- Top of page
The prevalence of trisomy 21 increases with maternal age1, 2. Until 1984, advanced maternal age was the only known recognized risk factor to justify amniocentesis for fetal karyotyping. In the mid-1980s, specific maternal serum biochemical markers in the second trimester of pregnancy were found to be of value in detecting trisomy 213–5. Maternal serum screening in the second trimester has become the most commonly performed screening method for trisomy 21 in non-selected populations and enables detection of approximately 55–75% of cases with trisomy 216, 7.
Most pregnant women in the industrialized world undergo a prenatal ultrasound examination during the second trimester as a routine part of their antenatal care. In Norway, this was formalized in 19868. The use of ultrasound imaging to examine the fetus in the second trimester led to the detection of both structural anomalies, such as duodenal atresia and atrioventricular septal defects, and the identification of soft markers that have been associated with trisomy 219–14. The sonographic markers associated with increased risk for trisomy 21 are macroglossia, nuchal thickening, renal pyelectasis, mild cerebral ventriculomegaly, echogenic bowel and shortened femur15–17. They are often subtle, have typically been difficult to detect and have a low likelihood ratio.
The most recent method of screening for trisomy 21 is ultrasound examination during the first trimester (11–13 weeks) to measure nuchal translucency (NT). Increased NT thickness has been shown to be associated with chromosomal abnormalities18. With a 5% false-positive rate, screening by a combination of maternal age and fetal NT identifies 75–80% of fetuses with trisomy 21 and other major chromosomal abnormalities19–21. The combination of first-trimester serum biochemistry (pregnancy-associated plasma protein-A and beta-human chorionic gonadotropin) and NT may lead to a detection rate of between 87 and 92%22–24 with a 5% false-positive rate. In the years to come, use of the scan at 11 + 0 to 13 + 6 weeks to detect trisomy 21 will most likely continue to increase worldwide, whereas the traditional screening methods for trisomy 21 using maternal age, and to some extent second-trimester ultrasound imaging, will be replaced by new screening protocols combining early ultrasound and biochemical markers.
In order to assess and compare the emerging techniques it is necessary to evaluate traditional practices. The objective of this study was to assess the contribution of the second-trimester routine ultrasound examination and maternal age to the prenatal detection of trisomy 21 in a large non-selected population in which no other screening methods were offered.
- Top of page
This follow-up study included all prospectively registered cases of cytogenetically confirmed trisomy 21 with a prenatal or postnatal diagnosis at St Olavs University Hospital in Trondheim during the period January 1987 to December 2004. The study population comprised a non-selected population in a geographically well defined area consisting of the city of Trondheim and eight surrounding municipalities. A total of 49 314 births (including miscarriages, terminations of pregnancy and intrauterine fetal deaths (IUFDs)) within the non-selected population were delivered beyond 16 gestational weeks. Within this population, approximately 97% of the pregnant women had a routine fetal examination at the National Center for Fetal Medicine (NCFM), Department of Obstetrics and Gynecology at St Olavs University Hospital in Trondheim, and later delivered there. Cases of trisomy 21 referred to the NCFM, a tertiary referral center, from outside our non-selected population were not included.
Specially trained sonographer/midwives performed the routine ultrasound examination, assumed to be at 18 gestational weeks based on the last menstrual period or early clinical assessment. The routine examination was performed between 16 + 1 and 22 + 5 gestational weeks. A national standardized protocol for the second-trimester ultrasound examination was used. Non-structural soft markers that were evaluated included mild ventriculomegaly (10–15 mm), choroid plexus cysts, nuchal thickening (> 6 mm), renal pelvic dilatation (> 10 mm), hyperechogenic bowel and short femur. All pathology found was presented to specialists in fetal medicine at the center, who performed all follow-up examinations and invasive procedures.
