First trimester ultrasound tests alone or in combination with first trimester serum tests for Down's syndrome screening

S Kate Alldred; Yemisi Takwoingi; Boliang Guo; Mary Pennant; Jonathan J Deeks; James P Neilson; Zarko Alfirevic

doi:10.1002/14651858.CD012600

. 2017 Mar 15;2017(3):CD012600. doi: 10.1002/14651858.CD012600

First trimester ultrasound tests alone or in combination with first trimester serum tests for Down's syndrome screening

S Kate Alldred ^1,^✉, Yemisi Takwoingi ², Boliang Guo ³, Mary Pennant ⁴, Jonathan J Deeks ², James P Neilson ⁵, Zarko Alfirevic ¹

Editor: Cochrane Pregnancy and Childbirth Group

PMCID: PMC6464518 PMID: 28295158

Abstract

Background

Down's syndrome occurs when a person has three, rather than two copies of chromosome 21; or the specific area of chromosome 21 implicated in causing Down's syndrome. It is the commonest congenital cause of mental disability and also leads to numerous metabolic and structural problems. It can be life‐threatening, or lead to considerable ill health, although some individuals have only mild problems and can lead relatively normal lives. Having a baby with Down’s syndrome is likely to have a significant impact on family life.

Non‐invasive screening based on biochemical analysis of maternal serum or urine, or fetal ultrasound measurements, allows estimates of the risk of a pregnancy being affected and provides information to guide decisions about definitive testing.

Before agreeing to screening tests, parents need to be fully informed about the risks, benefits and possible consequences of such a test. This includes subsequent choices for further tests they may face, and the implications of both false positive and false negative screening tests (i.e. invasive diagnostic testing, and the possibility that a miscarried fetus may be chromosomally normal). The decisions that may be faced by expectant parents inevitably engender a high level of anxiety at all stages of the screening process, and the outcomes of screening can be associated with considerable physical and psychological morbidity. No screening test can predict the severity of problems a person with Down's syndrome will have.

Objectives

To estimate and compare the accuracy of first trimester ultrasound markers alone, and in combination with first trimester serum tests for the detection of Down’s syndrome.

Search methods

We carried out extensive literature searches including MEDLINE (1980 to 25 August 2011), Embase (1980 to 25 August 2011), BIOSIS via EDINA (1985 to 25 August 2011), CINAHL via OVID (1982 to 25 August 2011), and The Database of Abstracts of Reviews of Effects (the Cochrane Library 2011, Issue 7). We checked reference lists and published review articles for additional potentially relevant studies.

Selection criteria

Studies evaluating tests of first trimester ultrasound screening, alone or in combination with first trimester serum tests (up to 14 weeks' gestation) for Down's syndrome, compared with a reference standard, either chromosomal verification or macroscopic postnatal inspection.

Data collection and analysis

Data were extracted as test positive/test negative results for Down's and non‐Down's pregnancies allowing estimation of detection rates (sensitivity) and false positive rates (1‐specificity). We performed quality assessment according to QUADAS criteria. We used hierarchical summary ROC meta‐analytical methods to analyse test performance and compare test accuracy. Analysis of studies allowing direct comparison between tests was undertaken. We investigated the impact of maternal age on test performance in subgroup analyses.

Main results

We included 126 studies (152 publications) involving 1,604,040 fetuses (including 8454 Down's syndrome cases). Studies were generally good quality, although differential verification was common with invasive testing of only high‐risk pregnancies. Sixty test combinations were evaluated formed from combinations of 11 different ultrasound markers (nuchal translucency (NT), nasal bone, ductus venosus Doppler, maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees); 12 serum tests (inhibin A, alpha‐fetoprotein (AFP), free beta human chorionic gonadotrophin (ßhCG), total hCG, pregnancy‐associated plasma protein A (PAPP‐A), unconjugated oestriol (uE3), disintegrin and metalloprotease 12 (ADAM 12), placental growth factor (PlGF), placental growth hormone (PGH), invasive trophoblast antigen (ITA) (synonymous with hyperglycosylated hCG), growth hormone binding protein (GHBP) and placental protein 13 (PP13)); and maternal age. The most frequently evaluated serum markers in combination with ultrasound markers were PAPP‐A and free ßhCG.

Comparisons of the 10 most frequently evaluated test strategies showed that a combined NT, PAPP‐A, free ßhCG and maternal age test strategy significantly outperformed ultrasound markers alone (with or without maternal age) except nasal bone, detecting about nine out of every 10 Down's syndrome pregnancies at a 5% false positive rate (FPR). In both direct and indirect comparisons, the combined NT, PAPP‐A, free ßhCG and maternal age test strategy showed superior diagnostic accuracy to an NT and maternal age test strategy (P < 0.0001). Based on the indirect comparison of all available studies for the two tests, the sensitivity (95% confidence interval) estimated at a 5% FPR for the combined NT, PAPP‐A, free ßhCG and maternal age test strategy (69 studies; 1,173,853 fetuses including 6010 with Down's syndrome) was 87% (86 to 89) and for the NT and maternal age test strategy (50 studies; 530,874 fetuses including 2701 Down's syndrome pregnancies) was 71% (66 to 75). Combinations of NT with other ultrasound markers, PAPP‐A and free ßhCG were evaluated in one or two studies and showed sensitivities of more than 90% and specificities of more than 95%.

High‐risk populations (defined before screening was done, mainly due to advanced maternal age of 35 years or more, or previous pregnancies affected with Down's syndrome) showed lower detection rates compared to routine screening populations at a 5% FPR. Women who miscarried in the over 35 group were more likely to have been offered an invasive test to verify a negative screening results, whereas those under 35 were usually not offered invasive testing for a negative screening result. Pregnancy loss in women under 35 therefore leads to under‐ascertainment of screening results, potentially missing a proportion of affected pregnancies and affecting test sensitivity. Conversely, for the NT, PAPP‐A, free ßhCG and maternal age test strategy, detection rates and false positive rates increased with maternal age in the five studies that provided data separately for the subset of women aged 35 years or more.

Authors' conclusions

Test strategies that combine ultrasound markers with serum markers, especially PAPP‐A and free ßhCG, and maternal age were significantly better than those involving only ultrasound markers (with or without maternal age) except nasal bone. They detect about nine out of 10 Down’s affected pregnancies for a fixed 5% FPR. Although the absence of nasal bone appeared to have a high diagnostic accuracy, only five out of 10 affected Down's pregnancies were detected at a 1% FPR.

Plain language summary

Screening tests for Down’s syndrome in first 24 weeks of pregnancy

Background   Down's syndrome (also known as Down's or Trisomy 21) is an incurable genetic disorder that causes significant physical and mental health problems, and disabilities. However, there is wide variation in how Down's affects people. Some individuals are severely affected whilst others have mild problems and are able to lead relatively normal lives. There is no way of predicting how badly a baby might be affected.

Expectant parents are given the choice to be tested for Down’s during pregnancy to assist them in making decisions. If a mother is carrying a baby with Down’s, then there is the decision about whether to terminate or continue with the pregnancy. The information offers parents the opportunity to plan for life with a Down’s child.

The most accurate tests for Down’s involve testing fluid from around the baby (amniocentesis) or tissue from the placenta (chorionic villus sampling (CVS)) for the abnormal chromosomes associated with Down’s. Both these tests involve inserting needles through the mother's abdomen and are known to increase the risk of miscarriage. Thus the tests are not suitable for offering to all pregnant women. Rather, tests that measure markers in the mother’s blood, urine or on ultrasound scans of the baby are used for screening. These screening tests are not perfect, they can miss cases of Down’s and also give a ‘high risk’ test results to a number of women whose babies are not affected by Down’s. Thus pregnancies identified as ‘high risk’ using these screening tests require further testing using amniocentesis (from 15 weeks' gestation) or CVS (from 10 + 0 to 13 + 6 weeks' gestation) to confirm a diagnosis of Down’s.

What we did   The aim of this review was to find out which of the first trimester ultrasound screening tests, with or without first trimester serum tests done during the first 14 weeks of pregnancy are the most accurate at predicting the risk of a pregnancy being affected by Down's. We looked at 11 different ultrasound markers and 12 different serum markers that can be used alone, in ratios or in combination, taken before 14 weeks' gestation, thus creating 60 screening tests for Down’s. We found 126 studies, involving 1,604,040 fetuses (including 8454 fetuses affected by Down's syndrome).

What we found   For the first 14 weeks of pregnancy, the evidence supports the use of first trimester ultrasound tests in combination with two serum (blood) markers ‐ especially pregnancy‐associated plasma protein A (PAPP‐A) and free beta human chorionic gonadotrophin (ßhCG) ‐ and maternal age, for Down's syndrome screening. In general, these tests are better than ultrasound markers on their own. They detect nine out of 10 pregnancies affected by Down's syndrome. Five per cent of women undertaking the test will have a high risk test result, however the majority of these pregnancies will not be affected by Down's syndrome.     Other important information to consider   The ultrasound tests themselves have no adverse effects for the woman, blood tests can cause discomfort, bruising and rarely infection. However some women who have a ‘high risk’ screening test result, and are given amniocentesis or CVS have a risk of miscarrying a baby unaffected by Down’s. Parents will need to weigh up this risk when deciding whether or not to have an amniocentesis or CVS following a ‘high risk’ screening test result.

Summary of findings

Background

This is one of a series of reviews on antenatal screening for Down's syndrome following a generic protocol (Alldred 2010) ‐ see Published notes for more details.

Target condition being diagnosed

Down’s syndrome

Down’s syndrome affects approximately one in 800 live‐born babies (Cuckle 1987). It results from a person having three, rather than two, copies of chromosome 21 — or the specific area of chromosome 21 implicated in causing Down's syndrome — as a result of trisomy or translocation. If not all cells are affected, the pattern is described as 'mosaic'. Down’s syndrome can cause a wide range of physical and mental problems. It is the commonest cause of mental disability, and is also associated with a number of congenital malformations, notably affecting the heart. There is also an increased risk of cancers such as leukaemia, and numerous metabolic problems including diabetes and thyroid disease. Some of these problems may be life‐threatening, or lead to considerable ill health, while some individuals with Down’s syndrome have only mild problems and can lead a relatively normal life.

There is no cure for Down’s syndrome, and antenatal diagnosis allows for preparation for the birth and subsequent care of a baby with Down’s syndrome, or for the offer of a termination of pregnancy. Having a baby with Down’s syndrome is likely to have a significant impact on family and social life, relationships and parents’ work. Special provisions may need to be made for education and care of the child, as well as accommodating the possibility of periods of hospitalisation.

Definitive invasive tests (amniocentesis and chorionic villus sampling (CVS)) exist that allow the diagnosis of Down's syndrome before birth but carry a risk of miscarriage. No test can predict the severity of problems a person with Down’s syndrome will have. Non‐invasive screening tests based on biochemical analysis of maternal serum or urine, or fetal ultrasound measurements, allow an estimate of the risk of a pregnancy being affected and provide parents with information to enable them to make choices about definitive testing. Such screening tests are used during the first and second trimester of pregnancy.

Screening tests for Down's syndrome

Initially, screening was determined solely by using maternal age to classify a pregnancy as high or low risk for trisomy 21, as it was known that older women had a higher chance of carrying a baby with Down’s syndrome (Penrose 1933).

Further advances in screening were made in the early 1980s, when Merkatz and colleagues investigated the possibility that low maternal serum alpha‐fetoprotein (AFP), obtained from maternal blood in the second trimester of pregnancy could be associated with chromosomal abnormalities in the fetus. Their retrospective case‐control study showed a statistically significant relationship between fetal trisomy, such as Down’s syndrome, and lowered maternal serum AFP (Merkatz 1984). This was further explored by Cuckle and colleagues in a larger retrospective trial using data collected as part of a neural tube defect (NTD) screening project (Cuckle 1984). This work was followed by calculation of risk estimates using maternal serum AFP values and maternal age, which ultimately led to the introduction of the two screening parameters in combination (Alfirevic 2004).

