Types of studies

We will include:

• retrospective case‐control studies recruiting a group of women with known Down’s syndrome pregnancies and a group of women with unaffected pregnancies, where the index test(s) are analysed and compared between the two groups;

• prospective and retrospective cohort studies including direct (head to head) studies in which all women from a given population have one or more index test(s) and reference standard, and in whom a pregnancy outcome is recorded; and

• randomised controlled trials in which women are randomised to one of several index tests / package of index tests and all receive a reference standard.

We will exclude studies in which no useful data can be extracted from the paper.

Participants

Pregnant women at less than 24 weeks gestation confirmed by ultrasound, who have not undergone previous testing for Down’s syndrome in this pregnancy, and those in special groups which pose either increased risk of Down’s syndrome or difficulty with conventional screening tests including:

• maternal age greater than 35 years old;

• multiple pregnancy;

• diabetes Mellitus; and

• family History of Down’s syndrome.

Index tests

Tests will be divided into serum tests, urine tests and sonographic tests and categorised by timing with regard to gestation.

• Serum tests at less than 24 weeks gestational age which may include, amongst others:

• • ßhCG / free ßhCG;

  • total hCG;

  • AFP;

  • uE3;

  • PAPP‐A;

  • Inhibin‐A; and

  • ADAM‐12.

• Urine tests which may include, amongst others:

• • invasive trophoblast antigen;

  • ß‐core fragment;

  • free ßhCG; and

  • total hCG.

• Sonographic tests at less than 14 weeks gestational age which may include, amongst others:

• • nuchal translucency;

  • absent nasal bone;

  • ductus venosus studies;

  • tricuspid regurgitation/A‐V regurgitation; and

  • abnormalities such as cyctic hygroma, as defined by individual studies.

We will look at comparisons of tests in isolation and in various combinations, including those which form part of tests involving more than one stage.  These may include tests such as double/triple/quadruple tests, first trimester and second trimester screening, combined test, integrated screening and sequential screening (i.e. independent, step‐wise, contingent and repeated measures screening).

Where tests are used in comparison we will look at the performance of test comparisons according to predicted probabilities computed using a risk equation.  Where individual patient data is available we will compute risks using a standardised equation.

Target conditions

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

Reference standards

We will consider one of several reference standards.  The primary reference standards will be genetic verification using the following:

• amniocentesis;

• chorionic villus sampling;

• postnatal karyotyping; and

• miscarriage with cytogenetic testing of the fetus.

Secondary reference standards will include birthing registers and Down’s syndrome registers, and diagnosis based on postnatal macroscopic inspection.  It is expected that most studies will suffer differential verification as they will only use genetic testing on fetuses considered high risk according to the screening test; the reference standard for most unaffected infants being observing a phenotypically normal baby.

Electronic searches

We will apply a sensitive search strategy to search the following databases using the text words and MeSH terms listed below, adapting the search strategy for each different database.

Databases to be searched include:

• MEDLINE via OVID (1980 to current date);

• EMBASE via Dialog Datastar (1980 to current date);

• BIOSIS via EDINA (1985 to current date);

• CINAHL via OVID (1982 to current date);

• the Database of Abstracts of Reviews of Effectiveness (The Cochrane Library);

• MEDION (http://www.mediondatabase.nl/);

• Database of Systematic Reviews and Meta‐Analyses in Laboratory Medicine (http://www.ifcc.org/);

• the National Research Register (http://www.nrr.nhs.uk/); and

• Health Services Research Projects in Progress database (http://www.nlm.nih.gov/hsrproj/).

The search strategy will combine three sets of search terms (see Appendix 1). The first set will be made up of named tests, general terms used for screening/diagnostic tests and statistical terms. Note that the statistical terms are used to increase sensitivity and are not used as a methodological filter used to increase specificity. The second set will be 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 will be combined with the Boolean operator OR and then the three sets will be combined using AND.  The terms used will be a combination of subject headings and free text terms.  The search strategy will be adapted to suit each database searched.

We will attempt to identify cumulative papers which report data from the same dataset, and contact 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 will examine references cited in studies identified as being potentially relevant, and those cited by previous reviews.  We will contact authors of studies where further information is required.  We will not apply a diagnostic test filter, and we will not apply language restrictions to the search. 