Karyotyping was offered to all women with structural anomalies diagnosed at the ultrasound examination, to women who would be 38 years or older at expected term, and for a few other reasons, such as a history of children with anomalies and/or chromosomal aberrations. During the study period, karyotyping for advanced maternal age was scheduled at approximately 15 completed gestational weeks. Karyotyping was preceded by a standard ultrasound examination. Since 2004 this was performed in week 12. Maternal biochemical serum testing for the detection of trisomy 21 was not available within the Norwegian healthcare system during the study period.
Data from the ultrasound examinations were stored in an electronic database. For each fetus with anomalies, video recordings, biochemical tests and karyotype were registered prospectively. After delivery or termination of pregnancy, further prenatal and postnatal data from the pregnancy, birth and neonatal development were recorded and, if available, autopsy reports and photographs were included. Fetal phenotyping, but not karyotyping, was performed in all cases of miscarriage beyond 16 weeks' gestation and IUFD. One pediatrician was responsible for examining all newborns on day 1. The registry used the nomenclature and coding system from the 9th and 10th revisions of the International Classification of Diseases (ICD-9, ICD-10).
The study period was divided into three 6-year periods: the first from 1987 to 1992, the second from 1993 to 1998, and the third from 1999 to 2004.
The SPSS 11 for Mac OS X software package (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Fisher's exact test and Chi-square test were used to test for significance of observed frequencies. The level of significance was set at 5%.
- Top of page
Within the study population, 88 cases of trisomy 21 were diagnosed by either prenatal or postnatal karyotyping during the period 1987–2004. The distribution and outcome of detected and undetected trisomy 21 fetuses are shown in Figure 1. The total prevalence of trisomy 21 in the non-selected population, including all prenatally and postnatally diagnosed cases, was 1.8 (95% CI, 1.4–2.2) per 1000 births. The prevalence of liveborn trisomy 21 infants (excluding termination of pregnancies and IUFDs) was 1.1 (95% CI, 0.8–1.4) per 1000 births.
Figure 1. Summary of all prenatally detected and undetected trisomy 21 cases in the non-selected population of 49 314 births. IUFD, intrauterine fetal death; TOP, termination of pregnancy.
Download figure to PowerPoint
The median maternal age of women with a trisomy 21 fetus was 34 (range, 19–45) years, and increased from 33 years to 35 years and 37 years in the first, second and third time periods, respectively. For the total non-selected population, a similar and significant increase in median maternal age was noted, from 28 years to 29 years and 30 years, respectively. The proportion of women in the non-selected population who would have been 38 years of age at expected term increased significantly in the three time periods from 2.7% to 3.7% and 4.7%, respectively (P < 0.001). However 72% (63/88) of the women with a trisomy 21 fetus/infant would have been less than 38 years of age at expected term. There was no significant change in the number of trisomy 21 cases during the three time periods.
Prenatally detected trisomy 21 cases
Trisomy 21 was detected prenatally in 38/88 (43%) cases, with no significant change in the detection rate over the three time periods (P = 0.3). A summary of the cases detected before and after birth within the three time periods is given in Table 1. The median maternal age of the women with prenatally detected cases was 36 years; 23/38 (61%) women were below 38 years of age at detection.
Table 1. Pre- and postnatal detection of trisomy 21 in the non-selected population
|Period||Prenatally detected||Postnatally detected|
|Accepted AC due to maternal age||First-trimester ultrasound||Second-trimester ultrasound||Third-trimester ultrasound||Total detection rate||Miscarriage/ IUFD||TOP*|
|1987–1992||3/5||0||8||1||12 (44)||0||9 (75)||15 (56)|
|1993–1998||6/13||1||8||1||16 (39)||4||11 (92)||25 (61)|
|1999–2004||3/7||1||6||0||10 (50)||1||7 (78)||10 (50)|
|Total||12/25||2||22||2||38 (43)||5||27 (82)||50 (57)|
Some 30% (26/88) of all cases of trisomy 21 were detected by particular ultrasound findings that led to karyotyping at median of 18 + 0 (range, 11 + 2 to 36 + 2) weeks. Two of the 26 women had an early ultrasound examination at a median of 13 + 2 gestational weeks (one had a previous history of miscarriage and the other of triploidy). At the second-trimester routine ultrasound examination, 25% (22/88) of cases were detected at median of 18 + 1 (range, 14 + 3 to 22 + 3) weeks. Six of the 22 cases detected in the second trimester did not have structural anomalies, only soft markers. After the second-trimester routine ultrasound examination, 2/26 women were referred for clinical indications, both with abnormal symphysis–fundal height at 22 + 6 and 36 + 2 weeks. The ultrasound findings that triggered further karyotyping in the 26 fetuses are listed in Table 2.
