Prenatal Testing or Screening?

Abstract

Over the past 50 years, the scope and extent of prenatal diagnosis and screening for genetic disorders have improved geometrically. There has been a pendulum like swing from testing to screening back and forth as new technologies emerge. The concurrent developments of cell free fetal DNA analysis of maternal blood has dramatically changed patient's choices towards screening. However, with the use of array comparative genomic hybridization of fetal DNA that requires diagnostic procedures (Chorionic villus sampling and amniocentesis), much more extensive diagnosis can be obtained. Until noninvasive methods can replicate what can be done with diagnostic procedures there still will be a “price to be paid” for opting for the non-invasive methods.

Introduction

“Mind the Gap”. Travelers to London will see this sign practically everywhere at the Underground (train stations). The sign literally means there is a gap between the end of the platform and the door to the train. If passengers do not pay attention when boarding, there could be an accident in which they fall down in the gap which potentially could have catastrophic consequences.

We argue that there is an analogy to the progress made over the past decade in noninvasive prenatal testing (NIPT) which originally was called diagnosis (NIPD), and more realistically is screening (NIPS) with concomitant advances in diagnostic testing from direct analysis of fetal tissue (eg, chorionic villi, amniotic fluid cells, fetal tissue biopsies).^1–13 However, there is overwhelming evidence of a juggernaut of increasing utilization of NIPT, and reliance upon it, with a concomitant decrease in the utilization of diagnostic procedures (DPs) since NIPT came on the market in many countries beginning about 2012.²

There have been many touted explanations for the changing pattern of utilization across many countries and societies. Over simplistically, proponents of NIPT have argued:

• (1) NIPT provides the same information as DP.
• (2) NIPT is safer, and DPs are dangerous.
• (3) DPs require skill to perform; NIPT does not.
• (4) NIPT focuses on disorders that patients are familiar with and to which they can relate.⁶

Detractors counter with:

• (1) DPs with “state of the art” diagnostic laboratory tests (eg, microarrays) provide far more information than NIPT. Whole exome sequencing (WES) will further push the envelope.
• (2) The incidence of these “ancillary” anomalies (eg, not common aneuploidies) is higher than “common” ones such as Trisomy 21 for the majority of patients.
• (3) At motivation against DPs is that it allows primary care providers the ability to not have to refer their patients to sub-specialists.
• (4) The “real” reason for NIPT is patient's desire for knowing the sex of their fetus/baby as early as possible.⁶

Indications for prenatal diagnosis and screening

Over the past 4 decades, a change in the pattern of clinically recognized indications for chromosomal prenatal diagnosis has occurred.¹ Overall, the most common indication for genetic counseling and prenatal diagnosis is still the risk for non-dysjunctional aneuploidy stemming from “advanced maternal age”. The risk threshold of which is still defined as being 35 years of age or older at delivery or having a numerical risk equivalent to that. Other “classic” indications to evaluate fetal karyotype include a previous, affected offspring, a balanced structural rearrangement of parental chromosomes, abnormal ultrasound, or abnormal screening tests.¹

In the 1980s and 1990s, increased use of biochemical serum screening and of ultrasonographic screening for fetal chromosome anomalies identified more young patients at risk.¹⁴ Previously, young pregnant women were considered to be at low risk for fetal aneuploidy. Positive screening results which increased their risk allowed them to consider having diagnostic prenatal testing. Various combinations of double, triple, or quadruple serum screening (multiplied by the a priori risk of their maternal age) selected a sub-group of patients among whom about 65% of chromosomally abnormal conceptions were identified. Using a risk cut off for fetal aneuploidy equal to that of age 35 years, some 5% of “young” pregnant patients would have a positive screening test. About 1 in 50 amniocenteses performed for this indication revealed a chromosomally abnormal conceptus.⁵ The positive predictive value (ie, the 1 in 50 = 2%) was not great, but it was notably higher than the chance that a 35 year old having an amniocentesis for only her age would have an abnormal result (about 1/140 = 0.7%).¹

