Aim of screening for fetal structural anomalies

The overall aim of fetal anomaly screening is to identify potential problems so that parents can make an informed choice and to improve the safety of birth.

Specifically, antenatal screening to identify fetal anomalies should allow women and their partners:

• reproductive choice (a choice about continuing with the pregnancy or choosing termination of pregnancy)
• time to prepare (for termination of pregnancy/postnatal treatment or palliative care/infant disability)
• managed delivery in specialist centre
• intrauterine therapy.

The criteria laid out by Wilson and Jungner (1968)¹⁰²⁰ to appraise the validity of any screening programme are that:

• disorders to be screened for should be clinically well defined
• the incidence of the conditions (individual malformations) should be known
• disorders to be screened for should be associated with significant morbidity or mortality
• effective treatment should be available, e.g. intrauterine treatment, birth managed in a specialist centre, or termination of pregnancy
• there should be a period before onset of the disorder (the antenatal period) during which intervention is possible to improve outcome or allow informed choice
• there should be an ethical, safe, simple and robust screening test, e.g. ultrasound appears safe, ethical and acceptable
• screening should be cost-effective.

However, it is important to note that many of the studies of antenatal screening for fetal anomalies evaluate ultrasound as a suitable test rather than examine the benefits for women and babies of screening for a range of fetal anomalies during pregnancy.

Description of included studies

One systematic review²⁹⁷ including 11 studies, and an additional 12 studies^727–741 were identified from the search. The 12 studies were critically appraised against the same criteria applied to the systematic review. Six studies were excluded either because of incomplete data or irrelevant study populations (e.g. high-risk populations). Details of the inclusion/exclusion process are provided on the accompanying CD-ROM. A new systematic review of all identified primary 17 studies, 11 studies in the systematic review and six newly identified studies, was conducted by the NCC-WCH. [EL = II]

Data from one RCT, nine prospective cohort studies and seven retrospective cohort studies were extracted. Four studies were conducted in the UK, while four were in the USA, four in Scandinavia, two in Belgium, two in Greece and one in South Korea. Details of the included studies are shown in Table 9.1. Meta-analyses of 11 studies on positive and negative likelihood ratios are presented in Figures 9.1 to 9.4.

Table 9.1

Description of included studies and detection rates of structural anomalies by antenatal ultrasound (first and second trimester).

Figure 9.1

Meta-analysis of positive likelihood ratios by routine ultrasound to detect fetal anomalies before 24 weeks.

Figure 9.2

Meta-analysis of negative likelihood ratios by routine ultrasound to detect fetal anomalies before 24 weeks.

Figure 9.3

Meta-analysis of overall positive likelihood ratios by routine ultrasound to detect fetal anomalies.

Figure 9.4

Meta-analysis of overall negative likelihood ratios by routine ultrasound to detect fetal anomalies.

Findings

Detection by RCOG category

Sensitivity (detection rate) for each condition according to the RCOG category⁷⁴² was also sought, and is presented in Table 9.2. Overall sensitivity for lethal anomalies was 83.6%, that for possible survival and long-term morbidity was 50.6%, that for anomalies amenable to intrauterine therapy was 100.0%, and that for anomalies associated with possible short-term/immediate morbidity was 16.1%. The sensitivity varies depending upon each condition.

Table 9.2

Prevalence and detection of congenital anomalies at second-trimester antenatal ultrasound according to RCOG subgroup.

Evidence summary

Second-trimester ultrasound seems to show high specificity but poor sensitivity for identifying fetal structural anomalies. Similarly, this test showed good summary value for positive likelihood ratio but poor negative likelihood ratio. However, these values ranged widely by centre and condition. The 100% detection rate for conditions amenable to intrauterine treatment is anomalous and arises from the fact that these conditions had to be identified before treatment could be considered.

Description of included studies

One review of literature included in an HTA²⁹⁷ and additional four studies^{300,743–746} were identified. However, only one^300,743 from the additional studies was included in this review owing to methodological weakness and incomplete data. [EL = III]

Findings

The review showed that there were relatively few data on screening an unselected or low-risk population, as most studies report results of screening in high-risk populations.²⁹⁷ Results on nuchal translucency measurement are presented later in the soft markers section. The review included five studies of first-trimester anomaly screening but could not draw any conclusion because of the methodological weakness of these studies.

The additional study was published in 1999, although the study did not specify the time when it was conducted.^300,743 Details of the study are presented in Table 9.1. This was a prospective cross-sectional study at a university hospital in the UK, and included 6634 unselected women carrying 6443 fetuses. All women underwent either transabdominal or transvaginal sonography at 11–14 weeks. Nuchal translucency and an anatomical survey were performed. There were six clinicians undertaking these examinations. The incidence of anomalous fetuses was 1.4%, and sensitivity (detection rate) was 59.0% (37/63 (95% CI 46.5% to 72.4%)). The specificity was 99.9%. Positive and negative likelihood ratios were 624.5 and 0.41. When first- and second-trimester scans were combined, the sensitivity was 81.0% (51/63 (95% CI 67.7% to 89.2%).

Evidence summary

There were only a few good-quality studies conducted which examine the diagnostic value of routine ultrasound in the first trimester. Although high specificity and positive likelihood ratio were reported, the sensitivity and negative likelihood ratio reported from a single centre in the UK were at a moderate level.