In 1987, in a small case‐control study of women carrying fetuses with known chromosomal abnormalities, Bogart and colleagues investigated maternal serum levels of human chorionic gonadotrophin (hCG) as a possible screening tool for chromosomal abnormalities in the second trimester (Bogart 1987). This followed the observations that low hCG levels were associated with miscarriages, which are commonly associated with fetal chromosomal abnormalities. They concluded that high hCG levels were associated with Down’s syndrome and because hCG levels plateau at 18 to 24 weeks, that this would be the most appropriate time for screening. Later work suggested that the ß subunit of hCG was a more effective marker than total hCG (Macri 1990; Macri 1993).

Second trimester unconjugated oestriol (uE3), produced by the fetal adrenals and the placenta, was also evaluated as a potential screening marker. In another retrospective case‐control study, uE3 was shown to be lower in Down’s syndrome pregnancies compared with unaffected pregnancies. When used in combination with AFP and maternal age, it appeared to identify more pregnancies affected by Down’s syndrome than AFP and age alone (Canick 1988). Further work suggested that all three serum markers (AFP, hCG and uE3) showed even higher detection rates when combined with maternal age (Wald 1988a; Wald 1988b) and appeared to be a cost‐effective screening strategy (Wald 1992a).

Two other serum markers, produced by the placenta, have been linked with Down’s syndrome, namely pregnancy‐associated plasma protein A or PAPP‐A, and Inhibin A. PAPP‐A has been shown to be reduced in the first trimester of Down’s syndrome pregnancies, with its most marked reduction in the early first trimester (Bersinger 1995). Inhibin A is high in the second trimester in pregnancies affected by Down’s syndrome (Cuckle 1995; Wallace 1995). There are some issues concerning the biological stability and hence reliability of this marker, and the effect this will have on individual risk.

In addition to serum and ultrasound markers for Down’s syndrome, work has been carried out looking at urinary markers. These markers include invasive trophoblast antigen, ß‐core fragment, free ßhCG and total hCG (Cole 1999). There is controversy about their value (Wald 2003a.

Screening and parental choice

Antenatal screening is used for several reasons (Alfirevic 2004), but the most important is to enable parental choice regarding pregnancy management and outcome. Before a woman and her partner opt to have a screening test, they need to be fully informed about the risks, benefits and possible consequences of such a test. This includes the choices they may have to face should the result show that the woman has a high risk of carrying a baby with Down’s syndrome and implications of both false positive and false negative screening tests. They need to be informed of the risk of a miscarriage due to invasive diagnostic testing, and the possibility that a miscarried fetus may be chromosomally normal. If, following invasive diagnostic testing, the fetus is shown to have Down’s syndrome, further decisions need to be made about continuation or termination of the pregnancy, the possibility of adoption and finally, preparation for parenthood. Equally, if a woman has a test that shows she is at a low risk of carrying a fetus with Down’s syndrome, it does not necessarily mean that the baby will be born with a normal chromosomal make up. This possibility can only be excluded by an invasive diagnostic test (Alfirevic 2003). The decisions that may be faced by expectant parents inevitably engender a high level of anxiety at all stages of the screening process, and the outcomes of screening can be associated with considerable physical and psychological morbidity. No screening test can predict the severity of problems a person with Down's syndrome will have.

Index test(s)

This review examined ultrasound and serum screening tests used in the first trimester of pregnancy (up to 14 weeks' gestation). The tests included the following individual ultrasound markers: nuchal translucency (NT), nasal bone, ductus venosus Doppler, maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees; and the following individual serum markers: inhibin A, AFP, free ßhCG, total hCG, pregnancy‐associated plasma protein A (PAPP‐A), uE3, a disintegrin and metalloprotease 12 (ADAM 12), placental growth factor (PlGF), placental growth hormone (PGH) invasive trophoblast antigen (ITA) (synonymous with hyperglycosylated hCG), growth hormone binding protein (GHBP) and placental protein 13 (PP13).

These markers can be used individually, in combination with age, and can also be used in combination with each other. The risks are calculated by comparing a woman's test result for each marker with values for an unaffected population, and multiplying this with her age‐related risk. Where several markers are combined, risks are computed using risk equations (often implemented in commercial software) that take into account the correlational relationships between the different markers and marker distributions in affected and unaffected populations.

Alternative test(s)

Down’s syndrome can be detected during pregnancy with invasive diagnostic tests such as amniocentesis or CVS, with or without prior screening. These tests are considered to be reference tests rather than index or screening tests. The ability to determine fetal chromosomal make up (also known as a karyotype) from amniotic fluid samples was demonstrated in 1966 by Steele and Breg (Steele 1966), and the first antenatal diagnosis of Down’s syndrome was made in 1968 (Valenti 1968). Amniocentesis is an invasive procedure which involves taking a small sample of the amniotic fluid (liquor) surrounding the baby, using a needle which goes through the abdominal wall into the uterus, and is usually performed after 15 weeks' gestation. Chorionic villus sampling involves taking a sample of the placental tissue using a needle which goes through the abdominal wall and uterus or a cannula through the cervix. It is usually performed between 10 and 13 weeks' gestation. Amniocentesis and CVS are both methods of obtaining fetal chromosome material, which are then used to diagnose Down’s syndrome. Both tests use ultrasound scans to guide placement of the needle. Amniocentesis carries a risk of miscarriage in the order of 1%; transabdominal CVS may carry a similar risk (Alfirevic 2003). A more recent systematic review suggests that the procedure‐related risk of pregnancy loss is lower than this (Akolekar 2015).

Recent developments in the use of cell‐free fetal DNA detection in maternal serum are paving the way for non‐invasive diagnosis of Down's syndrome and other trisomies, however these tests were not used as reference standards in any of the studies examined for this review, and were not included in the search strategy, which preceded their widespread introduction. A systematic review conducted by another group is currently in preparation, examining this newer screening technology ( Badeau 2015).

There are many different screening tests which are available and offered which are the subject of additional Cochrane reviews and there are other reviews looking at this area. Tests being assessed in the other Cochrane reviews include first trimester serum tests (Alldred 2015); urine tests (Alldred 2015a); second trimester serum markers (Alldred 2012); and tests that combine markers from the first trimester with markers from the second trimester (in press). Second trimester ultrasound markers have been assessed in a previous systematic review (Smith‐Bindman 2001).

Rationale

This is one of a suite of Cochrane reviews, the aim of which is to identify all screening tests for Down's syndrome used in clinical practice, or evaluated in the research setting, in order to try to identify the most accurate test(s) available, and to provide clinicians, policy‐makers and women with robust and balanced evidence on which to base decisions about interpreting test results and implementing screening policies to triage the use of invasive diagnostic testing. The full set of reviews is described in the generic protocol (Alldred 2010).

The topic has been split into several different reviews to allow for greater ease of reading and greater accessibility of data, and also to allow the reader to focus on separate groups of tests, for example, first trimester serum tests alone, first trimester ultrasound alone, first trimester serum and ultrasound, second trimester serum alone, first and second trimester serum, combinations of serum and ultrasound markers and urine markers alone. An overview review will compare the best tests, focusing on commonly used strategies, from each of these groups to provide comparative results between the best tests in the different categories. This review is written with the global perspective in mind, rather than to conform with any specific local or national policy, as not all tests will be available in all areas where screening for Down's syndrome is carried out.

A systematic review of second trimester ultrasound markers in the detection of Down’s syndrome fetuses was published in 2001 which concluded that nuchal fold thickening may be useful in detecting Down’s syndrome, but that it was not sensitive enough to use as a screening test. The review concluded that the other second trimester ultrasound markers did not usefully distinguish between Down’s syndrome and pregnancies without Down’s syndrome (Smith‐Bindman 2001). There has yet to be a systematic review and meta‐analysis of the observed data on serum, urine and first trimester ultrasound markers, in order to draw rigorous and robust conclusions about the diagnostic accuracy of available Down’s syndrome screening tests.

Objectives

The aim of this review was to estimate and compare the accuracy of first trimester ultrasound with and without serum markers for the detection of Down’s syndrome in the antenatal period, both as individual markers and as combinations of markers. Accuracy is described by the proportion of fetuses with Down’s syndrome detected by screening before birth (sensitivity or detection rate) and the proportion with a low‐risk screening test result (negative) from amongst babies born without Down's syndrome. We grouped our analyses to focus on investigating the value of adding increasing numbers of markers (comparing single, dual, triple, quadruple, quintuple and sextuple tests).

Investigation of sources of heterogeneity

We had planned to investigate whether a uniform screening test is suitable for all women, or whether different screening methods are more applicable to different groups, defined by advanced maternal age, ethnic groups and aspects of the pregnancy and medical history such as multiple (multifetal) pregnancy, diabetes and family history of Down's syndrome. We also planned to examine whether there was evidence of overestimation of test accuracy in studies evaluating risk equations in the derivation sample rather than in a separate validation sample.

Methods

Criteria for considering studies for this review

Types of studies

We included studies in which all women from a given population had one or more index test(s) compared to a reference standard. Both consecutive series and diagnostic case‐control study designs were included. Randomised trials where individuals were randomised to different screening strategies and all verified using a reference standard were also eligible for inclusion. Studies in which test strategies were compared head‐to‐head either in the same women, or between randomised groups were identified for inclusion in separate comparisons of test strategies. Studies were excluded if they included less than five Down's syndrome cases, or more than 20% of participants were not followed up.

Participants

Pregnant women at less than 14 weeks' gestation confirmed by ultrasound, who had not undergone previous testing for Down’s syndrome in their pregnancy were eligible. Studies were included if the pregnant women were unselected, or if they represented groups with increased risk of Down’s syndrome, or difficulty with conventional screening tests including maternal age greater than 35 years old, multifetal pregnancy, diabetes mellitus and a family history of Down’s syndrome.

Index tests

Improved diagnostic performance can be obtained by using several tests in combination, such as maternal age and serum marker combinations, or combinations of maternal age, serum markers and sonographic measurements. We examined individual first trimester ultrasound markers or combinations of these markers with one or more first trimester serum tests, with and without adjustment for maternal age.

The following ultrasound markers were examined: NT, nasal bone, ductus venosus Doppler, maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees.

The serum markers examined in different combinations with ultrasound markers were inhibin A, AFP, free ßhCG, total hCG, PAPP‐A, uE3, ADAM 12, PlGF, PGH, ITA (h‐hCG), GHBP and PP13.

We examined comparisons of ultrasound markers in isolation and in various combinations with or without serum markers. The combinations included one or two ultrasound markers with single (one marker), double (two markers), triple (three markers), quadruple (four markers), quintuple and sextuple (six markers) serum markers, with or without adjustment for maternal age.

Where tests were used in combinations, we examined the performance of test combinations according to predicted probabilities computed using risk equations and dichotomised into high risk and low risk at some standard high‐risk value. Risk equations are often coded into software to produce 'risk score' computations, which provide an individual's predicted probability of Down’s syndrome.

Target conditions

Down's syndrome in the fetus due to trisomy, translocation or mosaicism.

Reference standards

We considered several reference standards, involving chromosomal verification and postnatal macroscopic inspection.

Amniocentesis and chorionic villus sampling (CVS) are invasive chromosomal verification tests undertaken during pregnancy. They are highly accurate, but the process carries a 1% miscarriage rate, and therefore they are only used in pregnancies considered to be at high risk of Down's syndrome, or on the mother's request. All other types of testing (postnatal examination, postnatal karyotyping, birth registers and Down’s syndrome registers) are based on information available at the end of pregnancy. The greatest concern is not their accuracy, but the loss of the pregnancy to miscarriage between the urine test and the reference standard. Miscarriage with cytogenetic testing of the fetus is included in the reference standard where available. We anticipated that older studies, and studies undertaken in older women are more likely to have used invasive chromosomal verification tests in all women.