Table 2. Description of suspicious ultrasound findings that triggered karyotyping in 26 cases of trisomy 21
|Trimester||Number of fetuses||Ultrasound anomalies/ markers||n|
|First||2||Increased NT, hydrops||1|
| ||Increased NT, cystic hygroma, hydrops fetalis, hydrothorax||1|
| ||Cystic hygroma||6|
| ||Hydrops fetalis||10|
| ||Ventricular septal defect||1|
| ||Hypoplastic middle phalanx digit 5||2|
| ||Femoral shortening||2|
| ||Mild ventriculomegaly||1|
| ||Choroid plexus cysts||1|
| ||Renal pelvic dilatation||1|
| ||Echogenic focus in the fetal heart||1|
|Third||2||AVSD, hypoplastic middle phalanx digit 5, sandal gap, fetal growth restriction||1|
| ||Duodenal atresia, polyhydramnios||1|
Within the total study population of 49 314 births, amniocentesis was offered in 1811 (3.7%) cases as maternal age was expected to be 38 years or more at the time of delivery. The rate of women accepting this offer diminished significantly during the last period, from 51% to 50% to 36%, respectively. In the group aged 38 years or more with trisomy 21 fetuses/neonates, the rate of women opting for amniocentesis dropped in the three periods from 60% to 54% to 42%, respectively.
Some 12/38 (32%) prenatally detected cases of trisomy 21 in the population were detected by karyotyping performed because of advanced maternal age. Consequently, the ultrasound examination was responsible for the detection of 68% of all detected cases (P = 0.001). Of the 12 cases detected by maternal age screening, eight (67%) had ultrasound findings, such as increased NT (> 3 mm), hygroma colli, fetal hydrops, hydrothorax or atrioventricular septal defects, that would have led to the offer of karyotyping.
Of the 38 women with prenatally detected trisomy 21 fetuses, six continued their pregnancies. One woman was karyotyped at 15 weeks because of advanced age and decided to continue with the pregnancy. Three had suspicious ultrasound findings at the second-trimester routine ultrasound scan, leading to karyotyping between 15 and 22 weeks. All four could have discontinued the pregnancy according to Norwegian law, but elected not to. Two cases were detected by ultrasound examination later, at 25 and 35 gestational weeks. According to Norwegian law, both fetuses were beyond the legal gestational age for termination of pregnancy. Nevertheless, the mother of the trisomy 21 pregnancy detected at week 25 applied for a termination of pregnancy, which was not granted.
There were five IUFDs among the trisomy 21 cases. Three of these were diagnosed at an ultrasound examination approximately 1 week after karyotyping which had been performed in two cases for advanced maternal age and in one with a history of triploidy. Two of these three fetuses had signs of hydrops at ultrasound examination before karyotyping. The fourth intrauterine death was detected at the second-trimester routine examination at 16 weeks of gestation; the fetus then had generalized hydrops. The last case was the IUFD of a trisomy 21 fetus in a twin pregnancy at 25 weeks, detected at routine ultrasound examination. The mother had a normal vaginal delivery at 38 weeks, with birth of a healthy cotwin.