Routine versus selective ultrasound before 24 weeks

Routine versus no/concealed/selective ultrasound after 24 weeks

Routine versus no/concealed/selective Doppler ultrasound in pregnancy

Serial ultrasound plus Doppler versus selective ultrasound in pregnancy

First- versus second-trimester routine ultrasound in pregnancy

Diagnostic value of fetal echocardiography

Description of included studies

Studies examining the diagnostic value of fetal echocardiography on low-risk or unselected populations were reviewed. One systematic review including five studies was identified plus two additional studies.^749–751 A description of these studies is presented in Table 9.3.

Table 9.3

Diagnostic value of fetal echocardiography: description of included studies and reported sensitivity and specificity.

Effectiveness of routine use of fetal echocardiography

Nuchal translucency measurement

Description of included studies

Studies examining the diagnostic value of nuchal translucency measurement of low-risk or unselected populations on detecting cardiac anomalies were reviewed. One systematic review including eight studies and four additional studies was identified.^754–758 Since studies used different cut-off points, meta-analysis of these twelve studies to obtain summary likelihood ratios was conducted (Table 9.4 and Figures 9.5 and 9.6) Neither RCTs nor quasi-randomised controlled trials were identified to address the effectiveness of routine use of this measurement on clinical outcomes of women and their babies.

Table 9.4

Diagnostic value of nuchal translucency measurement on fetal cardiac anomaly.

Figure 9.5

Meta-analysis of positive likelihood ratios by nuchal translucency measurement to detect fetal cardiac anomalies.

Figure 9.6

Meta-analysis of negative likelihood ratios by nuchal translucency measurement to detect fetal cardiac anomalies.

Alpha-fetoprotein to detect neural tube defects

Description of included studies

The first study was a review²⁹⁷ [EL = 2++] which focused on women’s views and experiences of antenatal ultrasound. As the topic was very wide, it was decided to limit the review to studies where antenatal ultrasound was used for any purpose and direct data were obtained from pregnant women. Studies and reviews about prenatal screening and diagnosis were excluded. After a broad initial search to identify material related to women’s views in all screening and diagnostic tests, studies related to antenatal ultrasound use were selected after going through their abstracts. A series of six questions was prepared, targetting: (i) women’s knowledge about ultrasound and what a scan can do; (ii) women’s value about scans; (iii) her views about how ultrasound is conducted; (iv) impact of the result; (v) psychological impact of ultrasound; and (vi) wider impact of ultrasound on society. Studies were tabulated according to the question asked and data entered accordingly.

In the second study⁷⁶¹ qualitative interviews were conducted to determine women’s experiences and responses to detection of a minor structural variant, the choroid plexus cyst (CPC), in their fetuses on prenatal ultrasound. Thirty-four pregnant women with isolated CPC detected during a mid-trimester scan who had already been counselled by their physicians regarding the findings at a university-based hospital in the USA were enrolled for the study. Interviews lasting approximately 15 minutes were conducted by a trained research assistant or nurse clinician at 24 weeks of pregnancy, and no information was given about CPCs by the research team. The interview included both open-ended and more specific questions, and all were audiotaped and transcribed verbatim. Common themes were identified, and several categories of responses identified for each theme. Initial validation was undertaken by an independent qualitative study consultant not involved in the research. The t-test was used for comparing means and χ² for categorical variables. The results are reported as mean ± standard deviation. [EL = 3]

The aim of the third study⁷⁶² was to evaluate parental anxiety after diagnosis of a congenital malformation and to assess whether counselling by a consultant paediatric surgeon and a neonatal nurse practitioner could decrease parents’ psychological distress. Participants were all parents attending a fetal medicine unit in the UK with an antenatal diagnosis of surgical anomaly (principally abdominal wall defects and gastrointestinal and thoracic anomalies). Women unable to read English and those booked to give birth somewhere else were excluded. Anonymous questionnaires were used to gain information as well as the Spielberger State-Trait Anxiety Inventory (STAI) for measuring anxiety levels. The STAI consists of two parts – the STAI-S score measuring anxiety at the time of completing the inventory, and the STAI-T score measuring the inherent trait anxiety levels. Participants were asked to complete STAI after ultrasound at the fetal centre. Then each couple had a detailed consultation with the paediatric consultant and the clinical nurse specialist. Before leaving, the subjects were given a second STAI and asked to complete and return within 1 week. A control group comprising pregnant women with a normal ultrasound scan and uncomplicated pregnancy was recruited and asked to complete STAI as the other group. Non-parametric tests were used for comparison, and data are quoted as medians and interquartile ranges (IQRs). [EL = 3]

Findings

In the first study,²⁹⁷ a total of 82 reports representing 64 studies were selected (including five studies which were added later). There was wide variation among the selected studies in terms of questions addressed, methods used, and when and where they were conducted. The studies were not graded in terms of research quality or removed because of poor quality, although many had problems of design and reporting. This was done because, in spite of poor quality, these studies gave useful information. The main findings of the review are discussed below.

Antenatal ultrasound is very attractive to pregnant women and their partners as it provides early visual confirmation of pregnancy, direct contact with their baby and reassurance about fetal wellbeing. At the same time, these features may augment the potential for feelings of anxiety, shock and disappointment when the scan shows a problem.

Recent trends in the use of ultrasound have led to more findings of uncertain clinical importance, and this is likely to have important psychological and social consequences for women.

Although it was reported in earlier studies that some women feared that ultrasound might harm their babies, there is a paucity of evidence about it from the later studies.

Reports of a reduction in anxiety after ultrasound examination are likely to reflect increased anxiety before the scan rather than a real benefit.

No reliable evidence is available for any positive health behaviour (e.g. reduced smoking) as a consequence of antenatal ultrasound.

