregnancies had been identified as being at high risk for trisomies after screening by a combination of maternal age and fetal nuchal translucency thickness. Triple density gradient centrifugation, followed by incubation of the erythroblast-rich fraction with magneticallylabeled CD71 antibody, MACS and FISH were carried out. In 3% of cases, no fetal hemoglobin-positive cells were observed. In the chromosomally abnormal group, the percentage of cells demonstrating three-signal nuclei was higher than in the normal group but there was an overlap in values between the two groups (Figure 20).
a 21-chromosome-specific probe, three-signal nuclei were present in at
least 5% of the enriched cells from 61% of the trisomy 21 pregnancies
and in none of the normal pregnancies. For a cut-off of 3% of three-signal
nuclei, the sensitivity for trisomy 21 was 97% for a false-positive rate
of 13%. Similar values were obtained in trisomies 18 and 13 using the
appropriate chromosome-specific probe (Table 16).
The findings that, with the 21-chromosome-specific probe, three-signal nuclei were present in at least 5% of the enriched cells from about 60% of trisomy 21 pregnancies and in none of the normal pregnancies, suggest that this method could be associated with the same rate of detection of trisomy 21 as second-trimester serum biochemistry but with the advantage that the invasive testing rate may be as low as 0% rather than 5%. However, unlike serum biochemistry testing, which is relatively easy to apply for mass population screening, enrichment of fetal cells by triple density gradient centrifugation and MACS, followed by FISH, is both labor intensive and requires highly skilled operators. In the case of FISH, there are promising developments for automated computerized analysis of cells which are likely to simplify processing of the slides. The extent to which the techniques for enrichment of fetal cells could be improved, to achieve a higher yield of the necessary cells, as well as become automated, to allow simultaneous analysis of a large number of samples, remains to be seen.
On the basis of currently available technology, examination of fetal cells from maternal peripheral blood is more likely to find an application as a method for assessment of risk, rather than the non-invasive prenatal diagnosis of chromosomal defects. First-line screening by a combination of maternal age, fetal nuchal translucency and maternal serum free b-hCG with PAPP-A could detect 90% of trisomy 21 pregnancies for an invasive testing rate of about 5%158. One option in the management of the high-risk group is to carry out FISH on maternal blood enriched for fetal cells and reserve chorionic villus sampling only for those pregnancies where no fetal hemoglobin-positive cells are recovered and those where at least 3% of the cells demonstrate three signals with the 21-chromosome specific probe. Such a policy could potentially reduce the need for invasive testing to less than 1% of the whole population with a minor loss (about 3%) in the sensitivity for detection of trisomy 21.
Ager and Oliver reported a critical appraisal of all the studies on amniocentesis that were published during 1975–85205. There were 28 major national studies, each involving at least 1000 cases; the total post-amniocentesis fetal loss rate, including spontaneous abortion, intrauterine death and neonatal death was 2.4–5.2%. In four of the 28 studies, there were matched controls that did not undergo amniocentesis; the total fetal loss rate was 1.8–3.7%. On the basis of these data it was estimated that the procedure-related risk of fetal loss from amniocentesis was 0.2–2.1%205.
The only randomized trial was performed in Denmark206. In this study, 4606 low-risk, healthy women, 25–34 years old, at 14–20 weeks of gestation, were randomly allocated to amniocentesis or ultrasound examination alone. The total fetal loss rate in the patients having amniocentesis was 1% higher than in the controls. There were significant associations between spontaneous fetal loss and (1) puncture of the placenta, (2) high maternal serum a-fetoprotein and (3) discolored amniotic fluid. The Danish study also reported that amniocentesis was associated with an increased risk of respiratory distress syndrome and pneumonia in neonates. Some studies have reported an increased incidence of talipes and dislocation of the hip after amniocentesis, but this was not confirmed by the Danish study206.
