Recent developments in early pregnancy screening: are we getting closer to the Holy Grail?

Traditionally, expectant mothers have had their pregnancies predictively categorised as low risk or high risk, depending on the perceived probability of an adverse maternal or neonatal outcome. Although appealing in its dichotomous simplicity, such a categorisation does not reflect the spectrum of risk that exists for all pregnant women, nor does it acknowledge significant limitations that have, until recently, precluded the accurate prediction of obstetric risk, particularly among women who have never previously given birth. For example, an algorithm for the prediction of pre-eclampsia among women in their first pregnancy, based on maternal risk factors alone, yields only a 37% detection rate for a 10% false-positive rate.1 However, just as screening for fetal aneuploidy has evolved from using maternal age alone to using non-invasive prenatal testing of cell-free fetal DNA in maternal serum, so too has first-trimester testing been refined to permit, with significantly improved efficacy, the early prediction of other important obstetric concerns, such as fetal growth restriction and pre-eclampsia. With these developments, we are on the cusp of a new era in antenatal care, in which common and important pregnancy outcomes can be more reliably predicted from an early gestation, thereby promising improved triaging of patients, the institution of targeted surveillance and prophylactic therapies, and recruitment of a truly high-risk population to clinical research trials.

The current gold standard for aneuploidy screening is the first-trimester combined test, which assesses two proteins in the maternal serum — pregnancy-associated plasma protein A (PAPP-A) and the β subunit of human chorionic gonadotropin — and the fetal nuchal translucency (sonographic measurement of the fluid space behind the fetal neck). This test has a detection rate for trisomy 21 (Down syndrome) of 85%–90% for a 5% false-positive rate, although inclusion of multiple sonographic markers (eg, presence of the fetal nasal bone) in the testing algorithm reduces false-positive results by half, and increases detection to 93%–96%.2 Since its introduction early last decade, the uptake of this test has increased to over 50% of pregnancies in Australia, although previously identified substantial sociodemographic variations in its availability3 are likely to have persisted. Despite its improved performance, only a minority of pregnancies deemed high risk by this test will result in an aneuploid fetus, and around 5%–10% of Down syndrome pregnancies will remain undetected.

Late last year, aneuploidy screening in Australia was revolutionised by the clinical availability of cell-free fetal DNA tests, the basis of non-invasive prenatal testing (NIPT). These tests extract fetal genetic material from maternal blood, allowing for direct assessment of fetal chromosomes. From October 2013, five such tests have been commercially available in Australia, all of which are performed by laboratories abroad, but facilitated and promoted by local pathology and imaging practices. The tests vary in their costs (initially up to $1450), timing (from as early as 9 weeks up to the end of pregnancy), technology employed (whole-genome massively parallel sequencing, targeted approaches or single nucleotide polymorphism array), chromosomes tested (some include sex chromosomes in addition to chromosomes 21, 18 and 13), turnaround times (10 days to 2 weeks), and reported sensitivities and specificities for each aneuploidy assessed. For trisomy 21, each test is reported to have a greater than 99% sensitivity and specificity, thereby approaching the performance of a diagnostic test (but not replacing one).

The appeal of a non-invasive, highly sensitive and specific screening test in early gestation for aneuploidy is easily appreciated. Nevertheless, these tests have limitations that must be considered in determining their optimal use in clinical care. The populations in which the performance of these tests were validated included ones with very high rates of aneuploid pregnancies (up to one in eight), and relatively few studies have been performed that assess the performance of the test in populations with low-risk pregnancies or in multiple pregnancies.4 For these reasons, professional organisations (including the Royal Australian and New Zealand College of Obstetricians and Gynaecologists) have thus far advised against routinely offering NIPT to women at low risk of having a child with aneuploidy until further data are available. Additionally, the focused nature of this testing does not allow chromosomal anomalies other than those specifically analysed to be identified. This is in contrast to current first-trimester aneuploidy screening regimens, which also detect around 75% of all chromosomal abnormalities, and, by virtue of the 12-week ultrasound, permit early identification of major structural anomalies (such as anencephaly). On the other hand, expanding non-invasive fetal DNA testing to encompass the entire karyotype, potentially at the molecular level,5 is likely to generate many results of uncertain clinical significance, thereby increasing parental anxiety rather than providing reassurance. Including sex chromosomes in the testing panel, as is done by four of the five Australian providers, will also allow for de facto fetal sex selection if this information is routinely made available to parents.