None of the trials comparing ultrasound use with no ultrasound use has looked at its social and psychological impact on parents and babies.

In general, participants in the second study⁷⁶¹ were college educated (mean years of education 16.6 ± 2.5), married (85.7%), employed (100%) and had private insurance (97%). The mean maternal age was 32.2 ± 5.2 years. About 60% were primiparous and 80% had a planned pregnancy. Women’s responses have been organised into categories as listed below.

• Diagnostic situation. Mean gestational age at CPC detection was 18.86 ± 1.29 weeks. The majority of the participants (71%) were informed about CPC by an attending or local obstetrician at the conclusion of the ultrasound examination, and 35% of women were shown the CPC on ultrasound.
• Accuracy of knowledge. Most of the women (79%) had never heard of CPC before the diagnosis. When asked about the significance of the CPC, 82% felt that it was probably benign, 71% expressed it is a marker for trisomy, and 53% mentioned that it could be both. Among those who expressed it as a marker for trisomy, 79% understood that other factors (maternal age, serum markers) also influenced the probability of trisomy. Women with positive serum screening results were less likely to describe CPC as benign compared with women with a normal serum screen (OR 0.04, 95% CI 0.004 to 0.36; P < 0.001). No statistically significant difference was observed between the older women (> 34 years) and younger ones.
• Information seeking. Seventy-seven percent of women reported seeking additional information about CPCs beyond that given by their provider at the original scan, with the most common source being the internet. When asked about the usefulness of this additional information, 62% found it more useful than the primary information given at the time of ultrasound screening.
• Subsequent testing. The majority of women (65%) already had a serum screening test before detection of CPCs. After detection of an isolated CPC and in spite of accurate counselling about low risk, three women (9%) sought diagnostic tests purely for reassurance.
• Affective responses. When asked in an open-ended way to describe their emotions, 88% of women described an intensely negative immediate reaction, with most (68%) reporting their initial reaction as temporary. But only half of the women with a reassuring serum screen and none with an abnormal serum screen described their reaction as temporary. Sixty-eight percent of women revealed that they continued experiencing negative emotions even after receiving the diagnostic tests results, but neither increased maternal age nor visualisation of CPC on ultrasound were associated with persistence of the initial negative response. The later emotional responses included anxiety (23.5%), shock/grief (26.5%), decreased attachment (14.7%), decreased pleasure in pregnancy (14.7%), and thoughts of abortion/miscarriage (11.8%), confusion (8.8%), guilt (2.9%) and fear (5.9%).

Fifty-six pregnant women (subjects 26, control 30) completed the questionnaire in the third study.⁷⁶² The most common congenital malformation present was gastroschisis followed by diaphragmatic hernia and cystic adenomatoid malformation. Maternal age was significantly lower in subjects (median 26.5 years) than in the control group (median 32 years) (P = 0.006).

No significant difference was found between STAI-T scores of subjects and controls. No correlation was found between the score and maternal age or social class, or between maternal and paternal scores.

STAI-S scores of subjects were significantly higher than those of controls before paediatric consultation (P = 0.0004), but not after (P = 0.31). There was a significant reduction in the anxiety levels of both subjects (mothers and fathers) after consultation (on comparing their scores before and after paediatric consultation) (P = 0.01 for mothers, P = 0.006 for fathers). After grouping the subjects into fetal diagnostic groups, a significant decrease in anxiety levels was found for those with anterior abdominal defects but not with cystic adenomatoid malformation. No correlation was found between the scores and maternal age.

The study showed that there was a high anxiety state in both prospective mothers and fathers of fetuses diagnosed with congenital malformations on ultrasound which is over and above that associated with pregnancy. Counselling by specialist staff reduced levels of parental anxiety significantly.

Evidence summary

Results from a well-conducted structured review show that visual confirmation of fetal wellbeing is the primary reason why women seek ultrasound during pregnancy. There is a lack of evidence regarding its other benefits and harms.

Evidence from a qualitative study indicates that detection of an isolated choroid plexus cyst on antenatal ultrasound leads to negative emotions and anxiety in the majority of women, who then seek additional information from other sources. In spite of reassurance in the form of a negative serum screening test for Down’s syndrome, a few women also opt for an invasive test for confirmation.

Detection of surgically treatable congenital anomalies on antenatal ultrasound led to increased anxiety levels in the parents but counselling by specialist staff helped to alleviate it significantly.

Health economics evidence

In the NICE clinical guideline on diabetes in pregnancy⁶³⁶ an economic model was developed to compare the cost-effectiveness of screening for congenital cardiac malformations using a four chamber ultrasound scan versus the four chamber plus outflow tracts view. This was considered to be important because women with diabetes are at increased risk of having a baby with a cardiac malformation. It was felt that this model was also relevant for the antenatal care guideline and therefore it was adapted for the antenatal care population. The results are summarised here; futher details are provided in Appendix E.

The baseline analysis suggested that the four chamber plus outflow tracts view has an ICER of £24,000 relative to the four chamber view alone. This falls within the borderline cost-effectiveness range of £20,000 to £30,000 per QALY used by NICE.

For the health economics evidence for the combined Down’s syndrome and structural anomalies screening, please see Section 9.2 (Screening for Down’s syndrome)

GDG interpretation of evidence (screening for structural anomalies)

First-trimester studies

Findings

The first-trimester studies have been divided into the anomalies they looked at.

