Translational research advances a new era of prenatal diagnosis and newborn screening

1Evolution of prenatal and neonatal genetic diagnostic methods and screening

Prenatal and perinatal medicine advanced dramatically, though incrementally, during the second half of the twentieth century because of combination of research discoveries and technological developments applicable to the fetus and newborn. The important discoveries and their translation into patient care depended on interdisciplinary science and teamwork among health care providers. Perhaps no other area of medical practice has been so immediately and profoundly influenced by the genetics revolution as represented by the completion of the Human Genome Project in 2003, as cellular and molecular biology techniques developed in research laboratories to sequence the human genome became widely available in clinical practice. Thus, recent years have witnessed an extraordinary improvement in our ability to diagnose genetic conditions either prenatally or in the early neonatal period; almost all of these depend upon collaboration with clinical laboratories.

Prenatal genetics emerged in the 1960’s and 1970’s with amniocentesis used to assess fetal conditions. During the 1980’s, chorionic villus sampling technology with cell culture emerged as an important step forward. This was followed during the next decade by ultrasound-based assessment of fetal morphology, as originally applied by radiologists and subsequently by maternal-fetal specialists and reproductive geneticists. As novel genomic and proteomic technologies were applied, the precision of diagnostic technologies has improved dramatically and non-invasive screening methods now predominate in prenatal genetics practice. The interval from laboratory discovery to bedside application is no longer measured in years, as previously has been the case in medicine, but in months or even weeks. In recent years, the most significant advances have come about by applying both biochemical and molecular technologies. The burgeoning field of prenatal diagnosis has fostered a host of specialty journals, most notably Prenatal Diagnosis, books, and internet sites. Thus, driven by a combination of excellent diagnostic technologies and patient demand, and also augmented by growing proportion of first pregnancies in women beyond thirty years of age, a new era in prenatal diagnosis of genetic disease has been well established worldwide [11].

In parallel, the diagnosis of genetic disorders in newborn infants has improved for a wide variety of inborn errors of metabolism due to single gene defects. The advent of newborn screening is historically interesting because it depended primarily on development of a neonatal blood sampling technique rather than analytical biochemistry advances [12]. Specifically, to promote early diagnosis of phenylketonuria (PKU) in newborn populations, Robert Guthrie devised a technique for routine procurement of blood using a “heel stick” to obtain about 0.4 ml of blood on filter paper. During the early 1960’s, the potential opportunity to prevent mental retardation through early diagnosis aroused great interest, and Guthrie was motivated to focus his attention on PKU using blood phenylalanine levels determined microbiologically to ensure early diagnosis through population-based screening of all newborns. Driven by parent advocacy groups, mandatory statewide screening for PKU was first instituted in Massachusetts and Oregon during 1963, and two years later 28 more states had statutory requirements for PKU screening of all newborns. Next, galactosemia screening was implemented on the same dried blood specimens using microbiologic determinations and in the early 1970’s congenital hypothyroidism followed. Many other genetic disorders classified as inborn errors of metabolism were added in the 1980’s and 1990’s. With the recent advent of tandem mass spectrometry, more than 50 inborn errors of metabolism can be readily diagnosed using dried blood specimens. The analytical methods extended in the 1990’s to DNA analyses for carrier screening and diagnosis of cystic fibrosis (CF). Soon after nationwide newborn screening for CF began in the USA, another DNA-based test was introduced for Severe Combined Immunodeficiency (SCID). Because congenital metabolic and endocrine disorders have a latent period from onset to irreversible pathology, the opportunity exists for intervention with preventive therapies, and these have also been developing.

2Diagnostic strategies

Despite the incorporation of new and novel laboratory techniques, conclusively establishing the diagnosis of disease still typically depends on history, physical examination, and laboratory studies confirming preexisting suspicion. In the case of the fetus and asymptomatic neonate, laboratory tests are likely to increase in importance compared to other factors applicable to the newborn or infant. Screening for early detection of disease is especially valuable in prenatal and neonatal practice. According to Hennekens and Buring [13], “screening refers to the application of a test to people who are as yet asymptomatic for the purpose of classifying them with the respect to their likelihood of having a particular disease…the screening procedure itself does not diagnose illness…those who test positive are sent on for further evaluation by a subsequent diagnostic test or procedure to determine whether they do, in fact, have the disease.” As modern laboratory methods improve, however, some screening tests can actually become diagnostic tests as in the case of CF when DNA analyses on neonatal blood specimens reveal two F508del mutations of the cystic fibrosis transmembrane conductance regulator (CFTR) gene [14].

There are important differences between prenatal diagnostic strategies practiced by obstetricians and the early diagnosis of genetic disorders in newborns through population-based screening. In general, the former has relied on the traditional medical model of intervention in individuals (pregnant women), although the interventions are usually triggered by risk factors rather than illnesses. Newer approaches to population-based screening applied to pregnant women are represented by the use of massive parallel shotgun sequencing of cell-free fetal nucleic acids obtained from maternal blood to noninvasively screen pregnant women for fetal trisomy 21 and will likely be applicable to other fetal chromosomal abnormalities and Mendelian disorders [15, 16].