Studies undertaken in younger women and more recent studies were likely to use differential verification as they often only used prenatal karyotypic testing on fetuses considered screen positive/high risk according to the screening test; the reference standard for most unaffected infants being observing a phenotypically normal baby. Although the accuracy of this combined reference standard is considered high, it is methodologically a weaker approach as pregnancies that miscarry between the index test and birth are likely to be lost from the analysis, and miscarriage is more likely to occur in Down's than normal pregnancies. We investigated the impact of the likely missing false negative results in sensitivity analyses.

Search methods for identification of studies

Electronic searches

We applied a sensitive search strategy to search the following databases using the search strategies listed in Appendix 1. We used one generic search to identify studies for all reviews in this series.

We searched the following databases

MEDLINE via OVID (1980 to 25 August 2011)
Embase via Dialog Datastar (1980 to 25 August 2011)
BIOSIS via EDINA (1985 to 25 August 2011)
CINAHL via OVID (1982 to 25 August 2011)
The Database of Abstracts of Reviews of Effects (the Cochrane Library 2011, Issue 7)
MEDION (25 August 2011)
The Database of Systematic Reviews and Meta‐Analyses in Laboratory Medicine (www.ifcc.org/) (25 August 2011)
The National Research Register (archived 2007)
Health Services Research Projects in Progress database (HSRPROJ) (25 August 2011)

The search strategy combined three sets of search terms (seeAppendix 1). The first set was made up of named tests, general terms used for screening/diagnostic tests and statistical terms. Note that the statistical terms were used to increase sensitivity and were not used as a methodological filter to increase specificity. The second set was made up of terms that encompass Down's syndrome, and the third set made up of terms to limit the testing to pregnant women. All terms within each set were combined with the Boolean operator OR and then the three sets were combined using AND. The terms used were a combination of subject headings and free‐text terms. The search strategy was adapted to suit each database searched.

We attempted to identify cumulative papers that reported data from the same data set, and contacted authors to obtain clarification of the overlap between data presented in these papers, in order to prevent data from the same women being analysed more than once.

Searching other resources

In addition, we examined references cited in studies identified as being potentially relevant, and those cited by previous reviews. We contacted authors of studies where further information was required. We did not apply a diagnostic test filter, and we did not apply language restrictions to the search.

We carried out forward citation searching of relevant items, using the search strategy in ISI citation indices, Google Scholar and Pubmed ‘related articles’.

Data collection and analysis

Selection of studies

Two review authors screened the titles and abstracts (where available) of all studies identified by the search strategy. Full‐text versions of studies identified as being potentially relevant were obtained and independently assessed by two review authors for inclusion, using a study eligibility screening pro forma according to the pre‐specified inclusion criteria. Any disagreement between the two review authors was settled by consensus, or where necessary, by a third party.

Data extraction and management

A data extraction form was developed and piloted using a subset of 20 identified studies (from all identified studies in this suite of reviews). Two review authors independently extracted data, and where disagreement or uncertainty existed, a third review author validated the information extracted.

Data on each marker were extracted as binary test positive/test negative results for Down's and non‐Down's pregnancies, with a high‐risk result ‐ as defined by each individual study ‐ being regarded as test positive (suggestive or diagnostic of Down's syndrome), and a low‐risk result being regarded as test negative (suggestive of absence of Down's Syndrome). Where results were reported at several thresholds, we extracted data at each threshold.

We noted those in special groups that posed either increased risk of Down’s syndrome or difficulty with conventional screening tests including maternal age greater than 35 years old, multifetal pregnancy, diabetes mellitus and family history of Down’s syndrome.

Assessment of methodological quality

We used a modified version of the QUADAS tool (Whiting 2003), a quality assessment tool for use in systematic reviews of diagnostic accuracy studies, to assess the methodological quality of included studies. We anticipated that a key methodological issue would be the potential for bias arising from the differential use of invasive testing and follow‐up for the reference standard according to index test results, bias arising due to higher loss to miscarriage in false negatives than true negatives. We chose to code this issue as originating from differential verification in the QUADAS tool: we are aware that it could also be coded under delay in obtaining the reference standard, and reporting of withdrawals. We omitted the QUADAS item assessing quality according to length of time between index and reference tests, as Down's syndrome is either present or absent rather than a condition that evolves and resolves, and disregarding the differential reference standard issue, thus any length of delay is acceptable. Two review authors assessed each included study separately. Any disagreement between the two review authors was settled by consensus, or where necessary, by a third party. Each item in the QUADAS tool was marked as ‘yes’, ‘no’ or ‘unclear’, and scores were summarised graphically. We did not use a summary quality score.

QUADAS criteria included the following 10 questions.

Was the spectrum of women representative of the women who will receive the test in practice? (Criteria met if the sample was selected from a wide range of childbearing ages, or selected from a specified ‘high‐risk’ group such as over 35s, family history of Down’s syndrome, multifetal pregnancy or diabetes mellitus, provided all affected and unaffected fetuses included that could be tested at the time point when the screening test would be applied; criteria not met if the sample taken from a select or unrepresentative group of women (i.e. private practice), was an atypical screening population or recruited at a later time point when selection could be affected by selective fetal loss.)
Is the reference standard likely to correctly classify the target condition? (Amniocentesis, chorionic villus sampling, postnatal karyotyping, miscarriage with cytogenetic testing of the fetus, a phenotypically normal baby or birth registers are all regarded as meeting this criteria.)
Did the whole sample or a random selection of the sample receive verification using a reference standard of diagnosis?
Did women receive the same reference standard regardless of the index test result?
Was the reference standard independent of the index test result (i.e. the index test did not form part of the reference standard)?
Were the index test results interpreted without knowledge of the results of the reference standard?
Were the reference standard results interpreted without knowledge of the results of the index test?
Were the same clinical data (i.e. maternal age and weight, ethnic origin, gestational age) available when test results were interpreted as would be available when the test is used in practice?
Were uninterpretable/intermediate test results reported?
Were withdrawals from the study explained?

Statistical analysis and data synthesis

We initially examined each test or test strategy at each of the common risk thresholds used to define test positivity by plotting estimates of sensitivity and specificity from each study on forest plots and in receiver operating characteristic (ROC) space. Test strategies were selected for further investigation if they were evaluated in four or more studies or, if there were three or fewer studies, but the individual study results indicated performance likely to be superior to a sensitivity of 70% and specificity of 90%.

Estimation of average sensitivity and specificity

The analysis for each test strategy was undertaken first restricting to studies which reported a common threshold to estimate average sensitivity and specificity for each test at each threshold. Although data on all thresholds were extracted, we present only key common thresholds (historically reported in literature based on age‐related risk) close to risks of 1:384, 1:250 and the 5% false positive rate (FPR), unless other thresholds were more commonly reported. Where combinations of tests were used in a risk score, we extracted the result for the test combination using the risk score and not the individual components that made up the test.

Meta‐analyses were undertaken using hierarchical summary ROC (HSROC) models, which included estimation of random‐effects in accuracy and threshold parameters when there were four or more studies. When there was an insufficient number of studies to reliably estimate all the parameters in the HSROC model, univariate random‐effects logistic regression models were used to obtain pooled estimates of sensitivity and specificity. It is common in this field for studies to report sensitivity for a fixed specificity (usually a 5% FPR). This removes the requirement to account for the correlation between sensitivity and specificity across studies by using a bivariate model since all specificities are the same value. Thus, at a fixed specificity value, the summary estimate of sensitivity was obtained using a univariate random‐effects logistic regression model. This model was further simplified to a fixed‐effect model when there were only two or three studies and heterogeneity was not observed on the SROC plot. All analyses were undertaken using the NLMIXED procedure in SAS (version 9.2; SAS Institute, Cary, NC) and the xtmelogit command in Stata version 11.2 (Stata‐Corp, College Station, TX, USA).

Comparisons between tests

Comparisons between tests were first made utilising all available studies, selecting one threshold for each test from each study to estimate a SROC curve without restricting to a common threshold. The threshold for each test was chosen from each study according to the following order of preference: a) the risk threshold closest to one in 250; b) a multiples of the median (MoM) or presence/absence threshold; c) the performance closest to a 5% FPR or 95th percentile. The 5% FPR was chosen as a cut‐off point as this is the cut‐off most commonly reported in the literature. The analysis that used all available studies was performed by including the most evaluated or best performing test strategies in a single HSROC model. The model included two indicator terms for each test to allow for differences in accuracy and threshold. As there were very few studies for each test, a symmetric summary ROC curve was assumed. In addition, because the analysis failed to converge, we assumed fixed‐effect for the threshold and accuracy parameters. An estimate of the sensitivity of each test for a 5% FPR was derived from the SROC curve, and associated confidence intervals were obtained using the delta method.

Direct comparisons between tests were based on results of very few studies, and were analysed using a simplified HSROC model with fixed‐effect and symmetrical underlying SROC curves because the number of studies was insufficient to estimate between study heterogeneity in accuracy and threshold or asymmetry in the shape of the SROC curves. A separate model was used to make each pair‐wise comparison. Comparisons between tests were assessed by using likelihood ratio tests to test if the differences in accuracy were statistically significant or not. The differences were expressed as ratios of diagnostic odds ratios and were reported with 95% confidence intervals. As studies rarely report data cross‐classified by both tests for Down's and normal pregnancies, the analytical method did not take full account of the pairing of test results, but the restriction to direct head‐to‐head comparisons should have removed the potential confounding of test comparisons with other features of the studies. The strength of evidence for differences in performance of test strategies relied on evidence from both the direct and indirect comparisons.

Investigations of heterogeneity

If there were 10 or more studies available for a test, we had planned to investigate heterogeneity by adding covariate terms to the HSROC model (meta‐regression) to assess the effect of each factor stated in the Investigation of sources of heterogeneity section on accuracy and threshold.

Sensitivity analyses

Mothers with pregnancies identified as high risk for Down's syndrome by ultrasound and serum testing were often offered immediate definitive testing by amniocentesis, whereas those considered low risk were assessed for Down's syndrome by inspection at birth. Such delayed and differential verification will introduce bias most likely through there being greater loss to miscarriage in the Down's syndrome pregnancies that were not detected by the ultrasound and serum testing (the false negative diagnoses). Testing and detection of miscarriages is impractical in many situations, and no clear data are available on the magnitude of these miscarriage rates.

To account for potential bias introduced by such a mechanism, where possible, we performed sensitivity analyses by increasing the number of false negatives in studies where delayed verification in test negatives occurred (Mol 1999). We increased the number of false negatives in such studies by a multiplicative factor that we applied incrementally from 10% to 50%. The final value of 50% assumes the true number of false negatives is 1.5 times the observed number of false negatives, implying the observed number of false negatives.is 67% (i.e. 1/1.5) of the true number and the fetal loss rate is 33%. Since no increments were added to the number of true negatives, this represents a scenario where a third more pregnancies affected by Down’s syndrome is likely to miscarry compared to those unaffected by Down's syndrome. This is thought to be higher than the likely value.

We intended to conduct these sensitivity analyses on analyses investigating the effect of maternal age on test sensitivity. However, due to limited data, we performed the sensitivity analyses when comparing high‐risk populations with routine screening populations. This comparison was considered a proxy for the effect of maternal age because the main indication for referral for invasive testing was often increased risk due to advanced maternal age.

Results

Results of the search

After the results from each bibliographic database were combined and duplicates were removed, the search for the whole suite of reviews identified a total of 15,394 papers. After screening out obviously inappropriate papers based on their title and abstract, 1145 papers remained and we obtained full‐text copies for formal assessment of eligibility. From these, a total of 269 papers were deemed eligible and were included in the suite of reviews. A total of 126 studies (reported in 152 publications) were included in this review of first trimester ultrasound alone or in combination with first trimester serum screening. Since women with multifetal pregnancies were included in six of the 126 studies, where a study included multifetal pregnancies, we report fetuses rather than women or pregnancies. The review involved 1,604,040 fetuses including 8454 Down's syndrome cases.