Down’s syndrome and other chromosomal anomalies

Three studies evaluated the serum combined test^768–770 and three fetal nasal bone on ultrasound.^771–773 These studies have been tabulated in Tables 9.5 and 9.6, respectively. The additional six studies on evaluation of fetal nasal bone^{771,773–777} are given in Table 9.7.

Table 9.5

First-trimester screening for Down’s syndrome and other chromosomal anomalies using the serum combined test.

Table 9.6

First-trimester screening for Down’s syndrome and other chromosomal anomalies using nasal bone evaluation.

Table 9.7

First-trimester screening for Down’s syndrome using nasal bone evaluation – additional studies.

Results from a good-quality cohort with large sample size⁷⁶⁸ showed the serum combined test to have a detection rate of 92.6% at a false positive rate of 5.2% for the detection of Down’s syndrome, and a slightly lower detection rate for trisomy 18 or 13 and other chromosomal anomalies. Similar results were observed in another study,⁷⁷⁰ while the third study⁷⁶⁹ showed a lower detection rate but higher FPR for the combined test.

Conflicting results were seen for the diagnostic accuracy of fetal nasal bone evaluation (Table 9.6). While one study⁷⁷² showed fetal nasal bone to increase the detection rate of Down’s syndrome from 90% to 93% (fixed FPR 5%) compared with using the combined test only, the other study⁷⁷¹ showed it to have very poor diagnostic value. The third study⁷⁷³ had variable diagnostic accuracy results for the selected and unselected population.

Results from the additional six studies evaluated for fetal nasal bone were also inconclusive and wide variation was observed in them (Table 9.7). In two studies^779,780 it improved the detection rate compared with using the serum combined test alone, but in one study⁷⁷⁵ there was a reduction in the detection rate. The sensitivity and detection rate of fetal nasal bone alone in the rest of the studies varied from 32% to 70%.

From these nine included studies on nasal bone characteristics, various factors have been identified which seem to influence the finding of absent nasal bone on first-trimester ultrasound. These factors are experience and training of the ultrasound operator, gestational age at which ultrasound is conducted (ideally CRL to be more than 45 mm as ossification of nasal bone starts after this), type of population screened (low-risk or high-risk), and marker used for diagnosis (complete absence or hypoplasia of the nasal bone).

Down’s syndrome only

The diagnostic accuracy results of the three included studies for the serum combined test were similar (Table 9.8). While one multicentre study⁷⁷⁸ found a detection rate of 79.6% at an FPR of 2.9%, the other two showed detection rates of 90.3% and 82% at a fixed FPR of 5%.

Table 9.8

First-trimester screening for Down’s syndrome only.

Second-trimester screening

Compared with the first trimester only and first and second trimester together, few studies were found relating to serum screening tests done exclusively in the second trimester. Good-quality serum marker studies comparing both the first- and second-trimester tests have been grouped under the next section on combined first- and second-trimester screening. A number of studies were identified which evaluated the use of ultrasound for identifying ‘soft markers’ – nuchal fold thickening, choroid plexus cyst, echogenic intracardiac foci, renal pyelectasis and shortening of femur, but the general quality was low (EL = III).

Five studies were selected for inclusion in this section – three meta-analyses, one prospective cohort study and one retrospective cohort study. As these studies were quite different from each other, their data could not be tabulated and they have been described in a narrative manner.

The second-trimester studies have been further divided into the anomalies they looked at:

Combined first- and second-trimester studies

Description of included studies

Four good-quality studies were included: three prospective cohort studies^785–787 and one nested case–control study.³¹⁶ All the studies were multicentred with clearly defined objectives. One of the two studies with a selected population had first-trimester screen-positive and screen-negative women together in its sample population.⁷⁸⁷ In all studies the screening test and monitoring of quality measures were adequately explained. The reference test in all was a validated one (karyotyping/postnatal assessment/pregnancy records) (Table 9.9).

Table 9.9

First- and second-trimester screening for Down’s syndrome only.

Evidence summary

Findings from a descriptive study [EL = 3] shows that the integrated test, when implemented in practice, seems to be generally acceptable to pregnant women opting for it and results in a good uptake. But 25% of these women failed to attend for the second component of the test and this required sending them reminders.

Description of included studies

Two studies were identified which used modelling as a way of comparing different screening tests for Down’s syndrome detection.

To demonstrate the potential value of three-stage sequential screening for Down’s syndrome, DR and FPR were estimated by multivariate Gaussian modelling using Monte Carlo simulation.⁷⁸⁹ UK data were used for modelling. The protocol is known as ‘contingent screening’ and involves measuring free β-hCG and PAPP-A in all pregnant women at 10 weeks in the first stage. Those with low risk were screened negative at this stage, the remainder underwent NT measurement in the second stage and the risk was reassessed (for combined test). After the second stage, those with low risk were screened negative and those with very high risk were offered diagnostic tests. In the third stage, women with intermediate risk received a second-trimester quaduple test. Risk was reassessed according to the integrated test and high-risk women were offered diagnosis. [EL = III]

Using Monte Carlo simulation for modelling, another study⁷⁹⁰ compared the integrated test in three policies for screening: (i) integrated screening for all women; (ii) sequential screening (based on first-trimester tests, high-risk pregnancies to be diagnosed and remaining to undergo integrated test); and (iii) contingent screening.

Detection and false positive rates were estimated based on the data from a large cohort (nested case–control study) done in the UK. [EL = III]

Findings

The first modelling study suggested that, with full adherence to a three-stage policy, an overall detection rate of nearly 90% and a false positive rate of below 2% can be achieved. About two-thirds of the women can be screened on the basis of first-trimester biochemistry alone and about 80% by the combined test. The DR for first-trimester screening is about 60%.