Important screening test characteristics include two measures of validity, namely sensitivity and specificity, and also measures of probability. Sensitivity and specificity are inherent characteristics of screening tests with a defined cut-off value and are not affected by disease prevalence, whereas the predictive values (probability measures) of the tests are influenced by prevalence. The most important consideration in prenatal and neonatal genetics practice is that screening tests achieve close to 100% sensitivity, i.e. avoidance of false negative results. Unfortunately, it is impossible to achieve 100% sensitivity over a long duration because even when there are no biologic false negatives, the possibility always exists for laboratory or human error. Therefore, health care providers should always assume that no screening test is perfect, just as no laboratory test is infallible. Test specificity, or the true negative rate, indicates probability of the person without the disease will have a negative test result. As with sensitivity, specificity should ideally approach 100% but this is difficult to achieve while optimizing sensitivity. Finally, the positive predictive value of (PPV) should be regarded as extremely important because it provides a measure of the post-test probability or likelihood of disease following a positive screening test. PPV indicates the proportion of persons with positive test results who have the disease.

3Methods of prenatal diagnosis

A brief overview of the well-known conventional techniques is provided below and is followed by comments about newer preimplantation techniques.

A. Sampling Techniques

1. Chorionic villus sampling

Since amniocentesis (discussed below) is most commonly performed in the mid-second trimester (15 to 16 weeks), fetal diagnosis is not usually established prior to 16 weeks gestation. A technique that could be performed during the first trimester is most desirable to reduce the psychological stress of awaiting results until mid-pregnancy and to allow a safer method of pregnancy termination. Chorionic villus sampling (CVS) is such a technique. Following fertilization, the zygote differentiates first into the blastocyst, which contains an inner cell mass that develops into the fetus and an outer trophoblastic layer that develops into nonfetal structures such as amnion, chorion, and placenta. The genetic complement of the outer cell mass nearly always reflects the genetic constitution of the inner cell mass (i.e., the fetus) because both are derived from the same zygote. It follows that cytogenetic, DNA, or biochemical analysis on trophoblastic cells should provide information comparable with that obtained from cultured amniotic fluid cells. The one major exception is that assays requiring amniotic fluid, specifically alpha fetoprotein (AFP), require amniocentesis. The procedure is usually performed between 10 and 14 weeks’ gestation. Under ultrasound guidance, chorionic villi are aspirated mostly by passing a catheter transcervically into the placenta or passing a needle through the maternal abdomen and uterus into the placenta [17, 18]. Recent assessments suggest that the risk for CVS-related loss among skilled and experienced clinicians is not different than for amniocentesis, with an increased risk for loss at approximately 1 in 1,000 [19] although earlier studies on the safety of CVS suggested an approximate 1% increased loss of pregnancy [20]. The risk for fetal limb defects following CVS was initially believed to be increased but is now considered to be no greater among women undergoing CVS than those not undergoing invasive prenatal diagnosis [20, 21].

2. Amniocentesis

Biological specimens obtainable include amniotic fluid and cells. The procedure is performed from 14 weeks’ gestation onward. The technique involves passing a needle transabdominally into the amniotic cavity under continuous ultrasound monitoring.

Analyses are the same as for CVS. AFP can also be evaluated in amniotic fluid, as can additional metabolites (e.g., proteins, hormones). The procedure was previously associated with an increased risk of loss of 0.5% above the baseline risk of spontaneous loss or stillbirth, depending on the expertise of the physician performing the procedure [3, 22]. However, more recent assessments of the safety of amniocentesis now indicate a far lower risk of loss, with a large, prospective study indicating no significant increased risk of pregnancy loss among women undergoing second-trimester amniocentesis [23].

3. Percutaneous umbilical blood sampling

Access to the fetal circulation was initially accomplished by fetoscopy, a method of directly visualizing the fetus, umbilical cord, and chorionic surface of the placenta, using endoscopic instruments [24]. Fetoscopy for this purpose has now been replaced by ultrasound-directed percutaneous umbilical blood sampling (PUBS), also termed cordocentesis or funipuncture. Fetal blood chromosome analysis has been used to help clarify purported chromosome mosaicism detected in cultured amniotic fluid cells [25] or chorionic villi. More recently, fluorescent in situ hybridization (FISH) with chromosome-specific DNA probes has also been used for rapid prenatal diagnosis of aneuploidy using nucleated fetal blood cells from umbilical cord blood, as well as chorionic villus and amniotic fluid cells [26]. Fetal blood sampling appears to be a relatively safe procedure when performed by experienced physicians, albeit carrying more risk than CVS or amniocentesis. Collaborative data from 14 North American centers, sampling 1600 patients at varying gestational ages and for a variety of indications, revealed an uncorrected fetal loss rate of 1.6% [27].

B. Preimplantation Genetic Diagnosis

Preimplantation genetic diagnosis (PGD) is a prenatal diagnostic approach allowing for the detection of embryonic abnormalities that can prevent the implantation of an abnormal or affected conceptus [28, 29]. Many couples choose PGD rather than traditional prenatal genetic diagnostic approaches for assessing Mendelian disorders, aneuploidy, or chromosomal imbalance; however, such couples must undergo in vitro fertilization (IVF) in order to avail themselves of PGD. Unlike early amniocentesis, chorionic villus sampling, and other invasive techniques, PGD is not just an earlier option for prenatal diagnosis. It allows genetic diagnosis prior to the establishment of an in utero pregnancy, a preconceptional approach that has unique advantages. Moreover, potential application of PGD extends beyond traditional prenatal genetic diagnosis and improved pregnancy rates in assisted reproductive technology (ART).

A successful preimplantation genetics diagnostic program requires 1) high quality ART, 2) micromanipulation skills sufficient to obtain a specimen (polar body or blastomere) for analysis, and 3) molecular technology more sophisticated than that required for traditional prenatal diagnosis. Preimplantation genetic diagnosis requires access to gametes (oocyte) or embryos before 6 days post conception, the time when implantation occurs. There are three potential approaches: 1) blastomere biopsy or aspiration of 1 to 2 blastomeres from the 6 to 8 cell embryos (2 to 3 days), 2) polar body biopsy, and 3) trophectoderm biopsy from the 5- to 6-day blastocyst.

