Prenatal reflex DNA screening for trisomy 21, 18 and 13

KEYWORDS:

• Prenatal screening
• reflex DNA screening
• NIPT
• Down syndrome
• trisomy 21
• trisomy 18
• trisomy 13

1. Sequential DNA screening

The accuracy of a prenatal screening test is determined by the detection rate (proportion of affected individuals with a positive result) and the false-positive rate (the proportion of unaffected individuals with a positive result) [1]. Any modification in the method of screening, including a change in the screening protocol, that improves the detection rate and reduces the false-positive rate is desirable, provided it is affordable, since it improves efficacy and reduces the harm and distress arising from alerting people with false-positive results. DNA analysis of maternal plasma in prenatal screening for trisomies 21, 18, and 13 (affected pregnancies) has a high detection rate and a low false-positive rate, but it is costly and there is a technical failure rate ranging from about 9–0.2% depending on DNA method and the definition of a failed test [2,3]. These limitations have prompted a sequential testing approach to screening, in which women first have a conventional screening test such as the Combined test (the measurement of pregnancy-associated plasma protein-A and free beta-human chorionic gonadotrophin together with the ultrasound marker nuchal translucency, and maternal age); those with a risk of having an affected pregnancy above a specified level have a plasma DNA analysis performed. Because the DNA test has a high detection rate (about 98%), only about 2% of affected pregnancies are missed in the women who have a sequential DNA test, but the false-positive rate is substantially reduced, so that women offered an invasive diagnostic test (e.g. amniocentesis) have a high probability of having an affected pregnancy. Two screening protocols have been proposed to implement the sequential approach: the recall method and the reflex method.

2. Recall versus reflex DNA screening

In the ‘recall method’ [4], women provide blood for the Combined test and those designated screen-positive are recalled for counseling and offered a DNA screening test or an invasive diagnostic test. In the ‘reflex method’ [2,5,6] women provide blood for the Combined test and at the same venipuncture, a plasma sample is held for a DNA analysis to be performed if the Combined test yields a risk of having an affected pregnancy above a specified cut-off level (e.g. 1 in 800 which generates an approximate 10% reflexing rate). A positive Combined test result is not reported and there is no need to recall women for a second blood sample. Figure 1 illustrates the reflex method.

Figure 1. Flowchart of reflex DNA screening protocol (adapted from Wald et al., Genet Med) [2].

With the recall method, women are given a screening result based on the Combined test (positive or negative). Using a risk cut-off of 1 in 150 to define a positive Combined test, about 3% of women screened are recalled for counseling and worried by being told they have a positive result. With the reflex method, this worry is avoided because the Combined test result is only reported if it is negative; a positive test result is not reported at this stage. Instead, a DNA analysis on the plasma sample already provided is automatically triggered, that is, reflexed. The Combined test risk cut-off that triggers the DNA analysis is ideally lower than that used in the recall method, to increase the detection rate because, in contrast with the recall method, lowering the Combined test risk cut-off has a minimal effect on the false-positive rate [2]. The reflex DNA test result amalgamates information from the Combined test markers with the DNA marker, and a positive or negative screening result is issued accordingly. With the reflex method, few women with unaffected pregnancies receive a positive screening result and the cut-off used to trigger a DNA analysis can be reduced as DNA analysis becomes less expensive and more robust.

The recall method is essentially two screening tests in the same pregnancy and reports two risks, which is confusing and therefore can cause anxiety. The reflex method is a single test, reporting one risk per pregnancy. About 1 in 30 women with unaffected pregnancies receive a positive result with the recall method, compared with about 1 in 5000 with the reflex method. Using the same Combined test risk cut-off considerably fewer women are distressed with the reflex method, without loss of detection, a significant benefit, and by lowering the Combined test risk cut-off detection rates can be increased without having any material effect on the false-positive rate. Also with the reflex method fewer women with unaffected pregnancies have an invasive diagnostic test than with the recall method [2,4]. This arises because with the recall method some of the women recalled are sufficiently distressed to opt for a diagnostic test directly instead of having a DNA screening test [4].