A total of 60 different test strategies were evaluated in the 126 studies. These tests were formed from combinations of different ultrasound markers, serum tests and maternal age. The 11 individual ultrasound markers were nuchal translucency (NT), nasal bone, ductus venosus Doppler (ductus venosus a‐wave reversed, ductus venosus pulsivity index), maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees. The 12 individual serum markers were inhibin A, alpha‐fetoprotein (AFP), free beta human chorionic gonadotrophin (ßhCG), total hCG, pregnancy‐associated plasma protein A (PAPP‐A), unconjugated oestriol (uE3), disintegrin and metalloprotease 12 (ADAM 12), placental growth factor (PlGF), placental growth hormone (PGH), invasive trophoblast antigen (ITA) (h‐hCG), growth hormone binding protein (GHBP), and placental protein 13 (PP13). The strategies evaluated, with or without maternal age, included 13 single ultrasound markers; five combinations of two or more ultrasound markers; six ultrasound and single serum marker combinations; 22 ultrasound and double serum marker combinations; nine ultrasound and triple serum marker combinations; one ultrasound and quadruple serum marker combination; three ultrasound and quintuple serum marker combinations; and one ultrasound and sextuple serum marker combination. Seventy‐eight of the 126 studies only evaluated the performance of a single first trimester ultrasound or ultrasound and serum test or test strategy; 27 studies evaluated two tests, 10 evaluated three tests, four evaluated four tests, four evaluated five tests, one evaluated eight tests (Koster 2011), one evaluated 11 tests (Kagan 2010), and one evaluated 19 tests (Wald 2003).

The following test combinations were evaluated by four or more studies.

Ultrasound and triple serum markers

NT, PAPP‐A, free ßhCG, ADAM 12 and maternal age (four studies; 2571 women, including 256 Down's syndrome pregnancies)

Ultrasound and double serum markers

NT, PAPP‐A, free ßhCG and maternal age (69 studies; 1,173,853 fetuses, including 6010 Down's syndrome cases)

Ultrasound and single serum markers

NT, free ßhCG and maternal age (five studies; 10,795 women, including 421 Down's syndrome pregnancies)
NT, PAPP‐A and maternal age (five studies; 9,814 women including 372 Down's syndrome pregnancies)

Ultrasound markers alone

NT, nasal bone and maternal age (five studies, 29,699 women, including 221 Down's syndrome pregnancies)
NT and maternal age (50 studies; 530,874 fetuses including 2701 Down's syndrome cases)
Nasal bone and maternal age (four studies; 25,303 women, including 165 Down's syndrome pregnancies)
Ductus and maternal age (five studies; 5,331 women including 165 Down's syndrome pregnancies)
Nasal bone (11 studies; 48,279 fetuses including 290 Down's syndrome cases)
NT (13 studies; 90,978 fetuses, including 593 Down's syndrome cases)

Of the remaining test combinations, four were evaluated in three studies, six were evaluated in two studies and the remaining 40 in single studies only.

Methodological quality of included studies

The studies were judged to be of high methodological quality in most categories (Figure 1) and details are provided in the Characteristics of included studies. The spectrum of participants was judged to be representative in all study cohorts. The reference standard used was judged unclear in three studies (Hafner 1998; Krantz 2000; Orlandi 1997) and unacceptable in one study (Noble 1995). Due to the nature of testing for Down's syndrome screening and the potential side effects of invasive testing, differential verification is almost universal in the general screening population, as most women whose screening test result is defined as low risk (negative) will have their screening test verified at birth, rather than by invasive diagnosis in the antenatal period. Partial verification was avoided in 81 study cohorts (64%) and differential verification was avoided in 15 study cohorts (12%). Both differential and partial verification was avoided in 14 study cohorts (Biagiotti 1998; Borenstein 2008; Christiansen 2005; Cicero 2004a; De Graaf 1999; Hewitt 1996; Maiz 2007; Matias 1998; Matias 2001; Mavrides 2002; Molina 2010 high risk; Otaño 2002; Pajkrt 1998a; Prefumo 2005 ). Of the 14 study cohorts, the populations in 13 were high‐risk referral for invasive testing (prior to screening being undertaken), while one (Christiansen 2005) obtained maternal serum samples through screening programmes for syphilis and Down's syndrome. Reference standard results were unblinded in 124 study cohorts and unclear in three study cohorts. In contrast, index test results were blinded in 113 study cohorts and unclear in 14. It would be difficult to blind clinicians performing invasive diagnostic tests (reference standards) to the index test result, unless all women received the same reference standard, which would not be appropriate in most scenarios. Any biases secondary to a lack of clinician blinding are likely to be minimal.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

Most studies seemed to indicate 100% follow‐up, however there will inevitably be losses to follow‐up due to women moving out of area, for example. Studies sometimes accounted for these and it is unlikely that there were enough losses to follow‐up to have introduced significant bias. There was likely under‐ascertainment of miscarriage, and very few papers accounted for miscarriage or performed tissue karyotyping in pregnancies resulting in miscarriage. Some studies attempted to adjust for predicted miscarriage rate and the incidence of Down's syndrome in this specific population, but most did not. We have not attempted to adjust for expected miscarriage rate in this review. There is a higher natural miscarriage rate in the first trimester, however this will be uniform across studies and therefore unlikely to introduce significant bias.

Some studies which provided estimates of risk using multivariable equations used the same data set to evaluate performance of the risk equation as was used to derive the equation. This is often thought to lead to over‐estimation of test performance.

Findings

The results for the 10 most evaluated test strategies are presented in Table 1. Additional information and results at specific thresholds are provided below.

Summary of findings 1. Performance of the 10 most evaluated first trimester ultrasound markers alone or in combination with first trimester serum tests.

Review question	What is the accuracy of ultrasound based markers alone and in combination with maternal age and/or first trimester serum markers for screening for Down's syndrome?
Population	Pregnant women at less than 14 weeks' gestation confirmed by ultrasound, who had not undergone previous testing for Down’s syndrome. Some studies were undertaken in women identified to be at high risk based on maternal age.
Settings	All settings.
Numbers of studies, pregnancies and Down's syndrome cases	126 studies (reported in 152 publications) involving 1,604,040 fetuses of which 8454 were Down's syndrome cases
Index tests	Risk scores computed using maternal age and first trimester ultrasound and serum markers for ultrasound markers ‐ NT, nasal bone, ductus venosus Doppler, maxillary bone length, fetal heart rate, aberrant right subclavian artery, frontomaxillary facial angle, presence of mitral gap, tricuspid regurgitation, tricuspid blood flow and iliac angle 90 degrees ‐ and serum markers ‐ inhibin A, AFP, free ßhCG, total hCG, PAPP‐A, uE3, ADAM 12, PlGF, PGH, ITA (h‐hCG), GHBP and PP13.
Reference standards	Chromosomal verification (amniocentesis and CVS undertaken during pregnancy, and postnatal karyotyping) and postnatal macroscopic inspection.
Study limitations	116 studies only used selective chromosomal verification during pregnancy, and were at risk of under‐ascertainment of Down's syndrome cases due to pregnancy loss between administering the serum test and the reference standard.
Test strategy	Studies	Women (Down's cases)	Sensitivity (95% CI)	Specificity (95% CI)*	Consequences in a hypothetical cohort of 10,000 pregnant women assuming Down’s syndrome affects approximately one in 800 live‐born babies
Test strategy	Studies	Women (Down's cases)	Sensitivity (95% CI)	Specificity (95% CI)*	Missed cases	False positives
Nasal bone	11	48,279 (290)	49 (34, 64)	99 (99, 100)	7	100
NT	13	90,978 (593)	70 (61, 78)	95	4	500
NT and maternal age	50	530,874 (2701)	71 (66, 75)	95	4	500
Nasal bone and maternal age	4	25,303 (165)	68 (28, 92)	95	4	500
Ductus and maternal age	5	5331 (165)	68 (49, 83)	95	4	500
NT, nasal bone and maternal age	5	29,699 (221)	78 (55, 91)	95	3	500
NT, free ßhCG and maternal age	5	10,795 (421)	77 (72, 82)	95	3	500
NT, PAPP‐A and maternal age	5	9814 (372)	81 (75, 86)	95	3	500
NT, PAPP‐A, free ßhCG and maternal age	69	1,173,853 (6010)	87 (86, 89)	95	2	500
NT, PAPP‐A, free ßhCG, ADAM 12 and maternal age	4	2571 (256)	82 (75, 87)	95	3	500

Ratio of DORs (95% CI); P value (Studies)	Nasal bone	NT	Nasal bone and age	Ductus and age	NT and age	NT, nasal bone and age	NT, free ßhCG and age	NT, PAPP‐A and age	NT, PAPP‐A, free ßhCG and age
NT	–
Nasal bone and age	–	–
Ductus and age	1.19 (0.12, 11.4); P = 0.84 (K = 1)	–	0.85 (0.21, 3.41); P = 0.76 (K = 1)
NT and age	0.62 (0.13, 2.93); P = 0.50 (K = 2)	1.25 (0.90, 1.74); P = 0.17 (K = 3)	0.84 (0.48, 1.49); P = 0.52 (K = 3)	1.07 (0.51, 2.23); P = 0.85 (K = 3)
NT, nasal bone and age	0.61 (0.12, 3.10); P = 0.50 (K = 2)	–	4.01 (1.51, 10.6); P = 0.01 (K = 2)	0.95 (0.23, 3.97); P = 0.93 (K = 1)	1.05 (0.70, 1.56); P = 0.82 (K = 5)
NT, free ßhCG and age	–	2.15 (1.33, 3.50); P = 0.007 (K = 2)	–	–	1.47 (1.00, 2.15); P = 0.05 (K = 4)	–
NT, PAPP‐A and age	–	2.86 (1.73, 4.73); P = 0.001 (K = 2)	–	–	1.88 (1.27, 2.78); P = 0.004 (K = 4)	–	1.28 (0.84, 1.93); P = 0.23 (K = 4)
NT, PAPP‐A, free ßhCG and age	3.83 (0.89, 16.4); P = 0.07 (K = 2)	4.35 (2.00, 9.46); P = 0.015 (K = 4)	–	3.00 (0.42, 21.2); P = 0.19 (K = 1)	3.19 (2.19, 4.66); P < 0.0001 (K = 25)	1.23 (0.63, 2.40); P = 0.50 (K = 2)	2.06 (1.31, 3.22); P = 0.004 (K = 4)	1.61 (1.02, 2.55); P = 0.043 (K = 4)
NT, PAPP‐A, free ßhCG, ADAM 12 and age	–	–	–	–	–	–	–	–	0.87 (0.49, 1.52); P = 0.60 (K = 4)