Blastomere (6 – 8 cells) biopsy

Biopsy at this stage was the first technique developed. The approach is to aspirate 1-2 of the 6–8 cells contained within the zona pellucida. This can be accomplished by mechanical (razor) or chemical (pronase, EDTA) dissociation of the zona pellucida, followed by aspiration with a second pipette.

Polar Body Biopsy

A second general approach is to remove either the first or second polar body, or both. Polar body biopsy (PBB) is technically similar to blastomere biopsy, but the embryo is not entered per se. Literally this constitutes preconceptional diagnosis, as opposed to preimplantation diagnosis per se. Suppose a woman and her mate were both heterozygous for the same autosomal recessive disorder. A polar body having the mutant allele should be complemented by a primary oocyte presumed to have the normal allele. The normal oocyte could thus be allowed to fertilize in vitro and then transferred for potential implantation. Conversely, if the polar body were normal, fertilization would not be allowed to proceed because the oocyte would presumably contain the mutant allele. The same reasoning would apply for detecting aneuploidy. A polar body containing only one chromosome 21 (as determined by fluorescent in situ hybridization with a chromosome-specific probe) should be complemented by a primary oocyte also containing a single number 21 chromosome. If the first polar body failed to show a 21, the oocyte can be presumed to have two number 21 chromosomes, which would lead to a trisomic zygote.

Disadvantages include inability to assess the paternal genotype, precluding application if a father has an autosomal dominant disorder. Lacking information about paternal transmission also results in reproductive inefficiencies. Presence of a mutant autosomal recessive allele in a polar body precludes allowing that oocyte to be fertilized, whereas in reality the embryo resulting from fertilization of that oocyte would be affected only if the sperm also contained the father’s mutant allele. If it were known that the father’s normal allele were transmitted, the embryo would be heterozygous and phenotypically normal.

Another disadvantage is recombination, the meiotic phenomenon that occurs routinely between homologous chromosomes. Diagnostic hazards resulting from recombination are a greater problem for genes located nearer the telomeres because these genes display recombination frequencies approximating 50%.

Blastocyst (Trophectoderm) Biopsy

An underlying difficulty when aspirating either the polar body or a blastomere from the 6–8 cell embryo is that diagnosis must depend upon only a single cell, or occasionally two. Many more cells would be available if one biopsied the 5- to 6- day blastocyst, which contains hundreds of cells. The strategy is to biopsy the trophectoderm overlying the anembryonic pole. One major problem is that blastocysts are less readily obtainable as 6–8 cell embryos. Blastocysts can be obtained by culture in vitro, but this process has traditionally proved very inefficient. After 3 to 4 days in vitro, embryonic development is not sustained efficiently to the blastocyst stage.

In PGD, diagnosis may depend upon information from a single cell. Indeed, this unique characteristic of PGD is likely responsible for much of the diagnostic inaccuracy attributed to PGD, given the likelihood that one of these early cells may be characterized by a cytogenetic or genomic alteration not found in the rest of the zygote. Metaphase analysis is possible for only a single cell, but this is not optimal. Other approaches to determine chromosomal status include comparative genome hybridization (CGH) [30]. However, none of these approaches can prevent diagnostic inaccuracies resulting from inherent cellular mosaicism. The preferable approach for chromosomal analysis in PGD involves fluorescence in situ hybridization (FISH) with chromosome-specific probes. FISH is a molecular genetic technique based on analysis of DNA sequences that are chromosome-specific. Chromosome-specific probes can hybridize to denatured DNA, in either metaphase or interphase. In interphase cells the number of FISH signals should equal the number of a given chromosome. Using a chromosome 21-specific probe would detect trisomy 21 if three signals were visualized in the nucleus. This approach is also well suited for sex determination, X and Y probes being utilized concurrently. A diagnostic dilemma unique to PGD is the very high frequency of mosaicism in the early embryos.

Molecular analysis of a single cell requires techniques more sensitive than those necessary for analyzing chorionic villi or amniotic fluid cells. The small amount of DNA in one cell (6 pg), is insufficient for conventional multi-cell molecular diagnostic approaches. In traditional polymerase chain reaction (PCR) methods, a targeted DNA sequence flanked by specific primers is amplified to increase numbers of copies, usually 10⁵ to 10⁶ fold. Amplification from a single cell using a single set of primers is successful if a repetitive DNA sequence is being evaluated (e.g., Y-chromosome). However, technical modifications are necessary when analyzing unique sequence mutations because sensitivity for detecting unique sequence DNA cannot be readily achieved below approximately 100 pg DNA. Rather than a single round of PCR (20 to 30 cycles), nested primers are required. The second set of primers is constructed internal to the first. Following approximately 20 cycles of PCR with the outer primers, the product is subjected to PCR with a second (internal) set of primers. Using nested PCR, as little as 1 pg of DNA suffices for diagnosis.

The list of Mendelian disorders for which PGD has been clinically used has grown rapidly in recent years. These include cystic fibrosis (F508del homozygosity), Tay-Sachs disease, Rh (D), sickle cell anemia, certain β-thalassemias, phenylketonuria, spinal muscular atrophy and myotonic dystrophy, and adenosis polyposis coli and Li-Fraumeni syndrome. Of special note is that PGD has been accomplished for a variety of adult onset autosomal dominant disorders, including Huntington disease and Marfan syndrome. Any Mendelian disorder whose gene is localized or whose molecular basis is known is potentially detectable by linkage analysis if not otherwise.