Nuchal translucency (NT) screening/combined screening

The 1990s and 2000s featured the development and incorporation of NT screening combined with different biochemical analyte – particularly free β human chorionic gonadotropin (β hCG) and pregnancy associated plasma protein-A.¹⁴ Such “combined screening” had considerably improved statistical performance metrics than second trimester double, triple, or quadruple screening, but the quality control requirements for NT measurements were far higher and the actual performance far lower than standard laboratory techniques.^15,16 For example, the coefficient of variation of most lab tests was in the 5%–7% range, although for some lab tests, such as estriol, the coefficient varied from 5% in good labs to over 30% in poor ones which was the explanation of the huge variability of performance among triple and quad screening programs. Quality control programs for NT such as the British Fetal Medicine Foundation and the Nuchal Translucency Quality Review program of the American Society for Maternal Fetal Medicine measurements documented wide variability of performance, but that it could be improved with proper training and continuing evaluation and feedback. The reality was, however, that complete optimization could never be achieved with the diversity of quality and education in providers and training required for wide-spread screening.^15,16 An attempt to handicap (golf analogy) performance was developed by us,¹⁷ but it never caught on – particularly as NIPT was coming on the market which touted higher Down syndrome (DS) detection coupled with gender identification – albeit at much higher cost.

Noninvasive prenatal screening

The “holy grail” of prenatal screening for decades was the concept that fetal cells could be obtained from a maternal blood sample. Thus, the need and risk of an invasive DP for aneuploidy could be avoided.¹⁸ In 1997 Lo and Wainscoat first patented and then published a method for taking paternal DNA, amplifying it, and using it for the diagnosis of fetal gender.¹⁹ Over the years, several approaches have been attempted, including digital polymerized chain reaction of DNA and RNA and methylation differences. These approaches could not reliably identify and analyze free fetal DNA.²⁰ Beginning in 2011, the main approach has been the use of next-generation sequencing, also called massive parallel sequencing. DNA amplification is simultaneously performed millions of times with probes of approximately 36 base pairs long which provides enough specificity to accurately identify from which chromosome the excess fragments derive. Public health policy debates over the role of NIPS screening methods have been very stark. Public policy critics have argued that it is unreasonable to use methods whose costs are comparable to the total reimbursements in many jurisdictions from Medicaid for 9 months’ worth of care, including labor and delivery.²¹ The politically touchy assessment of financial costs, including the savings from pregnancies with serious problems that patients choose to terminate, are necessary to determine a true financial cost/benefit ratio.²² Ultimately, we believe that all patients should be offered diagnostic testing (discussed below). For those who choose screening, it is still debatable based upon improving detection and lower costs as to what is the best protocol.

Enhanced diagnostic testing and laboratory analysis

Diagnostic (invasive) procedures to obtain tissues for prenatal diagnosis of fetal disease are available throughout gestation from conception.¹ In the 1960s, amniocentesis permitted genetic diagnoses before ultrasound was even available. In the 1980s, chorionic villus sampling (CVS) moved up the timing of such diagnoses into the late 1^st trimester. Fetal tissue biopsies (such as muscle, skin, and liver) have had a limited but important place for specific diseases, and preimplantation genetic testing has developed into a significant component of in vitro fertilization. All have become considerably safer, more generally available, and more comprehensive over the years⁴ (Fig. 1).

Figure 1: Introduction of new diagnostic and laboratory procedures for prenatal diagnosis and screening. The unifying theme is increasing statistical performance metrics with increasing sensitivities and specificities for conditions interrogated. Amnio: amniocentesis; aCGH: Array comparative genomic hybridization; CVS: Chorionic villus sampling; DNA: Deoxyribonucleic acid; MICRO DEL: Microdeletions; NT: Nuchal translucency; RNA: Ribonucleic acid; WES: Whole exome sequencing; WGS: Whole genome sequencing.

The logarithmic advances have come in in the laboratory. Unbanded karyotyping of the 1960s improved with the development of banding, and more so with higher resolution banding. Molecular technologies, fluorescence in situ hybridization, quantitative fluorescence polymerase chain reaction, microarrays, and expanded Mendelian screening have all made their way into clinical practice vastly increasing the ability to determine the etiology of many anomalies that previously would have remained idiopathic. The next generation of diagnostics, WES, is now in research trials. Whole genome sequencing will be the next to be developed. The unifying theme for the past 50 years has been ever increasing diagnostic capabilities that not only increase our understanding of the current situation but allow for better counseling towards recurrence risks and related issues.^23–28

