Traditional prenatal screening, by the use of a combination of serum and ultrasound markers, has been the standard approach to prenatal screening for many years. Incremental improvements have led to the current model, which includes first- and second-trimester serum analytes combined with measurement of first-trimester nuchal translucency. This screening test provides numeric risk assessment for trisomies 21 and 18, as well as for neural tube defects. Approximately 5% of pregnancies are flagged as screen positive, and it has been recognized that “false-positive” screening results indicate risk for a host of other fetal and obstetric abnormalities or adverse outcomes. Cell-free DNA (cfDNA) screening, in contrast, is a far more precise test for a limited number of aneuploidies. Although clearly a better test for trisomy 21, this approach does not provide any information regarding risk for rare aneuploidies, obstetric complications, open fetal defects, or other fetal structural abnormalities. Furthermore, although the false-positive rate is very low, the exclusion of a large number of patients as the result of failed tests or the presence of other fetal abnormalities introduces bias in evaluating test performance. We whole-heartedly agree with Palomaki et al that “It is critical that groups reporting screening performance of any prenatal test carefully consider the impact of bias on the estimated detection rates, so that data used for counseling and policy-making will be accurate.”

In a recent publication, our group compared the detection of chromosomal abnormalities based on traditional sequential screening, as conducted in the California Prenatal Screening Program, with predicted performance outcomes had cfDNA screening been instead used as a primary screening test in this same cohort. We had recognized that sequential screening identifies many women as having “false-positive” results for trisomy 18 and/or 21 who are then found to have a fetus affected by a different disorder. Arguably, these tests are not “false” positive. We also wished to consider the implications of the careful curation of the included populations that has been characteristic of studies of cfDNA screening, in which a large number of women are excluded from analysis because of failed tests; these failed tests are somewhat comparable to “false-positive” serum results because they indicate an increased risk of adverse outcomes. The goal of our study was therefore to broaden the discussion regarding the relative benefits and limitations of each approach.

Traditional sequential screening and cfDNA screening produce different types of information, although both are screening tests for fetal abnormalities. Sequential screening is a very broad test that provides information about many fetal conditions, and therefore has better detection if the denominator is “all chromosome abnormalities” or “all birth defects.” It therefore may be a better test for women who are at low risk for aneuploidy but at average risk for a wide range of conditions. cfDNA, on the other hand, is a very precise test for 3 aneuploidies. It has a better detection rate if the denominator used is “all cases of trisomy 21” and may be a more appropriate test for women in whom this single condition is the primary concern. These women, however, need to be informed of the limited number of conditions this test detects and also need to be offered additional screening for other birth defects. The availability of 2 screening tests for trisomy 21, each with different benefits and harms, has led to vigorous debate as to which of these tests, separately or in combination, in which patients, should be used.

This debate is reminiscent of the discussions that occurred 20 years ago regarding the appropriate role of rapid aneuploidy testing with fluorescence in situ hybridization for trisomies 13, 18, 21, and the sex chromosomes, and whether full karyotyping was needed. As pointed out by Caine et al in a 2005 issue of The Lancet , “Replacement of full karyotyping with rapid testing for trisomies 13, 18, and 21 after a positive screen for Down’s syndrome will result in substantial numbers of liveborn children with hitherto preventable mental or physical handicaps, and represents a substantial change in the outcome quality of prenatal testing offered to couples.” This finding is directly applicable to our previous report of the tradeoff in use of cfDNA versus diagnostic testing for women who are screen positive based on sequential screening. Interestingly, these discussions were much more robust in European health systems, in which cost is a greater consideration and prenatal testing options are largely determined by government health policy. In the United States, the option of only providing the far more limited chromosomal information available through fluorescence in situ hybridization (information that is the same as that available now through cfDNA screening) was never considered seriously because it was assumed that women in the United States would prefer the more comprehensive information available through a full karyotype.

The importance of incorporating patient preferences into clinical decisions has been emphasized increasingly in recent years. Patient preferences are particularly germane to decisions regarding prenatal testing, as the values patients ascribe to the outcomes of these decisions are highly personal and value laden, and vary substantially from individual to individual. For many women, not having any form of testing is most aligned with their values. For some women who do desire fetal information, comprehensive detection of a wide range of conditions is most important, while for others, avoiding a miscarriage is the primary consideration. To enable women to make informed decisions, they need to understand the very real tradeoffs of the various testing options.

Clearly, prenatal screening has both benefits and limitations when compared to invasive diagnostic testing, and recent reassessment of miscarriage risks associated with prenatal diagnosis, which appear to be substantially lower than what has been traditionally quoted, impacts those tradeoffs. Furthermore, sequential screening has different tradeoffs compared with cfDNA screening. There is no question, however, that more comprehensive—and more accurate, less biased—pre- and posttest counseling is needed regarding prenatal screening and diagnostic testing options. Our goal in comparing the results of the California program with predicted results of cfDNA screening was to point out the biases and limitations in what is known about the impact of population implementation of primary cfDNA screening, and to spark discussion of how to assess and address these issues. Ideally, a comparative effectiveness study should be completed with inclusion of all enrolled patients and comprehensive assessment of outcomes in all patients. Such a study, however, would be very expensive and difficult to undertake in the current environment. In the absence of such a study, we welcome further discussion of patient education, preferences, and values, as well as of who should decide what conditions and test characteristics are most important to consider, and what the appropriate role of risk triage might be.

At a minimum, all providers of prenatal screening need to make an active effort to understand and, ideally, to participate in efforts to generate evidence on the outcomes of screening. In addition, practitioners should provide the best available information on ascertainment rates, detection rates, and false-positive rates to their patients, to help them make informed prenatal testing decisions. Assuming that “no news is good news” and that false-negative screening tests will always result in a report back to the laboratory is a wholly inadequate ascertainment approach. Editors of medical journals also have a role in assuring that accurate, unbiased, high-quality data are provided, and should require a minimum rate of follow-up and hold authors of cfDNA studies to a high standard comparable to those required in other areas of clinical research.