4Genetic conditions of special concern and their prenatal identification

Within the context of this chapter, prenatal diagnosis is defined as a process that rules out or detects and manages embryonic and fetal abnormalities or disease and that provides information to parents and physicians to optimize both pregnancy outcome and family well-being. Although prenatally diagnosable disorders may be categorized in many ways, it is simplest to consider only two groups of couples or families who may benefit from prenatal testing. Theoretically the two groups encompass all types of embryonic or fetal disorders that may be amenable to prenatal testing or management.

The first are those patients who themselves, or whose physicians, are (or theoretically can be) aware before conception of a known risk for fetal abnormality or unfavorable pregnancy outcome. This is the group of families in whose medical, family, or personal history there are known factors that might result in prenatally detectable or even preventable fetal disorders. They might include families of increased parental age, those with previous infants with chromosomal abnormalities or other genetically determined disorders, or families of specific ethnic origin that might increase their risk for abnormal pregnancy outcome.

The second are those who during pregnancy, without previously recognizable known risk factors, have an unexpected risk of fetal abnormality or potential for unfavorable outcome identified. Examples include those in whom a routine blood test or prenatal ultrasound reveals an unexpected abnormal finding.

1. Metabolic Single Gene Disorders Detectable by Biochemical Analysis

In considering pregnancy in association with biochemical disorders, most of which are inherited in autosomal recessive fashion, it is important to take into account the type of disorder, whether it is a parent who is affected or the fetus which is at risk, and whether and by what method the disorder is prenatally detectable. Clinical consideration of biochemical disorders can be divided into two categories: 1) carrier (heterozygote) screening to identify couples at risk for having an affected fetus and thereby candidates for prenatal diagnosis; and 2) diagnosis and management, of a first affected child.

Important technological developments in recent years have facilitated prenatal diagnosis of metabolic disorders. In fact, a number of metabolic disorders can now be diagnosed in the fetus prenatally or others through newborn screening [1] as listed in Table 1. Diagnosis relies on either the demonstration of a block in a metabolic pathway by examination of amniotic fluid cells, a sampling of chorionic villus biopsy, preimplantation diagnosis via blastomere of the 6- to 10-cell embryo, or through DNA studies that document the presence of a mutant gene or a marker associated with the mutant gene. Demonstration of a metabolic block requires either the measurement of the enzyme activity with natural or artificial substrates or the ability to measure flux through a metabolic pathway. Flux refers to the depletion rate of a labeled substrate that may feed into a metabolic pathway distant from the actual metabolic block. Alternatively, one can measure the metabolites that accumulate secondary to the metabolic block. The tissue used for prenatal diagnosis depends on the level of expression of the metabolic pathway, that is, tissue specificity, which may be influenced by the developmental age of the conceptus [2].

For disorders in which the specific enzyme defect can be demonstrated in cultured skin fibroblasts, prenatal diagnosis can be accomplished using cultured amniotic fluid cells or cultured or directly analyzed chorionic villus tissue. This approach requires determination of normal values of the specific enzymes, under identical conditions. Care must also be taken to ensure that heterozygote values do not fall within the range expected for affected individuals. In addition, in chorionic villus analysis, one must avoid maternal contamination by removing decidual tissue before culture or direct enzyme assay [3].

With our increasing knowledge and experience with the human genome, there are only a few diseases in which the enzyme is not normally expressed within chorionic villi or amniotic fluid cells and for which DNA assays have not yet been developed. Prenatal diagnosis of these disorders may rely on either detection of an abnormal metabolite in the cell-free amniotic fluid or measurement of the enzyme in fetal tissue known to express the enzyme. The latter approach requires a biopsy of fetal organs or collection of fetal blood by cordocentesis [3].

There are inherited metabolic diseases, such as neuronal ceroid lipofuscinosis (NCL), for which there are no specific biochemical tests. In these instances, morphologic diagnoses may be possible; for example, electron microscopy (EM) of uncultured amniotic fluid cells or chorionic villi has been diagnostically valuable for NCL. Amniotic fluid cells obtained at 16 weeks’ gestation in NCL contain curvilinear bodies. Curvilinear cytosomes are present in the placenta of the affected fetus [4]. Similarly, fingerprint cytosomes, characteristic of juvenile NCL, are present in the syncytiotrophoblast, making prenatal diagnosis possible by CVS.

Prenatal diagnosis of mucolipid storage diseases is possible by biochemical studies as well as by EM of uncultured and cultured amniotic fluid cells that contain numerous membranous and granular cytosomes [5, 6]. By using histochemistry, immunofluorescence, and EM to identify the peroxisomes in chorionic villi or cultured villus tissue, peroxisomal deficiency disease may also be diagnosed prenatally [7, 8].

The following discussion of selected disorders illustrates the principles and practices applied to prenatal diagnosis of metabolic disorders. More information is available in the references cited.

Tay-Sachs disease, a lysosomal sphingolipid storage disorder, is a typical example of a disorder that is detectable in the first pregnancy of a couple at risk. Its clinical manifestations, occurring more frequently in the Ashkenazi Jewish, Cajun, and some French Canadian populations, are caused by the absence of the alpha subunit of the enzyme hexosaminidase A (Hex A), which is a normal gene product. The clinical course of the infantile form is relentlessly fatal, usually within 4 to 6 years after birth. It is characterized by loss of developmental skills and milestones in an initially healthy-appearing infant; a typical cherry-red spot on the ocular macula; gradual deterioration (associated with intracellular accumulation of the abnormal metabolite, GM₂ ganglioside) of visual, auditory, and motor function; and frequent seizures. Hex A levels in the serum of heterozygotes are decreased by approximately 50% in comparison to those in unaffected homozygotes and are easily detectable in populations at risk. In women who are pregnant or on oral contraceptives, measurement of serum Hex A levels may lead to false positive results. Therefore, a leukocyte Hex A assay is used for these women.