Testing vs. screening

Since the early development of prenatal diagnosis in the late 1960s, there has been a pendulum constantly swinging back and forth between the primacy of screening vs. testing. DPs have progressively become safer as centers of excellence have developed with ever more sophisticated ultrasound machines and clinicians to visualize anatomy and experienced operators to perform them. At the same time, screening procedures have also improved with increasing rounds of improved sensitivity and specificity as generations of technology have progressed.^4,6

Over the past several decade, cell free fetal DNA for NIPS has emerged as the most popular method for screening for common aneuploidies and sex identification. Millions of cases have been performed, as the utilization of diagnostic testing (mostly amniocentesis) and other methods of screening such as the combined screen have plummeted.

In our own program for the past 35 years, about 80% of patients following genetic counseling have proceeded with a DP. That has not changed. What has changed is the proportion of patients even coming for consultations has dramatically decreased.^4,6

NIPS and microarrays (array comparative genomic hybridization (aCGH)) are both disruptive technologies that have had major impact upon the practice of obstetrics and gynecology and reproductive genetics. However, unlike most innovative changes in medicine, the introduction of NIPS has been primarily industry driven. Large scale clinical use studies were done only after wide-spread clinical introduction and extensive marketing such that the costs of the studies were mostly paid for by patient revenues.²⁹ Conversely, aCGH followed the more traditional paradigm of academic studies, grant funding, and large-scale multicenter investigations before introduction; it has not had the mass sales force efforts of NIPS. NIPS also allows the primary provider to be able to “work around” the sub-specialists and maintain more autonomy and control of their patients. However, we believe this practice generally does a disservice to patients who would desire a complete diagnostic evaluation, but who often are not told that there is more available than what they are being offered.^4,6,29

Our data are consistent with a general concept we have presented for several years that the focus of prenatal diagnosis needs to move beyond DS to consider all sources of neurologic and structural impairment.^4,6 With the advent of screening for pre-eclampsia, further progress is also being added to maternal health disorder screening.³⁰ For the majority of pregnant women, the incidence of abnormal CNVs is actually considerably greater than the standard aneuploidies.

For younger women, the detection of abnormal CNVs can be 10 times the expected yield of NIPS.^6,31 Eventually, with deeper next generation sequencing and WES, non-invasive methods may approximate the diagnostic capabilities of DPs and ACGH.^22,29,31 Until that time, however, literally tens of thousands of CNVs are not being detected because of reliance upon the screening practices of NIPS today.^29,32 The broad umbrella of public policy, cost/benefit analysis, and maximizing patient autonomy, includes a dispassionate analysis of the pros and cons of all options. To us, moving towards the direction of much higher diagnostic capabilities at the risk of complications including pregnancy loss with a ratio of 8.5/1 should be considered very compelling.

There will never be a uniform acceptance of any stance on this subject. For 50 years the concept of accepting a procedural risk for diagnostic capabilities has been at the center of genetic counseling and prenatal diagnosis.³³ The issues for NIPS vs. ACGH are in parallel with many that have come previously – just the names of the conditions which can be discovered, what resources are used, and risks taken to find them have evolved and will always continue to do so.

We have estimated that the cost of care of a child with an abnormal CNVs might be about US $500 000.^6,22 Unfortunately, hard data on the costs of numerous serious genetic disorders are not generally available.²² However, this number is half the commonly accepted $1 million societal cost for a baby with DS. If we compare these numbers against estimated costs for several other medical conditions, it becomes apparent that abnormal CNVs, for which we are doing very little as a public health measure, are costing the medical “system” far more than many issues for which we devote tremendous efforts to detect and parents often choose to prevent.^22,32–40 In toto, abnormal CNVs represent about 3 times the cost of DS and cerebral palsy.⁶ The incidences of auto fatalities, guns shot deaths, and human immunodeficiency virus are comparable to abnormal CNVs, and the cost of human immunodeficiency virus are roughly comparable (Table 1). Far more public health efforts are being directed at these other situations such that the tremendous expense for NIPS for DS represents a very inefficient use of health care expenditures.

Table 1: USA societal comparisons, estimates for United States of annual incidence, yearly costs, and anticipated lifetime costs of differing conditions.

