Due in great part to the more than 20 years of groundbreaking work by Lo et al, the long-awaited noninvasive prenatal diagnosis of trisomy 21 is almost at hand, perhaps lacking only a few technical enhancements to reduce or eliminate a small false error rate to be widely applicable. Even more exciting is the potential to examine the whole fetal genome in search of disorders not caused by aneuploidy.

Decades in the making, the testing of cell-free fetal deoxyribonucleic acid (DNA) is disruptive technology that challenges the billion-dollar-a-year market for aneuploidy screening and diagnosis. However, the range of clinical applications remains modest by definition: DNA-based disease, and as such, cell-free plasma DNA testing may be vulnerable to non-DNA methods that can identify both the most common types of clinically relevant fetal aneuploidy plus have the potential to identify non-DNA based diseases. The present report by Bahado-Singh and colleagues applying metabolomics to the diagnosis of trisomy 21 is a first effort and may prove an example of such technology.

The investigative team sought unique metabolic markers in maternal sera from a cohort of 90 pregnancies at 12 weeks enriched with 30 proven fetuses with trisomy 21. The research team used nuclear magnetic resonance (NMR) spectroscopy. Perhaps because NMR spectroscopy is relatively insensitive, only 40 metabolites were identified in both euploid and trisomy 21 pregnancies, and only 3 of the 40 metabolites were predictably altered in pregnancies with fetal trisomy 21: 3-hydroxybutyrate, 3-hydroxyisovalerate, and 2-hydroxybutyrate. Each of the metabolites was significantly increased in affected pregnancies.

By searching the Human Metabolomics Database, the investigators suggested potential explanations for the impact of trisomy 21. First, they noted that 3-hydroxybutyrate is both an energy source and a substrate for the synthesis of phospholipids and sphingolipids required for brain growth and myelination. Myelination is delayed in the brain of trisomy 21 fetuses compared with controls and decreased myelination often associated with mental retardation. Thus, the identified increase in 3-hydroxybutyrate could reflect decreased utilization by the trisomy 21 fetus.

Second, 3-hydroxyisovalerate is a marker of biotinidase deficiency, and biotin deficiency can be associated with findings similar to those of trisomy 21 including hypotonia, learning disability, seizures, and brain atrophy. The effect of trisomy 21 on biotin biosynthesis does not appear to have been directly examined, except from the standpoint of nutritional intake.

Lastly, 2-hydroxybutyrate is typically increased under conditions of oxidative stress during which it is a byproduct of glutathione synthesis. One of the key enzymes in that pathway, cystathione β-synthase is overexpressed in the brains of trisomy 21 patients, and oxidative stress is considered a likely cause of neurotoxicity in trisomy 21. However, we note glutathione S-transferase activity and reduced glutathione itself are dramatically reduced in children with trisomy 21. It is possible that as the tissue of the affected child ages, it loses the ability to quench oxygen-free radicals.

The finding that a panel of metabolic markers is altered within the maternal serum metabolome of pregnancies with a trisomy 21 fetus is not, as the authors noted, surprising. After all, trisomy 21 is specifically associated with altered trophoblast activity as reflected in the levels of numerous placental proteins found in the maternal serum. However, trisomy 21 is also associated with widespread abnormalities of transcription, suggesting that the impact of trisomy 21 on physiology is far wider than the function of genes located on the number 21 chromosome. From that perspective, it is somewhat surprising the investigators did not identify, with the technology used, any unique metabolites associated with trisomy 21. It is possible a more sensitive method, such as gas chromatography mass spectroscopy, might yield a different result. Time will tell.

The variety of methods available for aneuploidy screening and its diagnosis has dramatically increased over the past several decades, so it is reasonable to consider their respective roles in health care as we enter the era of value medicine. Pricing becomes the driving variable when tests have comparable performance.

Presently, universal risk assessment by combining maternal age with multiple biochemical and ultrasound markers provides a patient-specific risk with very high sensitivity and positive predictive values and a low screen-positive rate (eg, a 95% detection rate with a 2.5% screen-positive rate using a risk cutoff of 1 in 100).

Such screening performance is possible when either biochemical testing is performed in all pregnancies or by using a contingency approach in which the first stage of screening includes maternal age, fetal nuchal translucency, and either tricuspid or ductus venosus flow, with biochemical testing reserved for those at intermediate risk (about 20% of the total). Those women identified as high risk are then offered definitive testing; the smaller the number of women classed as high risk, the smaller the number of invasive procedures required.

In an attempt to further reduce the number of invasive procedures, cell-free plasma DNA diagnosis of trisomies 21, 18, and 13 is increasingly accepted by health care plans for use in high-risk but not low-risk women because of the low but real risk of a false-positive result. In addition, up to 4.5% of specimens submitted for cell-free DNA study may prove unusable, further delaying diagnosis.

Should test sensitivity either be 100% or closely approach it (as it appears for trisomy 21), the need for invasive testing is essentially eliminated when the cell-free DNA test is negative, although the need for a positive test remains. Unfortunately, some companies have priced their tests to match or be slightly less than the present maximum total cost of testing an individual woman (cost of screening plus the cost of invasive fetal testing). Thus, widespread use of cell-free DNA testing will increase the overall cost of care from a societal perspective until the tests achieve the high-performance characteristics required. In that instance, societal costs would decline because all women could be tested noninvasively with an expectation of a secure diagnosis, eliminating the cost of screening.

Returning then to the present example and speculating on an enhanced metabolomic approach would reveal several disease specific panels of markers diagnostic for trisomies 13, 18, and 21 plus the major sex chromosome abnormalities such as XO, XXY, and XYY. Let us also assume that when the methodology is scaled up for commercial use, a test can be delivered for a quarter to a third of the patient cost of cell-free plasma DNA testing. Because these chromosome abnormalities constitute 80-90% of those identified by amniocentesis or chorionic villus sampling and about 90% of the remaining fetuses would be discovered to have an abnormality on detailed ultrasound scanning, the incremental diagnostic information provided by cell-free plasma DNA testing might not represent added value for the vast majority of women.

One could also foresee the development of a clinical paradigm in which a noninvasive diagnosis using metabolomics is the first testing modality offered to all pregnant women and cell-free plasma DNA reserved for those in whom either the metabolomic panels are not informative for diagnosis or are normal but associated with an abnormal ultrasound. The cost advantages of metabolomics would be multiplied if metabolomic markers for other pregnancy diseases such as preeclampsia and preterm birth could be identified and included with modest additional cost. This is not possible yet, but in 5 years perhaps it could be.

Other technologies have the potential to empower the same paradigm hypothesized for metabolomics. For example, adults with trisomy 21 express multiple transcription errors at the exon level. It may be that panels of cell-free plasma RNAs originating from the placenta of the fetus with trisomy 21 can be used for diagnosis. In addition, there are already multiple studies suggesting the maternal plasma transcriptome may be altered spontaneous preterm birth, preeclampsia, and/or growth restriction. Thus, it is reasonable to foresee the development of a panel of cell-free plasma RNA markers using high-throughput, low-cost PCR that at 10-12 weeks constituted a pregnancy wellness examination providing accurate predictions of which women carried an abnormal fetus or were destined to develop preeclampsia or experience spontaneous preterm birth. It seems unlikely new tests such as metabolomics for disorders of pregnancy that complement cell-free DNA will be long in coming.