Ratio of DORs (95% CI); P value		Nasal bone	NT	Nasal bone and age	Ductus and age	NT and age	NT, nasal bone and age	NT, free ßhCG and age	NT, PAPP‐A and age	NT, PAPP‐A, free ßhCG and age
	DOR (95% CI) Studies	132 (71, 245) K = 11	45 (31, 67) K = 13	40 (7, 224) K = 4	41 (18, 92) K = 5	46 (37, 57) K = 50	66 (24, 180) K = 5	65 (51, 84) K = 5	80 (59, 109) K = 5	133 (114, 155) K = 69
NT	45 (31, 67) K = 13	0.34 (0.16, 0.71); P = 0.006
Nasal bone and age	40 (7, 224) K = 4	0.31 (0.05, 1.90); P = 0.18	0.90 (0.16, 5.05); P = 0.89
Ductus and age	41 (18, 92) K = 5	0.31 (0.11, 0.87); P = 0.03	0.90 (0.37, 2.20); P = 0.80	1.00 (0.11, 9.34); P = 1.00
NT and age	46 (37, 57) K = 50	0.35 (0.19, 0.66); P = 0.002	1.02 (0.66, 1.58); P = 0.92	1.14 (0.23, 5.61); P = 0.87	1.14 (0.52, 2.49); P = 0.74
NT, nasal bone and age	66 (24, 180) K = 5	0.50 (0.14, 1.81); P = 0.26	1.47 (0.47, 4.58); P = 0.48	1.64 (0.12, 21.5); P = 0.62	1.64 (0.33, 8.08); P = 0.46	1.43 (0.52, 3.98); P = 0.48
NT, free ßhCG and age	65 (51, 84) K = 5	0.49 (0.25, 0.98); P = 0.04	1.44 (0.89, 2.34); P = 0.12	1.61 (0.26, 10.1); P = 0.56	1.61 (0.65, 3.99); P = 0.26	1.41 (1.02, 1.96); P = 0.04	0.98 (0.30, 3.19); P = 0.98
NT, PAPP‐A and age	80 (59, 109) K = 5	0.61 (0.29, 1.25); P = 0.16	1.77 (1.05, 3.00); P = 0.04	1.98 (0.30, 13.1); P = 0.42	1.98 (0.76, 5.15); P = 0.14	1.73 (1.19, 2.53); P = 0.005	1.21 (0.35, 4.13); P = 0.73	1.23 (0.74, 2.05); P = 0.35
NT, PAPP‐A, free ßhCG and age	133 (114, 155) K = 69	1.00 (0.55, 1.84); P = 1.00	2.93 (1.96, 4.40); P < 0.0001	3.27 (0.68, 15.8); P = 0.14	3.27 (1.53, 7.00); P = 0.003	2.87 (2.21, 3.72); P < 0.0001	2.00 (0.73, 5.45); P = 0.17	2.03 (1.52, 2.72) P < 0.0001	1.65 (1.17, 2.34) P = 0.005
NT, PAPP‐A, free ßhCG, ADAM 12 and age	85 (58, 124) K = 4	0.64 (0.30, 1.37); P = 0.23	1.88 (1.07, 3.32); P = 0.03	2.10 (0.31, 14.1); P = 0.39	2.10 (0.78, 5.63); P = 0.12	1.84 (1.19, 2.84); P = 0.007	1.28 (0.37, 4.47); P = 0.65	1.30 (0.81, 2.09) P = 0.26	1.06 (0.61, 1.86) P = 0.81	0.64 (0.43, 0.96) P = 0.03