This protocol allows most of the Down’s syndrome pregnancies to be detected in the first trimester. Furthermore, it provides an efficient way of screening for Down’s syndrome where NT measurements cannot be performed in all women owing to scarcity of resources. However, it requires the selection of four different cut-offs during the three stages, each of which will affect the overall performance. Selecting a set of appropriate cut-offs is therefore complex and difficult to practise. The psychological impact of pregnant women possibly receiving four different results also needs to be evaluated.

The second modelling study concluded that integrated screening had the best screening performance. As the first-trimester test FPR was decreased, the performance of the other two policies approached that of the integrated screen. Setting the first-trimester risk cut-off to ≥ 1 in 300 with a fixed DR of 90%, sequential and contingent screening gave overall FPRs of 2.3% and 2.4%, respectively, and 66% of affected pregnancies were detected by the first-trimester tests. The integrated test on all women gave an FPR of 2.2%.

If pregnancies with a first-trimester risk of ≤ 1 in 2000 are classified screen negative and receive no further testing, then 99.5% of women with sequential screening or 30% with contingent screening would proceed to integrated screening.

Description of included studies

A multicentre RCT⁷⁹¹ in maternity care units affiliated to eight Swedish hospitals was carried out with an aim of comparing the effectiveness of two screening policies for detecting Down’s syndrome: routine ultrasound scan at 12–14 weeks by NT (12 week policy) versus routine ultrasound at 15–20 weeks (18 week policy). An unselected population with well-defined eligibility criteria was involved. After taking informed consent, the population was randomised block-wise at the level of maternity units using internet-based software. Appropriately trained operators carried out the ultrasound examination. Karyotyping was offered to all women with increased risk of Down’s syndrome (> 1 : 250 based on NT in the first group and on maternal age in the second), detection of a structural anomaly on scan, history suggestive of increased risk, or preference/desire of the woman due to worry. Follow-up of results (karyotyping, pregnancy outcome) was adequate. Evaluation of the primary outcome (number of babies born alive at ≥ 22 weeks with Down’s syndrome) and secondary outcomes (total number of babies born with Down’s syndrome, number of babies born with other chromosomal abnormalities, number of pregnancy terminations for Down’s syndrome, and rate of invasive tests for fetal karyotyping) was done using intention-to-treat analysis. The sample size was calculated to detect a difference of 0.1% in liveborn Down’s syndrome cases between the two groups at a 5% significance level with 90% power. χ² tests (for proportions) and Student’s two-sample test (for continuous data) were used for comparison. [EL = 1+]

The nested case–control study³¹⁶ has been discussed in the combined first- and second-trimester screening section above. Apart from evaluating the screening performance of various tests, it also examined their safety in terms of number of unaffected fetal losses per 100 000 women screened, and number of Down’s syndrome pregnancies detected for each procedure-related unaffected fetal loss. Both calculations were done at different detection rates. [EL = 2+]

A decision-analysis model⁷⁹² was used to compare five screening strategies: (i) first-trimester combined screen; (ii) second-trimester quadruple screen; (iii) second-trimester triple screen; (iv) integrated screen; and (v) sequential screen. A hypothetical cohort of 1 000 000 women below 35 years was analysed assuming the entire cohort would present for antenatal care before 10 weeks and accept prenatal screening for Down’s syndrome. After a positive triple or quadruple test, genetic sonogram would be performed and then prenatal diagnosis would be available. Four separate outcomes were examined: (i) overall cost-effectiveness; (ii) Down’s syndrome cases detected; (iii) Down’s syndrome live births averted; and (iv) euploid losses from invasive procedures. [EL = 3]

Clinical parameters used for modelling were synthesised from review of published data (mainly UK data). The prevalence of Down’s syndrome at 10 weeks of gestation was estimated as 1 in 595 pregnancies, with a baseline live birth rate of 1 in 1030. Seventy percent of women were estimated to opt for invasive diagnostic techniques after a positive screening test, and 90% to opt for termination of affected pregnancies. Baseline fetal loss after amniocentesis and chorionic villus sampling were estimated to be 0.9% and 1.6%, respectively, but this was also varied over a range. Spontaneous fetal loss of euploid pregnancies was estimated at 1% between 10 and 14 weeks, and an additional 1% between 15 weeks and delivery. The screening performance of various tests was derived from published data. [EL = 3]

Details of the fourth study³¹⁵ have already been discussed in the second-trimester screening section above.

The last study⁷⁹³ analysed the database from the FASTER trial (a multicentre prospective trial in the USA) to determine whether there is an NT measurement above which immediate invasive testing should be offered, without waiting for serum testing and computerised aneuploidy risk assessment. Pregnant women were eligible for inclusion if they were above 16 years of age, had a singleton pregnancy and a CRL of 36–79 mm (gestation 10 weeks 3 days to 13 weeks 6 days) at the time of first-trimester sonography for NT. Cases with cystic hygroma were excluded. NT was measured in the first trimester using a standardised protocol by specially trained ultrasonographers at the same time as when serum levels of PAPP-A and β-hCG were obtained. At 15–18 weeks, a quadruple serum screening test was also obtained, but the present study used only the risks as assessed from the first-trimester tests. A formal quality control programme was used throughout the study. [EL = 2+]

Findings

In the multicentre RCT a total of 39 572 women were randomised in the two groups (19 796 in the 12 week group, 19 776 in the 18 week group). Demographically the two groups did not differ in mean age, mean parity or other characteristics. In the 12 week group, NT measurement could not be carried out in 9% of the population owing to increased CRL or fetal demise, but was successfully measured in 96% of the remaining population. The prevalence of Down’s syndrome during the study period was 0.25% (98/39 572). The results are as follows:

In the second study, the safety of various tests was evaluated at a fixed DR of 85%. The integrated test had about one-fifth the number of fetal losses when compared with the combined and quadruple test, and half that of the serum integrated test. The number of cases of Down’s syndrome detected for each fetal loss was almost three times higher with the integrated test when compared with the combined and quadruple test.