The International Tay-Sachs Disease Network, through the participation of multiple centers around the world with well-organized, targeted, and accurate screening surveys, has been indirectly responsible for the dramatic reduction of this fatal disorder from about 60 newly diagnosed patients per year in the United States around 1970 to 3 per year in 1992 [31]. Most cases were prevented through prenatal diagnosis and pregnancy termination of affected fetuses of couples found through screening to be at 25% risk. The fetal diagnosis was initially established by quantitative Hex A determination in amniocytes or chorionic villus tissue; more recently it has become molecularly based and several mutations, as well as multiple forms of disease in addition to the infantile one, are known to exist. Thus, Tay-Sachs disease has become not only a preventable (and in its severe, infantile form, now almost extinct) disorder. In addition to Tay-Sachs disease, it is recommended that Ashkenazi Jewish individuals be offered screening for Canavan disease, familial dysautonomia, and cystic fibrosis. Additional tests that may be added to a screening panel include Fanconi anemia group C, mucolipidosis IV, Bloom syndrome, and Gaucher disease [32].

For most inborn errors of metabolism, carrier screening is not available. After the birth of an affected child the parents are identified as heterozygotes for a particular disease and prenatal diagnosis may be an option in subsequent pregnancies.

5Single gene disorders detectable by direct molecular analysis

There is an ever increasing list of disorders detectable by molecular analysis, which permits identification of the mutation directly in fetal (or embryonic) tissue. However, this should not be offered to any couple unless the couple is known to be at risk for the disorder in question, and unless testing is preceded by extensive, supportive, and informative counseling of all aspects and implications of potential results, as well as complications and ambiguities. Table 2 provides a selective list of disorders prenatally diagnosable by direct identification of the mutation, using a variety of DNA analytical methods, as well as the chromosomal location and mechanism of inheritance. Prototypes of more commonly occurring disorders within this group include cystic fibrosis (CF), sickle-cell disease, Duchenne muscular dystrophy.

Cystic fibrosis

CF is a multi-system disease inherited in autosomal recessive fashion. In the white population approximately 1 in 4000 individuals in North American and northern Europe are affected). The frequency of heterozygous carriers is about 1 in 29 individuals in Caucasians (Northern European origin) and Ashkenazi Jews. The heterozygote frequency is lower in Hispanic, Asians and African Americans. CF is a disease of exocrine glands that results in viscid secretions which most commonly and most severely affects the respiratory and gastrointestinal systems. In 1989, the CF gene was mapped to chromosome 7 [33]. Its gene product, a transport protein called cystic fibrosis transmembrane conductance regulator (CFTR), functions as a chloride channel activated by cAMP-dependent protein kinase A phosphorylation [33]. More than 2,000 different mutations have been discovered that can potentially result in an abnormal CFTR, but only about 10–15% of these actually cause the disease CF.

The most common disease-associated variant is F508del (originally referred to as ΔF508 and now more properly referred to as p.Phe508del), which is a deletion of 3 base pairs that code for phenylalanine. The only other mutation present in high frequency in any ethnic group in W1282X, the most common mutant allele in Ashkenazi Jews.

The pathogenesis in cystic fibrosis involves a blocked or closed chloride channel in the epithelial cell membrane. As result, chloride ions are trapped inside the cell. This draws sodium ions and water into the cell resulting in dehydration of mucus secretions. In the lungs, this can lead to airway obstruction and predisposition to infection, dyspnea, bronchiectasis and pulmonary fibrosis. The major cause of morbidity and mortality in these individuals is progressive, chronic bronchopulmonary disease. In the pancreas, the dehydrated secretions are also insufficiently alkalinized, hence causing not only duct obstruction but also a reduction in digestive enzymes and fibrosis. The fibrosis can involve the endocrine cells and may lead to not only pancreatic insufficiency but also diabetes mellitus. In the intestinal tract, the lack of digestive enzymes leads to malabsorption of fat and protein and the altered mucus can lead to bowel obstruction. Infants born with meconium ileus have a significantly higher risk of subsequent malnutrition. Sweat electrolytes are also abnormal with elevated concentrations of sodium and chloride. This in turn can lead to heat intolerance, hyponatremia, and a hypochloremic metabolic alkalosis.

The clinical presentation and course of individuals affected with CF is highly variable. The median age at diagnosis by traditional methods is 6–8 months. Some individuals have severe pulmonary and/or gastrointestinal disease, whereas others have mild expression with presentation in adolescence and young adulthood. There is a close correlation of genotype with phenotype in relation to pancreatic function; however, identification of the specific CFTR mutation has not proved highly predictive of the severity of pulmonary disease. Modifier genes and environmental factors undoubtedly play roles in this varied expressivity. Recently, a project known as CFTR2 has provided very helpful descriptions of the clinical signs and symptoms that can be expected with individual mutations [34].

With advances in medical treatment, the prognosis for individuals affected with CF has improved. In the United States, median survival now exceeds 40 years. Data from the Cystic Fibrosis Foundation indicate that about half of affected individuals are 18 years or older, and survival into the fifth decade is not uncommon (https://www.cff.org/2013_CFF_Annual_Data_ Report_to_the_Center_Directors.pdf).