Study	NT, PAPP‐A, free ßhCG and age	Nasal bone	NT and age	NT	Maternal age (range) in years	Reference standard	Population	Study design	Study location
Acacio 2001				X	Mean 35.8 (21‐45)	CVS biopsy, amniocentesis or blood or placenta used for fetal karyotyping	High‐risk referral for invasive testing	Retrospective study of patient notes	South America
Audibert 2001			X		Mean 30.1, all < 38, 86% < 35, 14% ≥ 35	Prenatal karyotype conducted (in 7.6% of patients) depending on presence of risk > 125, high maternal age, parental anxiety, history of chromosomal defects or parental translocation or abnormal second trimester scan age	Routine screening	Prospective consecutive series	France
Babbur 2005				X	Median 37 (19‐46)	Invasive testing offered to women with NT > 3 mm or risk > 1:250 as defined by combined NT and serum results (CVS from 11 weeks, amniocentesis from 15 weeks). Rapid in situ hybridisation test in patients with risk > 1:30. No details given of any follow‐up to birth	Women requesting screening (self‐paying service) and women attending on account of previous pregnancy history of fetal abnormality	Prospective cohort	UK
Barrett 2008	X				Mean 34.9 for screen positives, 30.5 for screen negatives	Karyotyping or follow‐up to birth	Routine screening	Cohort	Australia
Belics 2011					Mean 36.4 (15‐46) for Down's cases, 29.8 (15‐49) for unaffected pregnancies	Amniocentesis or CVS (85% of women) or follow‐up to birth	High‐risk referral for invasive testing	Cohort	Budapest
Benattar 1999			X		Mean 32 (16‐46), 8.3% > 35	Amniocentesis due to maternal age > 38 years (6.1% or women). Karyotyping encouraged for women with positive result on one or more index test. No details of reference standard for index test negative women	Routine screening	Prospective cohort	France
Bestwick 2010	X		X	X	Median 39 for Down's cases, 34 for unaffected pregnancies	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	UK
Biagiotti 1998	X		X		Unclear (maybe all ≥ 38)	Amniocentesis or CVS	High‐risk referral for invasive testing	Case control	Italy
Borenstein 2008					Median 35 (17‐49)	CVS	High‐risk referral for invasive testing	Prospective cohort	UK
Borrell 2005	X		X		Not reported	CVS (high‐risk women) or follow‐up to birth	Routine screening	Retrospective cohort	Spain
Borrell 2009	X				Mean 32	Karyotyping or follow‐up to birth	Routine screening and high‐risk referral	Prospective cohort	Spain
Brameld 2008	X				Median 31 (14‐47), 20% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Australia
Brizot 2001			X		Median 28 (13‐46), 19.4% ≥ 35	Antenatal karyotyping (5.9% of pregnancies: 62% of high‐risk, 29% of medium‐risk and 3% of the low‐risk women). Follow‐up to birth (85.3% of women)	Routine screening	Prospective cohort	Brazil
Centini 2005	X				≥ 35 (35‐44)	Amniocentesis in women high risk on screening (16.2%). Follow‐up at birth in women who were low risk on screening	High‐risk patients undergoing routine screening	Retrospective cohort	Italy
Chasen 2003			X		Median 33 (IQR 31‐36), 36.2% ≥ 35	Karyotyping or follow‐up to birth in 96.1% of patients	Routine screening	Prospective consecutive cohort	USA
Chen 2009					Median 30 (20‐44) for Down's cases, 32 (19‐40) for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	China
Christiansen 2005	X				Not reported	Karyotyping	Screening programmes for syphilis and Down's syndrome	Case control	Denmark
Christiansen 2009	X				Median 37.5 for Down's cases, 36.4 for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	Denmark
Christiansen 2010	X				Median 36 (25‐44) for Down's cases, 29 (17‐45) for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	Denmark
Cicero 2004a					Median 37 (16‐48)	CVS	High‐risk referral for invasive testing	Prospective cohort	USA
Cicero 2006		X			Median 35 (18‐50)	CVS or amniocentesis (in high risk women) or follow‐up to birth	Routine screening	Prospective cohort	UK
Cocciolone 2008 (first trimester screening cohort)	X				Median 31.3	Karyotyping or follow‐up to birth	Routine screening	Cohort	Australia
Cowans 2009	X				Mean 38 (16‐49) for Down's cases, 29 (13‐56) for unaffected pregnancies	Karyotyping or follow‐up to birth	Routine screening	Cohort	UK
Cowans 2010	X				Mean 37.0 (IQR 32.9‐40.5) for Down's cases, 32.4 (IQR 29.0‐35.9) for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	UK
Crossley 2002	X		X		Median 29.9, 15.4% ≥ 35	CVS (offered where women had high NT measurements), amniocentesis or follow‐up to birth	Routine screening	Prospective cohort	UK
De Graaf 1999	X		X		Not reported	CVS and amniocentesis	High‐risk referral for invasive testing	Case control	Netherlands
Ekelund 2008	X				Not reported	Karyotyping or follow‐up to birth	Routine screening	Cohort	Denmark
Gasiorek‐Wiens 2001			X		Median 33 (15‐49), 36.1% > 35	CVS, amniocentesis or follow‐up to birth	Routine screening	Prospective cohort	Germany, Switzerland and Austria
Gasiorek‐Wiens 2010			X		Median 35.1 (13.2‐46.7)	Karyotyping or follow‐up to birth	Routine screening	Cohort	Germany
Go 2005	X				49% ≤ 35, 51% ≥ 36	Invasive testing or follow‐up to birth	Routine screening	Retrospective cohort	Netherlands
Gyselaers 2005	X		X		Not reported	CVS, amniocentesis or follow‐up to birth	Routine screening	Prospective cohort	Belgium
Habayeb 2010					Median 35.4 (18‐49)	Karyotyping or follow‐up to birth	Routine screening	Cohort	UK
Hadlow 2005*	X				Mean 30.7, 21.2% ≥ 35	CVS, amniocentesis or follow‐up to birth	Routine screening	Prospective cohort	Australia
Hafner 1998*				X	Median 28 (15‐49) 6.9% ≥ 35	Amniocentesis or CVS in patients with previous Down’s pregnancy, > 35 years or with a positive biochemical test result. Other women underwent scan at 22 weeks and, if NT >2.5 mm special examination directed to examination of fetal heart. Follow‐up to birth	Routine screening	Prospective cohort	Austria
Has 2008	X	X	X		Median 28.3 (17‐45)	Karyotyping or follow‐up to birth	Routine screening	Cohort	Turkey
Hewitt 1996				X	Median 37 (21‐48)	CVS	High‐risk referral for invasive testing	Prospective cohort	Australia
Hormansdorfer 2011	X				Mean 31.1 (16‐46), 22% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Germany
Huang 2010	X				Median 30 (15‐47), mean 29.8 (SD 3.3)	Karyotyping or follow‐up to birth	Routine screening	Cohort	Taiwan
Jaques 2007	X				Mean 33 (16‐51), 18.5% ≥ 37	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Australia
Jaques 2010 FTS (first trimester screening)	X				Mean 16.3% ≥ 37	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Australia
Kagan 2010	X		X		Mean 35.4 (14.1‐52.2)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	UK
Kim 2006				X	Mean 29.9 (SD 3.3)	Amniocentesis or CVS in patients considered high risk (NT > 2.5, aged > 35 years, positive biochemical test result, history or chromosomal abnormality, fetal structural abnormality at ultrasound or other reason). Follow‐up to birth	Routine screening	Retrospective cohort	South Korea
Koster 2011	X				Median 37 (IQR 36‐39)	Karyotyping or follow‐up to birth	Routine screening	Case control	Netherlands
Kozlowski 2007 GC (Gynaecologists' practices)	X		X		Median 32 (15‐48), 26.4% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Cohort	Germany
Kozlowski 2007 PC (Prenatal centre)	X		X		Median 34 (14‐46), 43.2% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Cohort	Germany
Krantz 2000*	X		X		34.7% ≥ 35	Not reported	Routine screening	Prospective cohort	USA
Kublickas 2009	X				51% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Sweden
Kuc 2010	X				Not reported	Karyotyping or follow‐up to birth	Routine screening	Case control	Netherlands
Lam 2002			X		Mean 30.5 (19% ≥ 35) for unaffected pregnancies	Women considered high risk offered CVS (0.7%) or amniocentesis (11.8%).    Follow‐up to birth	Routine screening	Prospective cohort	Hong Kong
Leung 2009	X	X			Median 32 (IQR 30‐35), 27.4% ≥ 35	Amniocentesis or follow‐up to birth	Routine screening	Prospective cohort	China
MacRae 2008			X		Not reported	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	UK
Maiz 2007					Median 35 (17‐49)	CVS	High‐risk referral for invasive testing	Prospective cohort	UK
Maiz 2009					Median 34.5 (14.1‐50.1)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	UK
Malone 2004		X			Mean 30.1 (16‐47), 22.1% ≥ 35	Amniocentesis (in women considered high risk, n = 510) or follow‐up to birth	Routine screening	Prospective cohort	USA
Malone 2005	X				21.6% ≥ 35	Amniocentesis offered to women with positive results from any screening test. Follow‐up to birth	Routine screening	Prospective cohort	USA
Marchini 2010*	X				Median 31.3 (18‐45), 19.7% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Italy
Marsis 2004				X	Mean 37.8 (35‐43)	Amniocentesis (unclear in which patients this was conducted) or follow‐up to birth	Screening of patients ≥ 35 years of age	Prospective cohort	Indonesia
Marsk 2006	X		X		Mean 38.5 (SD 4.0) for Down's cases, 35.5 (SD 4.0) for controls	Not reported	Routine screening	Case control	Sweden
Matias 1998					Median 35 (17‐46)	Fetal karyotyping. In cases where NT above 95th percentile or abnormal ductus venousus flow, follow‐up scan conducted at 14‐16 weeks	High‐risk referral for invasive testing	Prospective cohort	UK and Portugal
Matias 2001					Median 35 (17‐46)	Fetal karyotyping. In cases where NT above 95th percentile or abnormal ductus venousus flow, follow‐up scan conducted at 14‐16 weeks	High‐risk referral for invasive testing	Prospective cohort	Portugal
Mavrides 2002			X		Median 35 (15‐42)	CVS or follow‐up	High‐risk referral for invasive testing	Prospective cohort	UK
Maxwell 2011 FTS (first trimester screening cohort)	X				Median 31 (14‐48), 24.3% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Australia
Maymon 2005					Mean 33.7 (SD 4.9) for Down's cases, 30.3 (SD 4.5) for controls	Amniocentesis (recommended for women with higher risk on first or second trimester testing) or follow‐up to birth	Routine screening	Case control	Israel
Maymon 2008	X		X		Not reported	Karyotyping or follow‐up to birth	Routine screening	Case control	USA
Merz 2011	X				Not reported	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Germany
Michailidis 2001				X	Mean 30.1 (13‐50), 21.1% ≥ 35, 11.9% ≥ 37	Karyotyping in women considered at risk due to index test results, age or family history or those with considerable anxiety (632 women, 8.5%) or follow‐up to birth	Routine screening	Prospective cohort	UK
Molina 2010 high risk (High‐risk cohort)		X			Mean 32.7 (16.7‐47.5)	CVS	High‐risk referral for invasive testing	Cohort	Spain
Molina 2010 screening (Screening cohort)	X				Not reported	Karyotyping or follow‐up to birth	Routine screening	Cohort	Spain
Monni 2005			X		Median 32 (14‐49)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Italy
Montalvo 2005	X				Mean 31.1 (14‐49), 25.9% ≥35	Invasive testing offered to women considered high risk from screening results or follow‐up to birth	Routine screening	Prospective cohort	Spain
Moon 2007		X			Mean 35.5 (SD 4.8) for Down's cases, 31.7 (SD 3.4) for unaffected pregnancies	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Korea
Muller 2003	X		X		Not reported	Invasive testing (offered to women with high NT measurement) or follow‐up to birth	Routine screening	Retrospective cohort	France
Nicolaides 1992				X	Median 38 (22‐47)	Fetal karyotyping by amniocentesis (52%) or CVS (48%)	High‐risk referral for invasive testing	Prospective cohort	UK
Nicolaides 2005	X				Median 31 (13‐49)	Amniocentesis or CVS (patients considered high risk based on screening). First trimester presence/absence of nasal bone, presence/absence of tricuspid regurgitation or normal/abnormal Doppler studies (patients of intermediate risk on first trimester screening and did not undergo CVS or amniocentesis. With the addition of information from these tests, if adjusted risk was high, CVS was performed). Follow‐up to birth	Routine screening	Prospective cohort	UK
Niemimaa 2001	X		X		17.5% ≥ 35	Invasive testing (patients considered high risk based on NT screening) or follow‐up to birth.	Routine screening	Prospective cohort	Finland
Noble 1995					Median 34 (15‐47), 47% ≥ 35	Karyotyping performed (27% of women) due to increased NT (14%), advanced maternal age (10%), previous chromosomally abnormal child (0.5%) or parental anxiety (2%).  Ultrasound examination at 20 weeks (65% of patients). Follow‐up to birth (9% of women)	Routine screening in a high risk population	Prospective cohort	UK
O'Callaghan 2000			X		Median 32	CVS, amniocentesis or neonatal karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Australia
O'Leary 2006	X		X		Median 31 (14‐47), 20% ≥ 35	CVS or amniocentesis (women assessed to be high risk on screening) or follow‐up to birth	Routine screening	Prospective cohort	Australia
Okun 2008 FTS (first trimester screening cohort)	X				Mean 34	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Canada
Orlandi 1997	X		X		Range 15 to 46, 35% ≥ 35	Not reported	Routine screening	Prospective cohort	Italy
Orlandi 2003		X			Median 31.7 (SD 4.0) for Down's cases, 36.5 (SD 4.1) for unaffected pregnancies	CVS or amniocentesis (women considered high risk on screening on the basis of NT and biochemical results, but not on nasal bone screening, or if requested due to age or anxiety) or follow‐up to birth	Routine screening (2 centres) or in referred patients (1 centre)	Prospective cohort	Italy and Netherlands
Orlandi 2005		X			Median 30.5 (SD 8.2)	Not reported	Routine screening	Retrospective cohort	Italy
Otaño 2002		X			Median 36 (19‐44)	CVS	High‐risk referral for invasive testing	Prospective cohort	Argentina
Pajkrt 1998			X		Mean 31.4 (SD 5.7), 24% ≥ 35	Prenatal karyotyping offered to patients considered high risk or maternal anxiety (conducted in 24%) or follow‐up to birth	Routine screening	Prospective cohort	Netherlands
Pajkrt 1998a				X	Mean 37.6 (22‐46)	Prenatal karyotyping	High‐risk referral for invasive testing	Consecutive cohort	Netherlands
Palomaki 2007 FTS (first trimester screening cohort)					Mean 32.3 (SD 4.6)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Canada
Perni 2006	X				Median 33.0 (IQR 31.0‐36.0)	CVS or amniocentesis. Cytogenetic testing in cases of miscarriage. Follow‐up to birth.	Routine screening	Retrospective cohort	USA
Prefumo 2005			X		Median 37 (19‐46)	CVS	High‐risk referral for invasive testing	Prospective cohort	UK
Prefumo 2006			X		Mean 31.4 (14.5‐50.2)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	UK
Ramos‐Corpas 2006		X			Mean 30.1 (15‐46) (SD 5.37), 18% ≥ 35	Invasive testing offered to patients considered high risk at screening (> 1:300) or follow‐up to birth	Routine screening	Prospective cohort	Spain
Rissanen 2007	X				29.5, 17.7% ≥35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Finland
Rozenberg 2002			X		Median 30.5 (18‐37)	Amniocentesis offered to patients with NT >3mm or serum marker risk was > 1:250, or follow‐up to birth	Routine screening	Prospective cohort	France
Rozenberg 2007	X				Mean 30.9 (SD 4.5)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Canada
Sahota 2010	X		X		Median 33.1, 30.1% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	China
Salomon 2010	X				Median 30.7 (18.0‐46.3)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	France
Santiago 2007	X		X		Mean 30.6 (14‐46)	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Spain
Sau 2001			X		Mean 28 (SD 5)	Invasive testing (women with high risk on screening) or follow‐up to birth	Routine screening	Prospective cohort	UK
Schaelike 2009	X		X		31.0% ≥35	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Germany
Schielen 2006*	X				Median 36.5 (18‐47)	Invasive testing or follow‐up to birth	Routine screening	Retrospective cohort	Netherlands
Schuchter 2001			X		Mean 28 (15‐46), 10.7% ≥ 35	CVS (offered to patients with first trimester NT > 3.5 mm), amniocentesis (offered to patients with first trimester NT 2.5‐3.4 mm, high risk on second trimester serum testing (> 1:250) and those > 35 years) or follow‐up to birth	Routine screening	Retrospective cohort	Austria
Schuchter 2002	X			X	13% > 35	CVS and amniocentesis (offered to patients with increased risk (> 1:400) at first trimester screening. CVS recommended when NT > 3.5 or when women did not want to wait until the 15^th week for amniocentesis), or follow‐up to birth	Routine screening	Prospective cohort	Austria
Schwarzler 1999			X		Mean 29.4 (16‐47)	Invasive testing (women considered high risk on screening) or follow‐up to birth	Routine screening	Prospective consecutive cohort	UK
Scott 2004	X		X		Median 32 (15‐44), 29% ≥ 35	Invasive testing or follow‐up to birth	Routine screening	Prospective cohort	Australia
Sepulveda 2007		X	X		Median 33 (14‐47), 35.4% ≥ 35	CVS, amniocentesis, cordocentesis or follow‐up to birth	Routine screening	Prospective cohort	Chile
Snijders 1998			X		Median 31 (14‐49)	CVS and amniocentesis (9.6% of women) or follow‐up to birth	Routine screening	Prospective cohort	UK
Sorensen 2011	X				Median 34 (23‐44) for Down's cases; mean 30.4 (16‐45), 16.5% ≥ 35 for unaffected pregnancies	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Denmark
Spencer 1999	X		X	X	Median 38 (19‐46) for Down's cases, 36 (15‐47) for controls	Invasive testing (high‐risk women) or follow‐up to birth	Routine screening	Case control	UK
Spencer 2002					Median 36 (20‐44) for Down's cases, 30 (16‐41) for controls	Not reported	Routine screening	Case control	UK
Spencer 2008	X				Median 35.8 for Down's cases, 29.3 for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	Denmark
Stenhouse 2004	X				Median 32 (14‐45), 27% ≥ 35	Invasive testing offered to women with screening risk of > 1:250 or follow‐up to birth	Routine screening	Prospective cohort	UK
Strah 2008			X		Median 28.6 (15‐42)	Karyotyping or follow‐up to birth	Routine screening	Cohort	Slovenia
Theodoropoulos 1998			X		Median 29 (16‐48), 7.8% ≥ 37	CVS or amniocentesis or follow‐up to birth. Unclear reference standard in cases of intrauterine death, miscarriages and terminations.	Routine screening	Prospective cohort	Greece
Thilaganathan 1999			X		Mean 29 (15‐45)	CVS (offered to patients considered high risk on screening) or follow‐up to birth	Routine screening	Prospective cohort	UK
Timmerman 2010					Mean 34.5 (19‐45)	Karyotyping or follow‐up to birth	Routine screening	Prospective cohort	Netherlands
Torring 2010	X				Mean 35 for Down's cases, 31 for controls	Karyotyping or follow‐up to birth	Routine screening	Case control	Denmark
Vadiveloo 2009	X				Median 33.1, 36.9% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	UK
Valinen 2007	X				Mean 29.6, 18.6% ≥ 35	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	Finland
Viora 2003					Median 32 (18‐47)	CVS or follow‐up to birth	Routine screening	Prospective cohort	Italy
Wald 2003	X		X	X	Not reported	Invasive testing (following second trimester screening) or follow‐up to birth	Routine screening	Case control	UK and Austria
Wapner 2003*	X		X		Mean 35 (SD 4.6), 50% ≥ 35	Invasive testing. Miscarriage with cytogenetic testing. Follow‐up to birth	Routine screening	Prospective cohort	USA
Wax 2009	X				Mean 36.7 (SD 3.2)	Karyotyping or follow‐up to birth	Routine screening	Retrospective cohort	USA
Wojdemann 2005	X		X		Mean 29, 10.8% ≥ 35	Invasive testing (in cases of increased risk) or follow‐up to birth	Referrals for screening	Prospective cohort	Denmark
Wortelboer 2009	X				Median 34.9 (15‐48)	Karyotyping or follow‐up to birth	Routine screening	Cohort	Netherlands
Wright 2008			X		Median 35.2 (16‐52)	Karyotyping or follow‐up to birth	Routine screening	Cohort	UK
Wright 2010	X				Median 31.9 (IQR 27.7‐35.8)	Karyotyping or follow‐up to birth	Routine screening	Cohort	UK, Denmark and Cyprus
Zoppi 2001			X		Median 33 (14‐48)	Amniocentesis, CVS or follow‐up to birth	Routine screening	Prospective cohort	Italy