Twenty-seven cases of trisomy 21 were terminated; the termination rate of the prenatally detected trisomy 21 pregnancies was 84% (27/32), excluding miscarriages and cases with no option of termination. The termination rate remained unchanged over time (Table 1). The termination rate among women who received their trisomy 21 diagnosis by opting for an amniocentesis for advanced maternal age was 90%, whereas it was 77% among the ultrasound-detected trisomy 21 fetuses.
Prenatally undetected trisomy 21 cases
A total of 50 trisomy 21 neonates were diagnosed postnatally in the three 6-year periods (15, 25 and 10 cases respectively). The latest postnatal diagnosis was made 3 weeks postpartum. The median maternal age of these 50 women was 31 years for the 18-year period. Some 86% (43/50) of all women with undetected trisomy 21 were aged less than 38 years; 7/50 (14%) were 38 years of age and above. Liveborn undetected trisomy 21 fetuses were delivered at a median of 38 (range, 30–41) weeks of gestation.
In 78% (39/50) of the trisomy 21 cases, the second-trimester ultrasound examination was described as normal. However, structural anomalies were found in 29/39 (74%) cases postnatally. The structural anomalies diagnosed postnatally among all 50 undetected trisomy 21 neonates are listed in Table 3. In addition, mild dysmorphic features were described in 10 of these newborns, such as sandal gap, hypoplastic middle phalanx of the fifth digit or low-set ears.
Table 3. Structural anomalies found postnatally among all 50 cases with prenatally undetected trisomy 21
|Atrial septal defect||9 (18)|
|Ventricular septal defect*||11 (22)|
|Atrioventricular septal defect||6 (12)|
|Double-outlet right ventricle||1 (2)|
|Tetralogy of Fallot||1 (2)|
|Pulmonary artery stenosis||1 (2)|
|Imperforate anus||2 (4)|
|Syndactyly between 3rd and 4th fingers||1 (2)|
In 11/50 prenatally undetected trisomy 21 cases, suspicious ultrasound findings were registered during pregnancy. In seven of these cases, the ultrasound examination was repeated, but trisomy 21 was not suspected. The ultrasound findings registered and additional structural anomalies found after birth in the 11 cases are listed in Table 4.
Table 4. Ultrasound findings among 11/50 cases in which trisomy 21 was undetected prenatally
|Time of detection||Ultrasound findings||Comments||Postnatal diagnosis in addition to trisomy 21|
|Second-trimester routine scan||Choroid plexus cyst, slight dilatation of cavum septi pellucidi||Normal scan at week 27||AVSD|
| ||Limbs < 5th percentile||Asian mother, scan concluded normal||ASD with pulmonary hypertension, rickets with spontaneous costa fractures|
| ||Echogenic intracardiac focus||Normal scan at week 20||VSD, PDA|
| ||Suspected thickening of the neck||Normal scan by second person||VSD|
| ||Talipes equinovarus||No karyotyping offered||None|
| ||Right-left cardiac chamber disproportion|| ||DORV, VSD, pulmonary hypertension|
|Third trimester|| |
| Week 34||Fetal growth restriction|| ||PDA, ASD, VSD, mitral valve incompetence, anemia, heart failure|
| Week 35||Fetal growth restriction and oligohydramnios|| ||Anal atresia, PDA, VSD, bilateral cataract|
| Week 36||Right ventricular cardiac hypertrophy, anomaly of ductus venosus and superior caval vein, fetal growth restriction|| ||ASD, portocaval shunt|
| Week 38||Fetal growth restriction and oligohydramnios|| ||ASD secundum with spontaneous closure|
| Week 38||Oligohydramnios|| ||AVSD, hypothyreosis|
Three children with prenatally undetected trisomy 21 died at the age of 2 and 3 years. One died shortly after surgery for tetralogy of Fallot. The second died from a prenatally diagnosed complex cardiac malformation involving right ventricular cardiac hypertrophy and anomaly of the superior caval vein and the ductus venosus. The child died from cardiac failure while under consideration for cardiac surgery. The third child, who had anal atresia, died shortly after surgery.