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

Selection of studies

Two review authors will screen the titles and abstracts, where available, of all studies identified by the search strategy.  We will obtain full text versions of all studies identified as being potentially relevant, and they will be independently assessed by two review authors for inclusion, using a study eligibility screening proforma a based on pre‐specified inclusion criteria.  Any disagreement between the two review authors will be settled by consensus, or by a third party where necessary.

Data extraction and management

We will develop a data extraction form, to aid collection of relevant and important data from included studies.  One review author will independently extract the data, and where difficulty or uncertainty exists, a second review author will validate the extracted information.

Whenever possible, data will be treated as dichotomous, 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 data are available at several cut‐points we will record these.  In addition, where combinations of tests are used, the overall result, from combining the test into a risk score, will be used to determine whether a test is positive or negative, and not the individual components that make up the test.

We will differentiate between data based on observation of test performance in patients in study groups, and data derived from statistical models (calculations of expected screening performance of tests).  Where possible we will extract summaries of distributions of markers and correlations of distributions of markers.  Individual patient data will be obtained, wherever possible, in order to compute classifications of risk in a standardised way.

Assessment of methodological quality

We will use a modified version of the QUADAS tool (Whiting 2003), which is a quality assessment tool for use in systematic reviews of diagnostic accuracy studies, to assess the methodological quality of included studies.  Two review authors will separately assess each included study.  Any disagreement between the two review authors will be settled by consensus, or by a third party where necessary. Each item in the QUADAS tool will be marked as ‘yes’, ‘no’ or ‘unclear’, and scores will be presented graphically and in tables.  We will not use a summary quality score.

QUADAS criteria will include the following questions.

1. Was the spectrum of patients representative of the patients who will receive the test in practice?

   1. Was the sample selected from a wide range of ages (i.e. childbearing age)? 

   2. Was the sample selected from a specified ‘high‐risk’ group? That is:

      • over 35;

      • family history of Down’s syndrome;

      • multiple pregnancy;

      • diabetes mellitus.

   3. Was the sample taken from a select or unrepresentative group of patients (i.e. private practice)/atypical screening population?

   4. Were all affected and unaffected fetuses included that could be tested at the time point when the screening test would be applied, or did recruitment occur at a later time point when selection could be affected by selective fetal loss?

2. Is the reference standard (amniocentesis, chorionic villus sampling, postnatal karyotyping, miscarriage with cytogenetic testing of the fetus or phenotypically normal baby) likely to correctly classify the target condition?

3. Did the whole sample or a random selection of the sample receive verification using a reference standard of diagnosis?

4. Did patients receive the same reference standard regardless of the index test result?

5. Was the reference standard independent of the index test result (i.e. the index test did not form part of the reference standard)?

6. Were the index test results interpreted without knowledge of the results of the reference standard?

7. Were the reference standard results interpreted without knowledge of the results of the index test?

8. 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?

9. Were uninterpretable/intermediate test results reported?

10. Were withdrawals from the study explained?

11. Were all patients found to be at high risk offered definitive testing/reference standard?

12. Other unspecified quality concerns include:

    1. source of medians used for age and laboratory standardisation? How frequently they are updated; are they based on retrospective data or study data; are they generated locally or published value?

    2. for evaluation of risk scores, was assessment made on a separate validation sample or using appropriate methods to correct for over‐optimism?

Statistical analysis and data synthesis

The methods used for meta‐analysis will be contingent on the data obtained.  Where dichotomised data are available, we will combine studies using bivariate hierarchical methods to obtain direct estimates or detection rates (sensitivity) and false positive rates (1‐specificity) and the associated uncertainty.  We will compare tests and test combinations by including co‐variates in the model.  Data on studies which make direct comparisons between tests will be analysed separately. 

Additional analyses will be undertaken for individual markers comparing distributions between unaffected and Down’s syndrome fetuses after appropriate transformation of parameters, using standard mean difference meta‐analytical methods.

Data on event rates, such as harms of testing, will be displayed using forest plots and pooled using standard meta‐analytical methods, taking account of heterogeneity between studies.

Investigations of heterogeneity

We will compare results between studies which derive risk scores and those that validate them; and between studies which report actual observed results and those in hypothetical standardised populations. 

Sensitivity analyses

We will consider the impact of age standardisation, improvements in technology and adjustment for fetal loss in sensitivity analyses.