Correction made for missing false negatives in studies with delayed verification of test negatives	NT			NT and maternal age			NT, PAPP‐A, free ßhCG and maternal age
	Ratio of DORs (95% CI); P value	Sensitivity at 5% FPR (95% CI) (studies)		Ratio of DORs (95% CI); P value	Sensitivity at 5% FPR (95% CI) (studies)		Ratio of DORs (95% CI); P value	Sensitivity at 5% FPR (95% CI) (studies)
	Ratio of DORs (95% CI); P value	Screening (n = 9)	High risk (n = 4)	Ratio of DORs (95% CI); P value	Screening (n = 46)	High risk (n = 4)	Ratio of DORs (95% CI); P value	Screening (n = 66)	High risk (n =3)
No FN correction	0.68 (0.26, 1.77); P = 0.40	73 (62, 81)	64 (45, 80)	0.34 (0.17, 0.69); P = 0.003	72 (68, 76)	47 (31, 63)	0.41 (0.16, 1.00); P = 0.05	88 (86, 89)	74 (54, 88)
FN increased +10%	0.69 (0.27, 1.78); P = 0.40	70 (59, 79)	62 (42, 78)	0.40 (0.20, 0.82); P = 0.01	69 (64, 73)	47 (31, 64)	0.48 (0.19, 1.20); P = 0.11	86 (84, 87)	74 (53, 88)
FN increased +20%	0.74 (0.29, 1.92); P = 0.50	69 (57, 78)	62 (42, 78)	0.43 (0.21, 0.89); P = 0.02	67 (63, 71)	47 (31, 64)	0.51 (0.20, 1.28); P = 0.15	85 (83, 87)	74 (54, 88)
FN increased +30%	0.81 (0.31, 2.09); P = 0.63	67 (55, 76)	62 (42, 78)	0.46 (0.22, 0.97); P = 0.04	66 (61, 70)	47 (30, 64)	0.55 (0.22, 1.38); P = 0.20	84 (82, 86)	74 (54, 88)
FN increased +40%	0.76 (0.29, 2.02); P = 0.55	66 (53, 76)	59 (39, 77)	0.50 (0.24, 1.02); P = 0.06	64 (60, 68)	47 (31, 64)	0.59 (0.24, 1.48); P = 0.26	83 (81, 85)	74 (54, 88)
FN increased +50%	0.81 (0.30, 2.15): P = 0.65	64 (52, 75)	59 (39, 77)	0.52 (0.25, 1.08); P = 0.08	63 (58, 67)	47 (30, 64)	0.62 (0.25, 1.56); P = 0.31	82 (80, 84)	74 (54, 88)

Test	No. of studies	No. of participants
1 Aberrant right subclavian artery	1	425
2 Frontomaxillary facial angle >95 percentile	1	242
3 Presence of mitral gap	1	217
4 Maxillary bone length, 5% percentile	1	927
5 Tricuspid regurgitation	1	312
6 Iliac angle 90 degrees	1	2032
7 Ductus venosus a‐wave reversed	1	378
8 Ductus venosus pulsivity index > 95 percentile	1	378
9 Nasal bone, mixed cut‐points	11	48279
10 NT, 2.5 mm	4	11835
11 NT, 3 mm	6	10381
12 NT, 5FPR	3	63885
13 NT, mixed cut‐points	13	90978
14 NT and age, risk 1:100	1	10668
15 NT and age, risk 1:250	10	79412
16 NT and age, risk 1:300	23	252811
17 NT and age, 1FPR	4	98453
18 NT and age, 3FPR	4	98453
19 NT and age, 5FPR	22	288853
20 NT and age, mixed cut‐points	50	530874
21 NT and nasal bone, Absent NB + NT ≥ 95th centile	1	486
22 Ductus and age, risk 1:250	1	3731
23 Ductus and age, 5FPR	2	3965
24 Ductus and age, mixed cut‐points	5	5331
25 Ductus, NT and age, risk 1:100	1	19736
26 Ductus, NT and age, risk 1:250	1	3727
27 Ductus, NT and age, 5FPR	2	3961
28 Ductus, NT and age, mixed cut‐points	3	23697
29 Age and nasal bone, mixed cut‐points	4	25303
30 Age, NT and tricuspid blood flow, risk 1:100	1	19736
31 Age, NT and nasal bone, risk 1:100	1	19736
32 Age, NT and nasal bone, risk 1:300	4	9963
33 Age, NT and nasal bone, mixed cut‐points	5	29699
34 Age, NT, nasal bone and ductus, risk NT>1:300 AND abnormal DV flow AND absent NB	1	544
35 Age, NT, nasal bone, free ßhCG and PAPP‐A, 1st trimester, 5FPR	1	20305
36 Age, NT, nasal bone, free ßhCG and PAPP‐A, 1st trimester, mixed cut‐points	3	41842
37 Age, NT and free ßhCG, 1st trimester, 5FPR	4	4986
38 Age, NT and free ßhCG, 1st trimester, risk 1:240	1	5809
39 Age, NT and free ßhCG, 1st trimester, mixed cut‐points	5	10795
40 Age, NT and PAPP‐A, 1st trimester, risk 1:100	1	1507
41 Age, NT and PAPP‐A, 1st trimester, risk 1:185	1	5809
42 Age, NT and PAPP‐A, 1st trimester, 5FPR	3	2498
43 Age, NT and PAPP‐A, 1st trimester, mixed cut‐points	5	9814
44 Age, NT and total hCG, 1st trimester, 5FPR	1	1110
45 Age, NT and AFP, 1st trimester, 5FPR	1	1110
46 Age, NT and ITA, 1st trimester, 5FPR	1	278
47 Age, NT and inhibin, 1st trimester, risk 1:100	1	40
48 Age, NT and inhibin, 1st trimester, risk 1:250	1	40
49 Age, NT and inhibin, 1st trimester, risk 1:400	1	40
50 Age, NT and inhibin, 1st trimester, 5FPR	1	1110
51 Age, NT and inhibin, 1st trimester, mixed cut‐points	2	1150
52 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:100	10	102332
53 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:150	5	177643
54 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:200	8	135768
55 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:220	1	2231
56 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:250	25	174712
57 Age, NT, PAPP‐A and free ßhCG, 1st trimester, risk 1:300	29	544681
58 Age, NT, PAPP‐A and free ßhCG, 1st trimester, 1FPR	7	88874
59 Age, NT, PAPP‐A and free ßhCG, 1st trimester, 3FPR	9	312680
60 Age, NT, PAPP‐A and free ßhCG, 1st trimester, 5FPR	24	391874
61 Age, NT, PAPP‐A and free ßhCG, 1st trimester, mixed cut‐points	69	1173853
62 Age, NT, PAPP‐A and uE3, 1st trimester, 5FPR	1	576
63 Age, NT, PAPP‐A and ITA, 1st trimester, 5FPR	2	11053
64 Age, NT, PAPP‐A and inhibin, 1st trimester, risk 1:100	1	40
65 Age, NT, PAPP‐A and inhibin, 1st trimester, risk 1:250	1	40
66 Age, NT, PAPP‐A and inhibin, 1st trimester, risk 1:400	1	40
67 Age, NT, PAPP‐A and inhibin, 1st trimester, 5FPR	1	1110
68 Age, NT, PAPP‐A and inhibin, 1st trimester, mixed cut‐points	2	1150
69 Age, NT, PAPP‐A and ADAM12, 1st trimester, 5FPR	2	1042
70 Age, NT, PAPP‐A and ADAM12, 1st trimester, risk 1:250	1	691
71 Age, NT, free ßhCG and ADAM12, 1st trimester, 5FPR	1	351
72 Age, NT, AFP and free ßhCG, 1st trimester, risk 1:250	1	1656
73 Age, NT, AFP and free ßhCG, 1st trimester, 5FPR	1	1110
74 Age, NT, AFP and free ßhCG, 1st trimester, mixed cut‐points	2	2766
75 Age, NT, AFP and PAPP‐A, 1st trimester, 5FPR	1	1110
76 Age, NT, total hCG and PAPP‐A, 1st trimester, 5FPR	1	1110
77 Age, NT, total hCG and inhibin, 1st trimester, 5FPR	1	1110
78 Age, NT, free ßhCG and inhibin, 1st trimester, 5FPR	1	1110
79 Age, NT, PAPP‐A, free ßhCG, 1st trimester serum, ductus venosus pulsivity index, 5FPR	1	7250
80 Age, free ßhCG and PAPP‐A, if risk 1:42‐1:1000, NT, final 1:250 risk	1	10189
81 Age, NT, ductus, free ßhCG and PAPP‐A, 1st trimester, risk 1:100	2	26986
82 Age, NT, ductus, free ßhCG and PAPP‐A, 1st trimester, risk 1:250	2	10325
83 Age, NT, ductus, free ßhCG and PAPP‐A, 1st trimester, 5FPR	2	10325
84 Age, NT, ductus, free ßhCG and PAPP‐A, 1st trimester, mixed cut‐points	3	30061
85 Age, NT, nasal bone, free ßhCG and PAPP‐A, 1st trimester, risk 1:100	1	19736
86 Age, NT, nasal bone, free ßhCG and PAPP‐A, 1st trimester, risk 1:300	1	1801
87 Age, NT, tricuspid blood flow, free ßhCG and PAPP‐A, 1st trimester, risk 1:100	1	19736
88 Age, NT, fetal heart rate, free ßhCG and PAPP‐A, 1st trimester, 5FPR	2	76385
89 Age, NT, fetal heart rate, nasal bone, free ßhCG and PAPP‐A, 1st trimester, risk 1:200	1	19736
90 age, NT, fetal heart rate, ductus, free ßhCG and PAPP‐A, 1st trimester, 5FPR	1	19614
91 Age, NT, fetal heart rate, tricuspid blood flow, free ßhCG and PAPP‐A,1st trimester, 5FPR	1	19736
92 Age, NT, AFP, free ßhCG and PAPP‐A, 1st trimester, risk 1:250	1	5483
93 Age, NT, AFP, free ßhCG and PAPP‐A, 1st trimester, 5FPR	2	1306
94 Age, NT, AFP, free ßhCG and PAPP‐A, 1st trimester, mixed cut‐points	3	6789
95 Age, NT, total hCG, inhibin and PAPP‐A, 1st trimester, 5FPR	1	1110
96 Age, NT, PAPP‐A, free ßhCG and PGH, 1st trimester, 5FPR	1	335
97 Age, NT, PAPP‐A, free ßhCG and GHBP, 1st trimester, 5FPR	1	335
98 Age, NT, PAPP‐A, free ßhCG and PIGF, 1st trimester, 5FPR	2	1443
99 Age, NT, PAPP‐A, free ßhCG and total hCG, 1st trimester, 5FPR	1	998
100 Age, NT, PAPP‐A, free ßhCG and PP13, 1st trimester, 5FPR	1	998
101 Age, NT, PAPP‐A, free ßhCG and ADAM12, 1st trimester, 5FPR	4	2571
102 Age, NT, PAPP‐A, free ßhCG and ADAM12, 1st trimester, risk 1:250	2	1222
103 Age, NT, PAPP‐A, free ßhCG and ADAM12, 1st trimester, mixed cut‐points	4	2571
104 Age, NT, free ßhCG, PAPP‐A and inhibin, 1st trimester, risk 1:100	1	40
105 Age, NT, free ßhCG, PAPP‐A and inhibin, 1st trimester, risk 1:250	1	40
106 Age, NT, free ßhCG, PAPP‐A and inhibin, 1st trimester, risk 1:400	1	40
107 Age, NT, free ßhCG, PAPP‐A and inhibin, 1st trimester, 5FPR	1	1110
108 Age, NT, PAPP‐A, free ßhCG, ADAM12 and PlGH, 1st trimester, 5FPR	1	998
109 Age, NT, total hCG, inhibin, PAPP‐A, AFP and uE3, 1st trimester, 5FPR	1	1110
110 Age, NT, free ßhCG, inhibin, PAPP‐A, AFP and uE3,1st trimester, 5FPR	1	1110
111 Age, NT, PAPP‐A, free ßhCG, ADAM12, total hCG and PlGF, 1st trimester, 5FPR	1	998
112 Age, NT, PAPP‐A, free ßhCG, ADAM12, total hCG, PlGF and PP13, 1st trimester, 5FPR	1	998
113 NT, free ßhCG and PAPP‐A, 1st trimester incidence rate 63.3%	1	6508
114 NT, PAPP‐A, free ßhCG and maternal age ‐ maternal age < 35 years	5	19057
115 NT, PAPP‐A, free ßhCG and maternal age ‐ maternal age ≥ 35 years	5	10980