The modelling study found sequential screening to be the most cost-effective. Compared with other screens, it was shown to detect antenatally most cases of Down’s syndrome and avert most live births of affected fetuses. But it also had the highest number of euploid losses due to diagnostic procedure. From the point of view of safety, the integrated screen performed the best with the lowest euploid losses. The addition of a genetic sonogram to the triple and quadruple screens increased the cost but brought the euploid losses down to very low levels.

The meta-analysis concluded that the number of fetal losses per case diagnosed when identified as an isolated ‘soft marker’ abnormality on ultrasound was highest with choroid plexus cysts (4.3) and lowest with thickened nuchal fold (0.6).

For others the values were femur length (1.2), humerus length (1.9), echogenic bowel (1.0), echogenic cardiac foci (2.0) and renal pyelectasis (2.6)

In the NT study, the sample population included 36 120 pregnancies with complete first-trimester results. The mean and median NT measurements increased from 10 through 13 weeks and there was considerable variation in the proportion of cases with NT ≥ 2.0 mm at each gestational week, but there was minimal gestational age variation in NT once a threshold of 3.0 mm was passed.

On comparison of outcome of pregnancies based on the various NT cut-offs, the following results were observed:

There were 32 women with NT ≥ 4 mm, and the addition of first-trimester serum markers to NT measurements did not reduce the final risk in any patients. In contrast, for patients with NT ≥ 3 mm, subsequent addition of serum markers reduced the final risk to less than 1 : 200 in only 8% of cases (ten women). For women with NT ≥ 2 mm, a large number of women (49%) had their risk reduced to less than 1 : 200 by the addition of first-trimester test results.

The authors concluded that the use of 4.0 and 3.0 mm cut-off of NT measurement for estimating pregnancies at risk of Down’s syndrome would lead to just 0.09% and 0.4%, respectively, of the population being subjected to invasive testing based on the two cut-offs. By waiting for serum assays and computerised risk assessment, no benefit (0%) was observed in the women with NT ≥ 4 mm and only a minimal benefit (8.0%) in women with NT ≥ 3 mm, that is, who had their final risk reduced to less than 1 in 200. This will increase the screen-positive rate for the whole population by a very small proportion, but will be beneficial in providing immediate results to the healthcare providers and reducing anxiety for the pregnant women.

Evidence summary

Reported evidence shows that the combined test in the first trimester has good diagnostic accuracy for Down’s syndrome and other chromosomal anomalies.

Among the currently available second-trimester serum tests, the quadruple test seems to have the best screening performance.

There is high-quality evidence to indicate that combining results of first- and second-trimester screening tests improves the diagnostic performance for Down’s syndrome and other chromosomal anomalies and is better than when either of them is used alone.

The integrated test seems to have a higher detection rate and a lower FPR compared with other currently used combined screening tests.

There is little evidence on the diagnostic value of other policies of combining first- and second-trimester results.

There is conflicting evidence regarding the performance of nasal bone ultrasound assessment as a screening tool for Down’s syndrome.

‘Soft markers’ on ultrasound have low sensitivity and LR+ when seen individually, except for nuchal fold thickening. When found in association with other anomalies, they seem to improve the diagnostic value but the evidence is not strong.

Retrospective analysis of a database from a high-quality prospective study shows that an NT measurement of 3 mm or more in the first trimester (any gestational age) identified the majority of pregnant women with Down’s syndrome, and increased the screen-positive rate/risk of invasive testing by only a small fraction compared with first-trimester risk evaluated by the combined test.

Description of included studies

A systematic review⁷⁹⁴ was carried out to understand the psychosocial aspects of genetic screening of pregnant women and newborns. The review aimed to address five broad questions concerned with: (i) knowledge; (ii) anxiety; (iii) other emotional aspects; (iv) factors associated with participation in the programmes; and (v) long-term sequelae of the results. Any genetic screening programme aimed at pregnant women or newborn babies was included. Both comparative and descriptive studies which reported data collected directly from pregnant women or parents were included. There were no geographical or methodological limits except that studies asking hypothetical questions, case reviews and those where ultrasound was done to detect structural anomalies only (and did not include chromosomal anomalies) were excluded. Five electronic databases and two journals were hand searched. The retrieved articles were equally divided among the five authors for quality assessment and data extraction, and these processes were completed using well-defined criteria and validated forms. A new quality score was devised for quality assessment which was not found to be useful later on. Literature on ‘other emotional aspects’ and ‘long-term sequelae’ was too fragmented (except in neonatal screening programmes) for useful conclusions to be drawn. [EL = 2++]