Women with CF have compromised fertility. They may show delayed onset of menarche; those with malnutrition and advanced lung disease often have anovulatory cycles with secondary amenorrhea. Dehydrated and thickened cervical secretions also may pose a physical barrier to sperm penetration of the cervical os. Nonetheless, as more women with CF enter reproductive age, they should be appropriately counseled about contraception, pregnancy and the potential increased risks associated with CF. Pregnancy rates among CF women have increased. During pregnancy these women are at risk for respiratory and cardiac compromise because their underlying lung disease has the potential to be further exacerbated by gestationally related hypoxemia and pulmonary hypertension. The changes in ventilatory drive, cardiac function, blood volume and increased nutritional requirements all pose potential problems for the woman with CF [35]. Most women with CF can complete a successful pregnancy. Careful monitoring of lung function and ensuring adequate energy intake is recommended. Affected women should be counseled about the symptoms and signs of preterm labor and routine cervical exams should be considered.

As mentioned above, over 2,000 mutations have been identified in the CFTR gene, but only about 10% of them are CF disease-causing [34]. The F508del mutation is represented in almost all populations, although its relative frequency varies among different geographic locations. The highest frequency for F508del is observed in white populations, especially of Northern European origin, where it accounts for 70–75% of the CF alleles. F508del is present in 48% of CF alleles in African Americans, 46% in Hispanics, 30% in Asian Americans, and 30% in Ashkenazi Jews. About 15 to 20 other mutations account for 2–15% of the other CF alleles, depending on the ethnic group studied. Most remaining mutations are rare. A panel of 23 mutations will detect 87% of cases in Caucasians and differing frequencies in other ethnic groups. In several populations, the combination of F508del with other ethnic-specific mutations allows for a 90% to 95% detection rate: Ashkenazi Jews, Celtic Britons, French Canadians from Quebec, and some Native Americans [36]. In Ashkenazi Jews, F508del and W1282X alone account for 85% of CF alleles.

An important implication of the various mutation detection rates among different populations is the number of couples in whom both partners can actually be identified as CF carriers thereby enabling unequivocal prenatal diagnosis. For example, if the mutation detection rate is 75% and one partner is determined to be a carrier and the other not, the risk of having an affected offspring is 1 in 400. By contrast, if the mutation detection rate is 90% and one partner is determined to be a carrier and the other not, the risk for an affected offspring is 1 in 1000. If both partners are screen negative, the risk is 1 in 240,000 [37]. Experience during the period of 2005-11 when CF newborn screening was being implemented nationwide demonstrated the practical ramifications of a sensitivity as low as 75%, i.e., the usual situation in which the mother is a carrier who tests positive for one CFTR mutation and the father/partner is negative. In that circumstance, after the prospective parents have been reassured, no matter how cautiously this communication occurs, countless pregnancies have been continued and numerous infants have been born with a positive newborn screening test and then a confirmatory sweat test, thus diagnosing CF. The implications of such outcomes need to be explored with regard to both prenatal and neonatal screening.

If amniotic fluid analysis were possible for the cystic fibrosis gene product (i.e., mutant protein), the inability to detect all mutations would not matter so much in situations where one parent carries an identified CF mutation and the other is screen negative, but still has a residual risk of carrying an unidentified CF mutation. Unfortunately, no such assay exists, meaning that prenatal diagnosis is not possible except to exclude the fetus who inherited the mutation from the known heterozygous parent. Measuring microvillus intestinal enzymes in amniotic fluid is specifically not helpful, for a low value is more likely to connote a false-positive than a true-positive value.

Another issue with carrier screening is genotype-phenotype correlations. Predictions of the severity and course of pulmonary disease are variable for specific mutations, especially for fetuses detected to be compound heterozygotes; however, prediction for pancreatic function is generally reliable [38]. The issue of phenotype being different from CF with pancreatic and pulmonary insufficiency becomes most relevant when one is confronted with a male having two CF alleles that results in congenital bilateral absence of the vas deferens (CBAVD).

Males with classic CF are virtually always sterile due to CBAVD. There is an increased frequency of CF mutations in otherwise healthy men with CBAVD. About 25% of these men with CBAVD are compound heterozygotes for two mutant alleles (one being a “mild” mutation); half have one identifiable mutation, and the remainder have no identifiable mutation [39]. Eighty-four percent with no identifiable mutation carry this splice site variant for a CF mutation called the 5T allele and 25% with no identifiable mutation carry a splice site variant. The 5T refers to the number of thymines present at the end of intron 8, which alters the likelihood of proper splicing of exon 9. The 7T allele is usually associated with normal splicing. The 5T allele is present in men with CBAVD at a four-fold higher frequency than the general population [40–42].

6Newborn screening: Principles and practices

Newborn screening may be defined [5] as “a population based public health program applying preventive medicine in defined regions to reduce newborn morbidly and mortality from certain biomedical and genetic disorders by using presymptomatic detection/diagnosis with dried blood specimens analyzed in central laboratories and employing automated procedures and linked to clinical follow-up systems.” It should be emphasized that each aspect of the sequence from blood collection to follow-up care is essential for quality assurance. The importance of consistent processes and procedures has been emphasized in a series of guideline publications of the Clinical and Laboratory Standards Institute (CLSI) that describe best practices [55–60].

