New assays for detecting chromosomal abnormalities with prenatal testing are being adopted clinically without enough data to allow appropriate interpretation (NEJM 2013; 369: 499-501).

The assays are cell-free DNA (cfDNA) tests which can be performed as early as 9 weeks and require only a maternal blood sample.  The potential advantages also include avoiding invasive testing, like amniocentesis or chorionic villus sampling, which can result infrequently in miscarriage.  The tests are reported to have high sensitivity and high specificity for detecting abnormalities.

The problem is that these assays have not been tested adequately in more general populations and the manufacturers have not provided positive predictive values (PPV) for these tests.

Here’s why the statistics matter:

If you perform testing with samples that have a chromosomal abnormality prevalence of 1 in 8 with sensitivity/specificity of 99.9% and 99.7% respectively, then you derive a PPV of 98% and negative predictive value of 99.99% which are excellent.  However, when one uses the exact same sensitivity/specificity, then when the prevalence drops to 1 in 200, the PPV drops to 62.6%.

As a result, professional organizations like the American Congress of Obstetricians and Gynecologists have recommended cfDNA only for “high-risk” pregnancies and to confirm positive results through invasive testing.  These recommendations negate much of the potential benefits of early screening and avoidance of invasive testing.  In addition, the costs of these tests range from $795 to more than $2000 (though some discounts or “introductory pricing” are available).

Ultimately, cfDNA may prove to be extremely useful.  Until more data are available in the general population, this remains unproven.