Outcome of all trisomy 21 newborns
Among the 56 newborns with trisomy 21, six cases were diagnosed prenatally and 50 cases postnatally. Structural anomalies were found in 43/56 (77%). Twenty-six of the 56 (46%) needed one or more surgical treatments; these are listed in Table 5.
Table 5. Surgical treatment performed in 26/56 trisomy 21 infants
|Reason for surgery||Prenatally detected (n)||Postnatally detected (n)||Number of infants|
|Cardiac malformation|| ||17|
| AVSD||2||7|| |
| PDA|| ||2|| |
| VSD|| ||4|| |
| ASD|| ||3|| |
| DORV|| ||1|| |
| Tetralogy of Fallot|| ||1|| |
| Pulmonary artery stenosis|| ||1|| |
|Anal atresia|| ||3||3|
|Esophageal reflux|| ||2||2|
|Duodenal atresia||1|| ||1|
Only one of the 56 cases was not followed up at St Olavs University Hospital. In addition to the follow-up of delayed physical and mental development, most of the children were frequently referred to the hospital in early childhood for dietary problems, infections of the respiratory tract, otitis media, and surgical procedures such as paracentesis and tonsillectomies. For the last 6-year period from 1999 to 2004 we have complete electronic registration of hospital visits, and in this period 13 trisomy 21 infants were delivered. The median number of contacts with the outpatient clinic per child was 16 during the first postnatal year and 10 during the second year. The median hospitalization time per child was 18 days for the first year and 2 days for the second year, excluding hospital admissions for cardiac surgery, which is centralized to the National Hospital in Oslo.
- Top of page
In our non-selected population, the prevalence of trisomy 21 of 1.8 per 1000 births after 16 weeks' gestation was similar to that found in other population-based studies, ranging between 1.7 and 1.8 per 1000 trisomy 21 cases12, 25–27. Although the median maternal age increased significantly from 28 to 30 years, the prevalence remained unchanged during the study period. A larger number of cases is probably needed to demonstrate the effect of increasing maternal age over time.
The objective of our study was to analyze the prenatal detection rate of trisomy 21 in a large non-selected population including 49 314 births, in which only one second-trimester ultrasound examination was offered to the pregnant women. In addition, karyotyping was offered to women who would be 38 years or older at expected term. Because no maternal serum program for the detection of trisomy 21 was available during the study period, it was possible to elucidate the effect solely of routine ultrasound examination and the offer of karyotyping to detect trisomy 21 because of maternal age. Such a program had a low overall detection rate of trisomy 21 of 43%, partly due to the fact that only 46% of women aged over 38 years accepted the karyotyping offered to them. In our total population, most cases were detected by ultrasound imaging. Advanced maternal age as an indication for karyotyping contributed only 14% of the detected cases, underscoring the inefficiency of such a screening program.