The fundamental principle for achieving benefits through newborn screening is illustrated in Fig. 1. In essence, early diagnosis through newborn screening provides children with an opportunity for better outcomes when there is a detectable preclinical stage of a congenital disorder, ideally soon after birth, and effective, accessible therapy exists that prevents irreversible abnormalities such as severe morbidity (e.g., mental retardation) or death. Often the biologic onset and symptomatic onset are chronologically close, and only limited time (days to weeks) exists from that stage to a “point of no return” where irreversible pathology has supervened. The goals of newborn screening are summarized in Table 3. For most genetic diseases included in newborn screening programs, prevention of mental retardation and developmental disabilities has been the primary objective and indeed was the driving force for the advent of this model public health initiative. Also, most of these disorders are autosomal recessive and very difficult to diagnose using the traditional history/physical exam strategy; even the very best pediatricians will fail to diagnose “accurately and promptly” enough to prevent long term adverse consequences [1].

Fig.1

Principles of achieving benefits for infants through early diagnosis and therapy.

As described previously, the first newborn screening programs were implemented in the early 1960’s for PKU [12], a disorder associated with phenylalanine hydroxylase deficiency. Initially voluntary, the PKU test became mandatory by legislative actions throughout the nation, such that by 1973 a total of 43 states had formal statutes. In general, state health departments, particularly their maternal and child health programs funded by Title V of the Social Security Act, assumed responsibility for implementing the new laws and ensuring neonatal blood collection and analysis. Thus, PKU became the prototype [61, 62] genetic disorder for newborn screening because it is very difficult to diagnose early in primary care practice, has a relatively short latent (preclinical) stage, a high risk for serious, irreversible pathology (mental retardation), and can be effectively treated using dietary phenylalanine restriction [63]. Although early implementation efforts were fraught with problems, attention to the entire newborn screening system led to quality improvement as centralized laboratories became responsible for the lab tests and a network of centers assumed the essential follow-up role. All states screen every newborn for PKU, but the analytical tests have varied considerably from the original bacterial inhibition assay of Guthrie [5] to automated methods using sophisticated tandem mass spectrometry (MS/MS) [4, 58]. As described subsequently under “The Future of Newborn Screening” the antiquated methods from the 1960’s are being replaced by MS/MS [58].

All states also screen their newborn populations for congenital hypothyroidism using the same dried blood specimen and a variety of radioimmunoassay methods [4]. The testing protocols have varied and included simple measurement of blood thyroxin, a combination of thryopin and thyroid stimulating hormone (TSH), or TSH alone. There has been a trend to use primary TSH screening in recent years because it can be both effective and cost-effective [5]. As shown in Table 4, congenital hypothyroidism is five times more common than PKU, and thus was very attractive in the 1970’s to add to newborn screening panels, especially in view of the curative therapy available (thyroxin) [64]. In the U.S., about 1200 confirmed cases per year are identified with an assumed sensitivity of 100%, specificity of 98.5%, and PPV of about 2%, i.e., 50 false positives for every diagnosis [65]. A variety of problems were encountered as implementation occurred nationwide and quite rapidly [5, 66, 67] (e.g., imperfect sensitivity of laboratory protocols and follow-up problems) but congenital hypothyroidism screening has been a great success. In retrospect, the lessons learned from flawed implementation of PKU screening promoted an expeditious, constructive response of leaders that was very effective. AAP guidelines include the recommendations provided below [68].

Another hereditary endocrinopathy, congenital adrenal hyperplasia (CAH), is now screened for in about half the states as part of newborn screening programs. The goal of early diagnosis through neonatal screening is to prevent deaths associated with the salt wasting form of the disease. This disorder is identified by radioimmunoassay of 17-hydropyprogesterone and is readily treated with cortisol replacement [5, 69]. The major problem with CAH screening is the relatively low PPV of 0.5%, which leads to large proportion of false positive tests, especially in premature infants [5], i.e., up to 200 for each diagnosis [65].

Most regions also screen newborns for galactosemia in an effort to prevent mental retardation. A variety of methods are used [4]. Classical galactosemia is caused by a deficiency of galactose-1-phosphate uridyl-transferase. A growing number are now screening for biotinidase deficiency, which develops because of an ability to recycle biotin. Both of these disorders occur with an incidence of 1:70,000 and thus are somewhat infrequently diagnosed, i.e., about 50 cases of each disorder per year among the approximate 4 million newborns in the USA. Delays in diagnosis, however, can be catastrophic for those 50 infants.

Other hereditary metabolic disorders are being included in newborn screening panels, although the value of early diagnosis is debatable for some conditions. For instance both homocystinuria, which is caused by defective metabolism of sulfur – containing amino acids, and Maple Syrup Urine Disease (branchedchain ketoaciduria) screening were temporarily discontinued in Wisconsin because of their very low incidence (about 1:225,000) and uncertain benefits. Indeed, these two disorders do not meet the criteria used in Wisconsin [70, 71, 72] (listed below) or the World Health Organization (WHO) criteria [72]. With their low incidence only about 18 cases of each disorder are likely to be diagnosed per year in the United States. Unfortunately, preventing mental retardation may not be feasible through screening newborns. On the other hand, these disorders will progress and be life threatening or fatal with long delays in diagnosis and their detection may be more feasible with new analytical, automated technology.