A prospective study involving 1301 singleton pregnancies compared early amniocentesis with chorionic villus sampling at 10–13 weeks of gestation207. The procedures were performed (1) for the same indication, (2) at the same gestational range, (3) by the same group of operators, (4) using essentially the same technique of transabdominal ultrasound-guided insertion of a 20-G needle, and (5) the samples were sent to the same laboratories. Successful samplings resulting in a non-mosaic cytogenetic result were the same for both early amniocentesis and chorionic villus sampling (97.5%). Furthermore, the intervals between sampling and obtaining results were similar for the two techniques. The main indication for repeat testing in the chorionic villus sampling group was mosaicism, whereas, in the early amniocentesis group, it was failed culture; this failure was related to gestation at sampling: 5.3% at 10 weeks and 1.6% at 11–13 weeks. Spontaneous loss (intrauterine and neonatal death) after early amniocentesis was approximately 3% higher than after chorionic villus sampling. The gestation at delivery and birth weight of the infants were similar after both procedures, and the frequencies of preterm delivery or low birth weight were not higher than those that would be expected in a normal population. In the early amniocentesis group, the incidence of talipes equinovarus (1.63%) was higher than in the chorionic villus sampling group (0.56%)207.
A randomized study in Denmark involving 1160 pregnancies compared transabdominal chorionic villus sampling at 10–12 weeks with early amniocentesis at 11–13 weeks using a filtration technique; randomization was at 10 weeks208. Fetal loss after chorionic villus sampling was 4.8% and after early amniocentesis it was 5.4%, but this difference was not significant. The study was stopped early because interim analysis of results demonstrated a significantly higher rate of talipes equinovarus (1.7%) after early amniocentesis than after chorionic villus sampling (0%)208.
A randomized study in Canada involving 4374 pregnancies compared early amniocentesis at 11–13 weeks with amniocentesis at 15–17 weeks using a 22-G needle; randomization was at 9–12 weeks209. Total fetal loss in the early amniocentesis group (7.6%) was significantly higher than in the late amniocentesis group (5.9%). Furthermore, early amniocentesis was associated with a significantly higher incidence of talipes (1.3% compared to 0.1%) and postprocedural amniotic fluid leakage (3.5% compared to 1.7%)209.
On the basis of existing data, it is therefore clear that amniocentesis should not be carried out before 13 weeks of gestation. The extent to which early amniocentesis performed after 13 weeks will prove to be safer than chorionic villus sampling is currently under investigation by an NIH-sponsored study in the USA.
Chorionic villus sampling was first attempted in the late 1960s by hysteroscopy, but the technique was associated with low success in both sampling and karyotyping and was abandoned in favor of amniocentesis. In the 1970s, the desire for early diagnosis led to the revival of chorionic villus sampling, which was initially carried out by aspiration via a cannula that was introduced ‘blindly’ into the uterus through the cervix. Subsequently, ultrasound guidance was used for the transcervical or transabdominal insertion of a variety of cannulas or biopsy forceps.
randomized studies have examined the rate of fetal loss following first-trimester
chorionic villus sampling compared to that of amniocentesis at 16 weeks
of gestation (Table 17)210–213. In total, about 10000
pregnancies were examined and the results demonstrated that, in centers
experienced in both procedures, fetal loss is no greater after first-trimester
chorionic villus sampling compared to second-trimester amniocentesis.
The most likely explanation for the increased loss after chorionic villus
sampling in the European study is the participation of many centers with
little experience in this technique.
In 1991, severe transverse limb abnormalities, micrognathia and microglossia were reported in five of 289 pregnancies that had undergone chorionic villus sampling at less than 10 weeks of gestation214. Subsequently, a series of other reports confirmed the possible association between early chorionic villus sampling and fetal defects; analysis of 75 such cases demonstrated a strong association between the severity of the defect and the gestation at sampling215. Thus, the median gestation at chorionic villus sampling was 8 weeks for those with amputation of the whole limb and 10 weeks for those with defects affecting the terminal phalanxes. The background incidence of terminal transverse limb defects is about 1.8 per 10000 live births216, and the incidence following early chorionic villus sampling is estimated at 1 in 200–1000 cases. The types of defects are compatible with the pattern of limb development, which is essentially completed by the 10th week of gestation. Possible mechanisms by which early sampling may lead to limb defects include hypoperfusion, embolization or release of vasoactive substances, and all these mechanisms are related to trauma. It is therefore imperative that chorionic villus sampling is performed only after 11 weeks by appropriately trained operators. The data from the International Registry on chorionic villus sampling are disputing the association between this procedure and limb reduction defects217.
The 11-14-week scan
Copyright © 2001 by KH Nicolaides, NJ Sebire, RJM Snijders, RLS Ximenes & G. Pilu