A prospective cohort study⁷⁹⁵ was carried out in four antenatal clinics in Australia to assess informed choice in pregnant women to participate in second-trimester serum screening using a validated measure, and to compare anxiety levels in women who are well informed compared with those who are poorly informed. Participants included pregnant women at between 8 and 14 weeks of gestation attending at their first prenatal visit and with sufficient English to complete a written questionnaire. Written and oral information was provided to all participants as per the existing hospital policy. Informed choice was measured by Multidimensional Measure of Informed Choice (MMIC), a validated measure of informed choice which assesses knowledge and attitude dimensions and also confirms whether a woman’s participation in a screening test matches her expressed attitude towards it. The Hospital Anxiety and Depression Scale (HADS) was used to measure anxiety and this scale specifically distinguishes between anxiety and depression. Both the scales were administered at the booking visit and HADS was repeated at 20 weeks (after participation in the test) and at 30 weeks using postal questionnaires. [EL = 2+]

In the third study, a smaller sample drawn from the RCT described above⁷⁹¹ was used to study the effect of screening on women’s anxiety during pregnancy and after birth, with a specific focus on worries about the health of the baby.⁷⁹⁶ The 12 week group was the intervention group and the 18 week group acted as the control. The principal outcome of women’s worries about the ‘possibility of something being wrong with the baby’ was measured by the Swedish version of the Cambridge Worry Scale questionnaire including 16 items of common concerns during pregnancy. The State-Trait Anxiety Inventory (a validated tool for evaluating general anxiety) and the Edinburgh Postnatal Depression Scale (validated for evaluating anxiety in the antenatal/postnatal period) were also used. Information was collected at three different timings – the first questionnaire was filled at the antenatal clinic, the second was sent at 24 weeks of gestation (mid-pregnancy), and the last was posted 2 months after delivery. The same instruments were used for all three questionnaires. [EL = 3]

A cross-sectional survey⁷⁹⁷ was carried out in three Canadian cities to investigate the relationship between MSS use and maternal attachment to pregnancy following the receipt of a favourable result (i.e lowered risk ratio). Building on the preliminary evidence that MSS results are not reassuring to women, it was predicated that favourable MSS results would not be sufficient to allow women to move beyond tentative pregnancy stage. Hence it was hypothesised that:

• there would be no difference in prenatal attachment between women receiving favourable amniocentesis results (amniocentesis group) and who opt against testing (no testing group)
• there would be a lower level of attachment among women who receive favourable MSS results and did not undergo amniocentesis (MSS group) compared with the other two testing groups, and this difference would be evident in the second and third trimesters.

Participants included high-risk pregnant women (maternal age > 35 years) who opted for MSS or amniocentesis or did not opt for any testing. Informational posters were placed at various places (physician offices, laboratories, maternity stores), and interested women who met the eligibility criteria were enrolled. The instrument used to collect information was a self-administered questionnaire by mail, and prenatal attachment was measured by a 21-item Prenatal Attachment Inventory (PAI) score (construct validity and reliability of this scale were established). The three groups were compared using ANOVA and ANCOVA for statistical analysis. [EL = 3]

To address the question of whether there are social and ethnic inequalities in the offer and uptake of prenatal screening and diagnosis in the UK, a systematic review⁷⁹⁸ was carried out employing a broad search strategy. In order to address the review question, studies were assessed in terms of:

• utilisation – number of women screened as a proportion of those eligible
• offer – number of women offered screening as a proportion of those eligible
• uptake – number of women screened as a proportion of those offered screening.

Studies were reviewed and summarised by one reviewer. Two key aspects of the studies were assessed independently by two reviewers and summarised as indicators of quality – non-participation rate and whether the distinction between utilisation, offer and uptake was recognised in the study. Owing to heterogeneity, meta-analysis could not be performed. [EL = 2+]

A prospective descriptive study⁷⁹⁹ was carried out in two UK district hospitals to find out reasons for lower uptake of screening tests in women from minority ethnic groups and socio-economically disadvantaged sections of society. Screening uptake was evaluated from hospital records. Attitudes towards undergoing the test were assessed by women’s responses to a structured question with four items. Knowledge about the test was assessed using an eight-item questionnaire deemed important in professional guidelines for informed consent in screening. Choices were classified as ‘informed’ depending on the consistency between test uptake, women’s attitude towards the test, and their knowledge about it. [EL = 3]

Another cross-sectional survey⁸⁰⁰ was carried out in six UK maternity units (three in Scotland, three in England) to ascertain by means of a structured questionnaire women’s preference for type of screening test. Pregnant women attending antenatal clinics were asked to put in order of preference four different approaches for screening (all with FPR of 5%):

• first-trimester testing – 90% detection with results available in 1 hour
• first-trimester testing – 90% detection with results within 2–3 days (combined test)
• first-trimester plus second-trimester detection – 93% detection and results within 2–3 days of second test (integrated test)
• second-trimester testing – 75% detection and results available within 2–3 days. [EL = 3]

Findings

In the first systematic review 106 out of 288 identified studies met the eligibility criteria – 78 concerned with antenatal screening and 28 with neonatal screening. Only results pertaining to antenatal screening programmes are discussed below. Findings from antenatal carrier testing for cystic fibrosis and other diseases prevalent in minority ethnic groups are similarly also not discussed here.

Most of the antenatal studies were descriptive and only 33% (26/78) were RCTs or comparative. A questionnaire was the most common instrument used to collect data (in 79% of studies), either alone or together with other methods. Only 16 studies (20%) included both women who were tested and those who were not. Fifty-four studies were concerned with screening for Down’s syndrome and other chromosomal anomalies. The sample size of studies varied from ten to 6442 participants. Data were collected after the test results in 40 studies, and in just three studies was it collected at three different times – before test, after test and after test results. A large number of studies assessed knowledge (64.6%), anxiety (46.8%) or attitudes/beliefs (46.2%). Thirty-four antenatal studies (43.6%) had an apparent input from a psychologist or a social scientist. The various findings have been divided into three sections.