How NIPT might best be integrated into current aneuploidy screening strategies remains to be determined; national guidelines are still in development. Various models have been proposed, including reserving NIPT for pregnancies deemed high risk by conventional screening, offering it to all in addition to a 12-week scan, and applying a contingent approach in which NIPT is performed in pregnancies deemed to be at intermediate risk on conventional screening.6 Detailed cost–benefit analyses will be required before the optimal contribution of NIPT can be established.

It has long been recognised that abnormal levels of maternal serum analytes assessed in aneuploidy screening have a statistically significant association with adverse obstetric outcomes in euploid pregnancies. For example, a PAPP-A level of less than 0.42 multiples of the median level in the first trimester (ie, less than the 5th centile) confers an adjusted odds ratio of 2.81 for low birthweight (less than the 5th centile).7 However, its sensitivity is only 12.23%, with a positive predictive value of 9.5%. No single biomarker has been identified that has enough predictive value for complications later in pregnancy to be of clinical utility. Adequate test performance is only achieved by assessing multiple parameters in combination, including baseline maternal characteristics.

The development of such tests is best illustrated in predicting pre-eclampsia, the commonest serious medical disorder of pregnancy. A seminal article published in 2009 described a test incorporating maternal factors, mean arterial pressure, uterine artery Doppler pulsatility index, placental growth factor and PAPP-A that detected 93% of early-onset pre-eclampsia for a false-positive rate of 5%.8 A similar strategy has recently been validated in Australia,9 in which 91.7% of early-onset pre-eclampsia was detected for a false-positive rate of 10%. This testing regimen has now been introduced into clinical practice in some centres. In contrast, other multiparametric regimens have not demonstrated predictive utility, with a recent Australian study showing that soluble fms-like tyrosine kinase-1 (sFlt-1) and placental growth factor (PlGF) do not add predictive value to a model based on maternal factors alone.10 Aspirin has been shown to reduce the risk of pre-eclampsia, especially if therapy is started early in pregnancy; its utility in lowering the risk of pre-eclampsia in those predicted by these tests to be at high risk of this condition is the subject of ongoing randomised trials (http://controlled-trials.com/ISRCTN13633058).

Research into the early prediction of other causes of adverse pregnancy outcome is ongoing, and will likely lead to clinically useful strategies for these as well.11 For example:

• combining measurement of the cervical length at 11–13 weeks’ gestation with maternal factors can predict over half of all spontaneous preterm births earlier than 34 weeks’ gestation;12

• assessing maternal serum analytes such as adiponectin and sex hormone-binding globulin in the first trimester may identify up to 75% of those who will go on to develop gestational diabetes;13 and

• almost three-quarters of fetal growth restriction necessitating delivery before 37 weeks’ gestation can be identified using a screening algorithm incorporating maternal factors and biophysical and biochemical markers.

Given the strong predictive value of past obstetric history, the group who stand to benefit most from these screening tests are women in their first pregnancies, who do not have the benefit of prior obstetric outcomes from which future risks can be predicted.

These advances in early pregnancy screening promise a major shift in the focus of early antenatal care. To promote acceptance of these tests by patients, they would ideally be integrated into a unified testing strategy, with clearly articulated positive and negative predictive values for each outcome. It is essential, however, that the established principles of screening15 be applied to these advances, both to NIPT and to screening for adverse pregnancy outcomes. Robust prospective analysis of these developments in testing in the Australian context is required, with particular attention to cost-effectiveness, and the extent to which these testing strategies improve pregnancy outcomes (eg, by avoiding invasive testing, initiating prophylactic therapy, or instituting closer surveillance). The cost of these tests, if borne privately, will run the risk of exacerbating health inequity in Australia, and commercial imperatives may well lead to them being introduced before there is widespread agreement on a framework for their use, and to them being performed without appropriate pre-test and post-test counselling. Detailed economic modelling will be required to justify allocating scarce public resources to such tests, and attention must be paid to ensuring that women living in regional and remote areas — for whom early prediction of an adverse pregnancy outcome may be of particular benefit — are not disadvantaged.

Given their central role in antenatal care, especially in the public system, it is vital that general practitioners be well apprised of the indications for and performance of these testing strategies, to facilitate optimal pre-test counselling, and post-test referral for additional care when indicated. With this approach, we may finally be in a position to grasp the holy grail of antenatal care: the highly accurate early prediction of a wide range of pregnancy complications, and thus a much improved distinction between high-risk and low-risk pregnancies.