The reflex method therefore achieves a higher detection rate than the recall method by allowing the use of a low reflexing Combined test risk cut-off to trigger a DNA test, and this has a minimal impact on the much lower false-positive rate of the reflex method [2,4].

The reflex DNA method can be configured to be more cost-effective than the recall method. The extra cost of holding plasma samples on all women is modest if EDTA tubes are used. Experience at the Wolfson Institute has shown that the test failure rate remains low if centrifugation of the blood sample occurs within 60 h. Against this modest cost the much lower false-positive rate means that there is a significant saving in counseling costs, first after the Combined test, and again after the DNA test. Also, the higher cost of some women in the recall method choosing an invasive diagnostic test after a positive Combined test result instead of a DNA screening test is avoided.

3. Implications for screening policy

Reflex DNA screening has been implemented in five maternity units in England on a routine service basis, with a 95% detection rate and a 0.02% false-positive rate [2]. In the project, a Combined test risk cut-off of 1 in 800 was used to trigger a reflex DNA analysis. This screening performance is substantially better than with the Combined test alone (81% detection rate with a 2.4% false-positive rate [2]), and better than the use of the Integrated test (90% detection rate for Down syndrome with a 1.8% false-positive rate) [6]. Universal DNA screening achieves a higher detection rate than reflex DNA screening (98% vs. 95%) [2,3], but the financial cost is greater, more women are faced with technical assay failures, and the false-positive rate is higher. If 10% of women are reflexed to a DNA analysis, technical failures and false-positives affect 10-fold fewer screened women. If the cost of DNA analysis were to decrease sufficiently and technical failures became much rarer, DNA analysis could replace existing methods of screening.

In 2016, the UK National Screening Committee issued a statement recommending implementing the recall method [7]. Since this recommendation was issued, the reflex method has been implemented in practice and published [2] with the method shown to be acceptable with the high screening performance predicted from previous modeling studies [5,6]. It is therefore reasonable for screening authorities throughout the world to consider the new evidence and for the National Screening Committee to review its position.

4. DNA analysis as the measurement of a screening marker

Prenatal DNA analysis is widely called a non-invasive prenatal test (NIPT). However, this is confusing because screening tests in pregnancy are by their nature noninvasive, so stating this implies a difference that does not exist, and the ‘T’ (for Test) creates an ambiguity as to whether the test is a screening or a diagnostic test. It is better to regard DNA analysis as the measurement of a screening marker, rather than a test in itself, in the same way that specific protein and steroid measurements are used in current screening practice. Regarding a DNA analysis as the measurement of a marker can endure at least until such time as DNA analysis may replace all other markers in prenatal screening for trisomy 21, 18, and 13. All screening tests for trisomy 21, 18, and 13 involve multiple markers, some of which can be measured at different times in pregnancy. From the perspective of screening therefore, DNA measurement is not conceptually different from other markers used together in a test. What is new is that the DNA analysis is performed in a reflex way on a high risk group of women following measurement of the Combined test markers, not on all women. It is straightforward when seeking consent to screening to indicate that the test offered is the reflex DNA screening test, with sufficient information to indicate what is involved and with the provision of an information leaflet.

5. Summary

The reflex DNA screening method can achieve a 95% detection rate for trisomies 21, 18, and 13, with a 0.02% false-positive rate, a high screening performance that exceeds that of the recall method and avoids causing needless distress to many women. The screening performance also exceeds that of screening tests used in current routine practice (i.e. the Combined test, the Quadruple test, and the Integrated test). Reflex DNA screening has been implemented in routine prenatal care with the high screening performance predicted from a previous modeling study. The method warrants consideration as the routine method of prenatal screening for trisomy 21, 18, and 13 until DNA analytical methods become simpler and cheaper with a minimal technical failure rate.

Declaration of interest

NJ Wald is director of Logical Medical Systems, which produces the software alpha, for the interpretation of prenatal screening tests. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.