We believe that it is time to take a step back and view prenatal screening and diagnosis from the wider perspective of its impact upon society in more than just the strict statistical performance metrics. However, it is impossible to address an epidemic until one recognizes that it exists.

Because the yield of diagnosed abnormalities is essentially twice that of karyotypes in dysmorphic children, chromosomal microarrays (ACGH) have for all practical purposes in modern pediatrics replaced karyotypes.⁴¹ The United States National Institute for Child Health and Human Development trial and many others have shown that the detection of CNV's of well documented pathology is about 1.5% in fetuses with normal ultrasounds and karyotype.^10,11 For over 40 years, high risk has been defined as age 35, so, in fact, now “everyone's risk” is actually over 35. This is why we offer DPs to essentially all pregnant women regardless of their age.⁴

For NIPS, never before have we seen one technology substantially replace another (diagnostic tests) when it is both simultaneously less efficacious and ultimately more expensive.⁴ The shift from diagnostics to screening is not new, however. Over two decades, ultrasound reliance has also decreased diagnostic testing, but has resulted in more anomalies being missed.⁴² We have voiced serious concerns about quality control of NT and have proposed methods to improve it. Overall, laboratory tests have higher performance metrics than ultrasound, but nothing will ever be perfect.^3,43

A prominent argument for NIPS is that it carries no procedural risks. While true, in experienced hands, the risks of both CVS and amniocentesis are very low (1/800–1/1 000); CVS is just as safe (or safer) than amniocentesis; the reality is much lower than NIPT advocates often claim.^44,45 We counsel all patients that in essentially the middle 99%, it does not actually matter if they have screening or testing; everything will be “fine”. The issue is, if they are going to be wrong, which way would they rather be wrong? Would they rather take a small risk of having a baby with a serious problem, or a small risk of a complication because they wanted to know “that”. Not surprisingly, there is a strong liberal/conservative divide on how patients act.⁴⁶

We have modeled the choices, for every 1 million CVS/amnios that there would be about 2 000 losses (or less). However, about 17 000 serious problems could be identified (ratio 8.5/1). We believe that most couples would find this ratio to be acceptable – provided they are actually presented with these data.^6,22 In our experiences, most patients never hear such a discussion.

NIPS has high performance for DS, but the sensitivity is much less for other conditions.⁶ Although originally proposed for only the high-risk population, its use quickly migrated to being used mostly on low risk, routine patients. By definition, all screening tests have poorer performance for lower incidence conditions and are therefore less cost beneficial. Our own work and others have suggested that the cost of finding additional DS cases over combined screening may be about US $3 million per case.⁶ In the United States, the Aetna Insurance company began coverage for NIPS for “routine” pregnancies in 2015, but in 2017 reversed their decision. We believe this was a reasonable decision. The cost/benefits of routine NIPS vs. combined screening (β hCG, pregnancy associated plasma protein-A, and NT) will continually evolve as the costs of NIPS and diagnostic laboratory tests inevitably decrease with new technology and higher volumes.

The spectrum of ACGH analysis has improved considerably. About 1% of all children develop neurological handicaps/developmental delays (ND), and over 1% develop autism.⁴⁷ ACGH studies have suggested as much as 40% of ND and 20% of autism can be identified.^4,6,29 The 2012 National Institute for Child Health and Human Development data showed, at the most conservative, 0.5% of patients with no other findings had a well-documented, pathological CNVs.⁴⁸ One point seven percent had a CNV that was either well documented or likely pathological. We believe 1.5% is a reasonable, conservative number. Modeling variable detection, costs, and patient choices, allowed us to develop both high and low-cost estimates under differing scenarios.

Hypothetically, at one extreme, if all patients underwent CVS and ACGH and all with anomalies chose to terminate, the health care system (as a whole) could save about $7 billion per year per 1 million patients. In the United States with approximately 4 million births per year, theoretically, the savings could be $28 billion. Even reducing costs by 50% would still suggest $14 billion in savings.^4,6 One can multiply the number of births for any country multiply by the anticipated costs and patient's decisions to create local models.

In differing countries with differing medical coverage models, the costs of care, non-medical societal costs, ethical belief structures, norms, and patient's options vary considerably. No short piece such as this can model those. However, this analysis can serve as the first step towards the creation of such on a country by country assessment.

Funding

None.

Conflicts of Interest

None.