Clinical features and settings	High‐risk referral for invasive testing
Participants	230 participants Brazil ‐ private centres Dates not specified Pregnant women Mean age 35.8 years (21‐45 years) Singleton pregnancies Karyotyping performed at same time as NT 10‐14 weeks' gestation
Study design	Diagnostic validation study to determine the best ROC cut‐off for NT Retrospective study of patient notes
Target condition and reference standard(s)	Down's syndrome: 12 cases Reference standards: chorionic villus biopsy, amniocentesis or blood or placenta used for fetal karyotyping
Index and comparator tests	NT with cut‐off of 2.5 mm (found to be optimum cut‐off from ROC) (Sequoia, Aspen 128XP10‐Acuson and Toshiba SH140)
Follow‐up	100% karyotyping
Aim of study	To define the best fixed cut‐off point for NT, and the accuracy of this cut‐off for all fetal aneuploidy screening and for trisomy of chromosome 21
Notes
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Selective testing of high‐risk women as done in practice
Acceptable reference standard?   All tests	Yes	Karyotyping
Partial verification avoided?   All tests	Yes	All patients received a reference standard
Differential verification avoided?   All tests	No	Women had different reference standard
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Yes	Index test interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	No	No details given for test failures/uninterpretable measurements
Withdrawals explained?   All tests	No	No details of withdrawals given

Clinical features and settings	Routine referral
Participants	6906 participants with complete outcome data Germany ‐ gynaecologists practices January 2000 ‐ December 2003 Pregnant women Median maternal age 32 years (15‐48 years), 26.4% ≥ 35 years 11‐14 weeks' gestation
Study design	Cohort study
Target condition and reference standard(s)	Down's syndrome: 19 cases in gynaecologists practices Reference standard: karyotyping or follow‐up to birth
Index and comparator tests	Maternal age First trimester NT (FMF certified gynaecologists) First trimester free ßhCG and PAPP‐A (Kryptor analyser, Brahms) Risk cut‐point 1:300
Follow‐up	Data on pregnancy outcome were obtained by contacting the patient or their general gynaecologist. Women without complete outcome data (36%) were excluded from the study
Aim of study	To evaluate and compare the screening performance for fetal trisomy 21 in the first trimester of pregnancy in general gynaecologists practices and specialised centres for prenatal care in Germany
Notes
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Routine screening of typical pregnant population
Acceptable reference standard?   All tests	Yes	Karyotyping or follow‐up to birth
Partial verification avoided?   All tests	Yes	All women received a reference standard
Differential verification avoided?   All tests	No	Choice of reference standard depended on index test results
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Yes	Index test interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	No	No details given for test failures/uninterpretable measurements
Withdrawals explained?   All tests	Yes	146 women (including 11 with down's syndrome) excluded as results could not be assigned to gynaecologists' or prenatal centre group

Clinical features and settings	Routine screening in a high‐risk population
Participants	2529 participants UK October 1994 to April 1995 Pregnant women Singleton pregnancies Median age 34 years (15‐47 years), 47% ≥ 35 years 10‐14 weeks' gestation
Study design	Prospective cohort
Target condition and reference standard(s)	Down's syndrome: 61 cases Reference standards: karyotyping performed (27% of women) due to increased NT (14%), advanced maternal age (10%), previous chromosomally abnormal child (0.5%) or parental anxiety (2%). Ultrasound examination at 20 weeks (65% of patients). Follow‐up to birth (9% of women)
Index and comparator tests	Maternal age First trimester NT (methods not stated) Fresh serum (or serum frozen over a weekend) tested for: First trimester free ßhCG (immunoradiometric assay, CIS)
Follow‐up	Pregnancy outcome obtained from maternity units or the patients themselves. Follow‐up information only appears to have been obtained in 9% of cases (second trimester ultrasound used as reference standard for other women)
Aim of study	To measure the contribution of maternal serum free beta hCG in a screening programme for fetal trisomy 21 based on fetal NT in the first trimester of pregnancy
Notes	No proper results data are presented for this study
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Routine screening of typical pregnant population
Acceptable reference standard?   All tests	No	Invasive testing, ultrasound at 20 weeks or follow‐up to birth
Partial verification avoided?   All tests	Unclear	Unclear if all women received a reference standard
Differential verification avoided?   All tests	No	Choice of reference standard depended on index test results
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Yes	Index test interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	No	No details given for test failures/uninterpretable measurements
Withdrawals explained?   All tests	No	No details of withdrawals given

Clinical features and settings	Routine screening of general and high‐risk women
Participants	2010 participants (744 in subgroup undergoing NT testing) Italy Dates not reported Recruited through private physician or genetic counselling program for women of advanced maternal age Pregnant women Aged 15‐46 years, 35% ≥ 35 years Singleton pregnancies 9‐13 weeks' gestation
Study design	Prospective cohort
Target condition and reference standard(s)	Down's syndrome: 11 cases (7 in subgroup with NT testing) Reference standards: not reported
Index and comparator tests	Maternal age First trimester NT (37% of patients) (FMF methods, Toshiba SSA 250A or Acuson XP 10) First trimester free ßhCG and PAPP‐A (all patients) (dried blood samples, enzyme‐linked immunosorbent assays) Risk cut‐point 1:380
Follow‐up	Not reported
Aim of study	To evaluate first trimester combined screening for Down's syndrome
Notes	Unclear as to what reference standard (if any) was used. All cases of Down's syndrome identified had been picked up by screening
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Routine screening of typical pregnant population
Acceptable reference standard?   All tests	Unclear	Reference standard not reported
Partial verification avoided?   All tests	Unclear	Unclear if all women received a reference standard
Differential verification avoided?   All tests	Unclear	Unclear if the choice of reference standard depended on screening results
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Yes	Index test interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	No	No details given for test failures/uninterpretable measurements
Withdrawals explained?   All tests	Yes	Details given of women undergoing NT but not biochemical testing

Clinical features and settings	Women referred for invasive testing or self‐referred for screening
Participants	1156 participants: 210 cases and 946 controls matched for gestational and maternal age UK ‐ fetal medicine research centre Dates not specified Pregnant women Median maternal age 38 years (19‐46 years) (cases) and 36 years (15‐47 years) (controls) 10‐14 weeks' gestation
Study design	Case‐control study
Target condition and reference standard(s)	Down's syndrome: 210 cases Reference standards: invasive testing (high‐risk women) or follow‐up to birth
Index and comparator tests	Maternal age First trimester NT (methods not reported) Frozen serum samples tested for: First trimester free ßhCG and PAPP‐A (Kryptor analyser, time resolved amplified cryptate emission (TRACE))
Follow‐up	Stated that pregnancy outcome was ascertained in all women
Aim of study	To examine the potential impact of combining maternal age with fetal NT thickness and maternal serum free ßhCG and PAPP‐A in screening for trisomy 21 at 10‐14 weeks of gestation
Notes
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Selective testing of high‐risk women as done in practice
Acceptable reference standard?   All tests	Yes	Karyotyping or follow‐up to birth
Partial verification avoided?   All tests	Yes	All women received a reference standard
Differential verification avoided?   All tests	No	Choice of reference standard depended on index test results
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Unclear	Unclear if all index tests interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	No	No details given for test failures/uninterpretable measurements
Withdrawals explained?   All tests	No	No details of withdrawals given

Clinical features and settings	High‐risk referral
Participants	445 fetuses with increased risk based on NT or biochemical testing and information available on pregnancy outcome The Netherlands ‐ fetal medicine unit September 1996 ‐ March 2008 Mean maternal age 34.5 years (19‐45 years) First trimester
Study design	Prospective cohort
Target condition and reference standard(s)	Down's syndrome: 72 cases Reference standards: karyotyping or follow‐up to birth
Index and comparator tests	First trimester ductus venosus pulsatility index and Ductus venosus a‐wave (methods reported elsewhere)
Follow‐up	Pregnancy outcome was obtained from standard follow‐up forms filled in and returned by patients, maternity wards or midwife practices and by reviewing neonatal, pathology and clinical paediatric notes. When the baby was born without structural defects or dysmorphic features, the chromosomes were assumed to be normal. In all cases of enlarged NT or antenatal suspicion of abnormal development, the infant was investigated by a neonatologist, paediatric cardiologist or geneticist.
Aim of study	To investigate if ductus venosus pulsatility index for veins and a‐wave measurements can increase the accuracy of first trimester Down's syndrome screening in a high‐risk population
Notes
*Table of Methodological Quality*
Item	Authors' judgement	Description
Representative spectrum?   All tests	Yes	Selective testing of high‐risk women as done in practice
Acceptable reference standard?   All tests	Yes	Karyotyping or follow‐up to birth
Partial verification avoided?   All tests	Yes	All women received a reference standard
Differential verification avoided?   All tests	No	Choice of reference standard depended on index test results
Incorporation avoided?   All tests	Yes	Reference standard was independent of the index test
Reference standard results blinded?   All tests	No	Reference standard interpreted with knowledge of index test results
Index test results blinded?   All tests	Yes	Index test interpreted without knowledge of reference standard results
Relevant clinical information?   All tests	Yes	Information available as would be in standard clinical practice
Uninterpretable results reported?   All tests	Yes	Satisfactory waveform measurements made in 98% of cases
Withdrawals explained?   All tests	No	No details of withdrawals given

PERMALINK

First trimester ultrasound tests alone or in combination with first trimester serum tests for Down's syndrome screening

S Kate Alldred

Yemisi Takwoingi

Boliang Guo

Mary Pennant

Jonathan J Deeks

James P Neilson

Zarko Alfirevic

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Target condition being diagnosed

Down’s syndrome

Screening tests for Down's syndrome

Screening and parental choice

Index test(s)

Alternative test(s)

Rationale

Objectives

Investigation of sources of heterogeneity

Methods

Criteria for considering studies for this review

Types of studies

Participants

Index tests

Target conditions

Reference standards

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of methodological quality

Statistical analysis and data synthesis

Estimation of average sensitivity and specificity

Comparisons between tests

Investigations of heterogeneity

Sensitivity analyses

Results

Results of the search

Ultrasound and triple serum markers

Ultrasound and double serum markers

Ultrasound and single serum markers

Ultrasound markers alone

Methodological quality of included studies

1.

Findings

Summary of findings 1. Performance of the 10 most evaluated first trimester ultrasound markers alone or in combination with first trimester serum tests.

2.

3.

4.

Summary of findings 2. Performance of other first trimester ultrasound markers alone or in combination with first trimester serum tests.

Comparative analysis of the 10 selected test strategies

5.

1. Direct (head‐to‐head) comparisons of the diagnostic accuracy of the 10 most evaluated first trimester ultrasound markers alone or in combination with first trimester serum tests.

2. Indirect comparisons of the diagnostic accuracy of the 10 most evaluated first trimester ultrasound markers alone or in combination with first trimester serum tests.

Investigation of heterogeneity and sensitivity analyses

3. Summary of study characteristics.

4. Investigation of the effect of type of population.

6.

7.

Discussion

Summary of main results

Strengths and weaknesses of the review

Applicability of findings to the review question

Authors' conclusions

Implications for practice.

Implications for research.

Notes

Acknowledgements