Other studies reporting on the effectiveness of routine ultrasound examination for prenatal detection of trisomy 21 in non-selected populations have shown large variations in detection rates. Detection by ultrasound imaging in the second trimester relies principally on the identification of gross anatomical defects. Between 1989 and 1991, a Scandinavian multicenter study of 27 844 low-risk women between 18 and 34 years of age had an overall detection rate for trisomy 21 of 6.3% at second-trimester obstetric ultrasound examination28. In a European study from 19 centers across Europe, fetal ultrasonographic examination (without NT screening) resulted in prenatal detection of 26% of the trisomy 21 cases, in the time period 1996–199813. In a French registry study, carried out in 1998 in the county of Isère, the mean detection rate for trisomy 21 was 51%, rising from 41% to 67% between 1990 and 199525. However, three ultrasound examinations were offered routinely, and the estimated rate of amniocentesis in the group with advanced maternal age was between 65 and 70%. The increased detection of trisomy 21 during the study period was thought to result from the introduction of first-trimester screening25. Howe et al. retrospectively evaluated the diagnostic sensitivity of routine mid-trimester ultrasound examination in a maternity unit in Southampton, UK between 1993 and 199812. A total of 31 259 pregnancies were screened, with a detection rate for trisomy 21 of 68%. The detection rate in younger women (< 35 years) was 53%. The rate decreased to 41% after exclusion of cases with privately arranged serum or NT screening. In this population, 61% of the women aged 35 years or more had an amniocentesis. In many published studies it is difficult to find exact data about second-trimester ultrasound detection rates, because most countries offer several ultrasound examinations often in combination with second-trimester biochemical tests. In fact, many Western countries have included biochemical markers in the second trimester, achieving detection rates of between 55% and 75% with decreased numbers of amniocenteses7, 24, 29–31. Today several countries additionally offer fetal NT screening with biochemical screening, resulting in detection rates of up to 90%19, 20.
In 1987, Benacerraf et al. indicated that fetuses with trisomy 21 could be identified with a sensitivity of 75% by ultrasound imaging in the second trimester32. More recently, Nyberg et al. reported some type of sonographic finding in 69% of fetuses with trisomy 2133. Other studies from tertiary referral centers report detection rates of up to 91%34. These studies are biased by a selected material with a high prevalence of trisomy 21. Thus, the sensitivity of such studies cannot be compared with our data, which are based on a non-selected population.
In the last 6-year study period, the ultrasound examination performed before an amniocentesis was moved from week 15 to week 12, with a simultaneous significant reduction in the number of women opting for amniocentesis because of maternal age. Apparently, an increasing rate of women then relied exclusively on early ultrasound examination without further testing. During the whole study period our amniocentesis rate was low compared with that in studies from other countries. In France, the amniocentesis rate among women aged 38 years and older was 65–70% in 1986, increasing to 84% in 199525.
The termination rate in our study was 84%, with no change over time. A systematic review, including studies from different countries such as the US, UK and other European countries, showed that the corresponding termination rate for trisomy 21 was approximately 92%35. The attitude towards termination of a trisomy 21 pregnancy in our non-selected population therefore seems quite similar to that in other European studies.
Better ultrasound equipment and improved teaching and training of sonographers has had a significant impact on detection rates for congenital heart defects36. However, it did not have any impact on detection rates of trisomy 21 at the second-trimester routine ultrasound examination in our study. Trisomy 21 was not detected prenatally in many fetuses in which there were additional structural anomalies. In all, 52% of fetuses in which trisomy 21 was not detected prenatally had heart defects that were undetected prenatally. The majority of the heart defects are detectable. Improved detection of cardiac defects might be one of the most important ways to improve rates of detection of trisomy 21. Even so, it appears unlikely that detection rates of trisomy 21 can be improved significantly without extending our prenatal program, for example with first-trimester NT screening in combination with biochemical tests.
Our data showed that many liveborn trisomy 21 infants required surgery or other forms of treatment and long follow-up, resulting in considerable demands on healthcare resources and the families.
In conclusion, the detection rate of trisomy 21 cases remained unchanged throughout the 18-year study period. The data can be considered as a future reference standard for population screening for trisomy 21 using maternal age and second-trimester ultrasound examination without the influence of confounding factors such as private healthcare, biochemical serum screening and first-trimester NT measurements. Amniocentesis offered because of advanced maternal age had both a poor detection rate and a high false-positive rate, and ought to be considered obsolete. Second-trimester ultrasound imaging had a higher detection rate, but remained unable to improve the total detection rate of trisomy 21 to an adequate level. The termination rates for trisomy 21 in this study were similar to those in other Western countries, reflecting the women's own decisions. Such decisions should be based on the best possible information. If improvement of the detection rates of trisomy 21 is desired, additional programs are necessary, such as first-trimester ultrasound examinations.