Evidence summary

There is high-quality evidence to indicate that pregnant women do not have sufficient knowledge to make the informed decisions that need to be made regarding Down’s syndrome screening and they find the concept of risk calculation particularly difficult to understand. Providing them with more information does not lead to an increase in their anxiety level.

Good evidence from a cohort study shows that women taking part in prenatal screening programme are inadequately informed regarding aspects of testing and the further pathway of management when an increased risk is identified.

Results from a cross-sectional study indicate that women undergoing a serum screening test for Down’s syndrome develop less attachment for the baby owing to the uncertainty surrounding interpretation of the test result.

Evidence from a review of literature shows that pregnant women from South Asia have a lower rate of uptake, acceptance and utilisation of screening tests.

For the screening tests in general, white women and women from socio-economically advantaged sections of society have a higher uptake, better knowledge, more consistency of actions related to positive attitude, and a higher rate of informed decision making when compared with women from South Asia and socio-economically disadvantaged sections of society.

Antenatal screening for Down’s syndrome

A systematic search of the literature was conducted to identify economic evaluations of screening for Down’s syndrome. The search identified 132 abstracts, of which 40 full papers were retrieved for further consideration. Six studies are included in the review.

One study⁸⁰¹ was conducted to examine the performance of integrated Down’s syndrome screening (first- and second-trimester measurements integrated into a single screening test) when ratios of the levels of the same serum markers measured in both these trimesters (cross-trimester ratios) are added as new screening markers. The addition of cross-trimester ratios to an integrated test significantly improves the efficacy and safety of prenatal screening for Down’s syndrome. So, the addition of cross-trimester ratios is cost-effective and could be usefully introduced into screening programmes.

Another UK study⁸⁰² was conducted to compare the effects, safety and cost-effectiveness of antenatal screening strategies. The main outcomes of the study were the number of liveborn babies with Down’s syndrome, miscarriages due to chorionic villus sampling or amniocentesis, healthcare costs of the screening programme, and additional costs and additional miscarriages per additional affected live birth prevented by adopting a more effective strategy. Compared with no screening, the additional cost per additional liveborn baby with Down’s syndrome prevented was £22,000 for measurement of NT. The cost of the integrated test was £51,000 compared with the measurement of NT. All other strategies were more costly and less effective, or cost more per additional affected baby prevented. Depending on the cost of the screening test, the first-trimester combined test and the quadruple test would also be cost-effective options. The main conclusions of the study were that the choice of screening strategy should be between the integrated test, first-trimester combined test, quadruple test or NT measurement depending on how much service providers are willing to pay, the total budget available and values on safety. Screening based on maternal age, the second-trimester double test and the first-trimester serum test was less effective, less safe and more costly than these four options.

One HTA study³¹⁶ was conducted to identify the most effective, safe and cost-effective method of antenatal screening for Down’s syndrome using NT, maternal serum and urine markers in the first and second trimesters of pregnancy and maternal age in various combinations. The cost-effectiveness analysis showed that the screening using the integrated test is less costly than might be expected because the extra screening costs tend to be offset by savings in the cost of diagnosis arising from the low FPR. It was estimated that to achieve an 85% detection rate the cost to the UK NHS would be £15,300 per Down’s syndrome pregnancy detected. The corresponding cost of using the second-trimester quadruple test would be £16,800 and using the first-trimester combined test it would be £19,000.

Antenatal screening for Down’s syndrome + structural anomalies

One HTA²⁹⁷ study was conducted and one of the aims of this study was to refine and update a decision model of cost-effectiveness of options for routine scanning for fetal anomalies. The initial eight options considered were reduced to three dominating options: one second-trimester scan alone, one third-trimester scan alone and a combination of the one second-trimester scan followed by one third-trimester scan. More representative cost data are required before precise estimates of the additional costs and benefits of alternative options can be determined. Also, it is clear from the analysis that one second-trimester analysis scan emerged as a clear reference case, being one of the cheapest options yet still detecting a significant number of anomalies. When termination is acceptable and available, a third-trimester scan alone or the combination of one second- with one third-trimester scan, although comparable in economic terms, may be impractical because of the delay in identifying anomalies.

Another study⁸⁰³ was conducted to compare the cost-effectiveness of different programmes of routine antenatal ultrasound screening to detect four key fetal anomalies: serious cardiac anomalies, spina bifida, Down’s syndrome and lethal anomalies. The study showed that there was a substantial overlap between the cost ranges of each screening programme demonstrating considerable uncertainty about the relative economic efficiency of alternative screening programmes consisting of one second-trimester ultrasound scan. The cost per target anomaly detected (cost-effectiveness) for this programme was in the range £5,000–109,000, but in any 1000 women it would also fail to detect between 3.6 and 4.7 target anomalies. The model highlighted the weakness of the available evidence and demonstrated the need for more information both about current practice and costs.

Finally, a study⁸⁰⁴ was conducted in the UK to determine the most clinically and cost-effective policy of scanning and screening for fetal anomalies in early pregnancy. The number of anomalies detected and missed, the number of iatrogenic losses resulting from invasive tests, the total cost of strategies and the cost per abnormality detected were compared between strategies. First-trimester screening for chromosomal anomalies costs more than the second-trimester screening but results in fewer iatrogenic losses. Strategies which include a second-trimester ultrasound scan result in more anomalies being detected and have lower costs per anomaly detected.

GDG interpretation of evidence