The implementation of tandem mass spectrometry has promoted screening for numerous disorders of amino acid and organic acid metabolism as well as disturbances in fatty acid oxidation (Table 5). Tandem mass spectrometry is arguably the most significant procedural advance in newborn screening since the “Guthrie card” method of obtaining dried blood specimens was introduced in the 1960’s. Consequently, we have included a detailed review of this technological advance and recommend reading the CLSI guidelines [58] for more information on laboratory aspects such as quality control. Prior to MS/MS, the laboratory strategy was to use a separate biochemical test or even two tests for each disorder. Automated, computer-assisted MS/MS transformed that paradigm into a modern strategy of using that method to screen for multiple hereditary metabolic disorders. Nearly a quarter century of evolutionary advances, however, was needed to reach the current status of routine MS/MS in newborn screening laboratories. The early phase depended on a combination of liquid chromatography and mass spectrometry detection of metabolite derivatives (analytes). By the 1990’s, instrumentation advances with techniques such as fast atom bombardment and fast ion bombardment coupled to a series of units controlling separation, fragmentation and detection set the stage for creating a high throughput system employing electrospray ionization as the optimal sample delivery interface. With the addition of computer-assisted data monitoring, it became possible for a newborn screening laboratory to analyze up 500 samples per day and screen for 50 or more disorders (Table 5).

MS/MS-detectable hereditary metabolic disorders include aminoacidopathies, fatty acid oxidation deficiencies, organic academia disturbances, and urea cycle defects. The detection of aminoacidopathies and urea cycle disorders relies on determination of elevated amino acids such as phenylalanine for PKU and lencine/isoleucine/valine for MSUD. The abnormalities in fatty acid oxidation and organic acid metabolism are identified by measuring acylcarnitines, an intermediate metabolite consisting of a fatty acid or organic acid and carnitine, which is required for transfer across the mitochondrial membrane for oxidation. Each disorder shown in Table 5 has a distinct profile of abnormal levels of specific acylcarnitines. Although these disorders are individually uncommon, when MS/MS as a single test is used to facilitate their early diagnosis, the cumulative incidence among newborns is approximately 1:4000 and thus is similar to congenital hypothyroidism and CF (Table 4). Decisions to add MS/MS and screen newborns with an expanded panel have been made in different regions based on a variety of criteria, including technical capability and financial capacity. When Wisconsin decided to proceed in the 2000, it was clear that predetermined criteria [70] summarized below were met [73]. Since then most states have implemented MS/MS and excellent guidelines for its application published [58].

Medium chain acyl-CoA dehydrogenase deficiency (MACD) is the prototype disorder in this category and causes potentially fatal vomiting, lethargy, encephalopathy, hepatomegaly, seizures and coma [74]. The approximate incidence is 1:10,000. Effective treatment to prevent a metabolic crisis depends on a low fat, high carbohydrate diet. Early results of screening are promising.

When it became clear that penicillin prophylaxis could be life saving for children with sickle cell disease (SCD) [45], hemoglobinopathy screening was incorporated into many programs and now is offered almost universally in the USA. Originally depending on protein electrophoresis, after using cellular acetate, programs now use other molecular diagnostic methods such as isoelectric focusing and HPLC [4]. Prevention of deaths associated with pneumococcal infections can be readily achieved in children with SCD [4, 53] and thus it seems surprising that some regions don’t include hemoglobinopathy screening in their panels. Although a lower incidence (e.g., 1:40,000) of SCD associated with regional demographic differences (i.e., a lower proportion of African-Americans) is offered as an explanation, some of these states screen for hereditary metabolic disorders that are even less prevalent [8]. This situation is typical of the regional disparities currently evident in the newborn screening programs of North America and suggests the need for national policy development [8, 71, 72, 73].

In the case of cystic fibrosis (CF), a quarter century of extensive, well-organized research in the US, Western Europe, New Zealand, and Australia [14] led to national recommendations [75] in 2004 by leaders of the Centers for Disease Control and Prevention (CDC) following their earlier conclusion that more research was needed [76]. After CDC recommended more CF newborn screening, all states began planning efforts and by 2011 the entire USA and many European countries had implemented such programs. The methods that have been used commonly are described in Table 6 and more detail is available in the CLSI Guideline document [60]. The role of the CDC in reviewing evidence and issuing policy recommendations for CF newborn screening was unique. In addition, the long term emphasis on prospective research related to CF newborn screening is also noteworthy because in general only limited systematic outcomes research has occurred regarding other genetic conditions included in newborn screening panels. Other features are also special such as the screening test used most commonly which involves DNA analysis. Specifically a two tiered method employing immunoreactive trysinogen (IRT) is used first and this is followed by DNA analysis for CFTR mutations when the IRT level exceeds a cutoff selected to maximize sensitivity. The IRT/DNA test was the first molecular genetics method applied to complete populations in community-based screening [70]. Because there are more than 2000 CFTR mutations, the DNA component has been challenging but fortunately about 70% of CF chromosomes have the principal and first mutation discovered [33], namely F508del (deletion of 3 bases at codon 508 and loss of phenylalanine in the mutant chloride channel protein). Furthermore, about 90% of CF patients in North America, Europe, and Australasia have at least one F508del mutation, so that simple IRT/DNA (F508del) testing detects most of them and presumptively diagnoses half the CF population who are homozygous for F508del [1, 30, 77]. However, the recent CF Founation consensus guidelines for diagnosis emphasizes that even these infants need to have a diagnostic sweat test performed [78]. The IRT/DNA (F508del) screening test has a high PPV (17%) compared to other disorders on routine newborn screening panels [79]. In addition, there are advantages to using IRT/DNA (CFTR) multi-mutation analysis, and development of such tests has proceeding rapidly [1, 31]. New analytical technologies such as next generation sequencing [80] provide an opportunity to detect more than 100 CFTR mutations simultaneously within a few hours. As shown in Table 4, CF is relatively common among disorders included in newborn screening programs. In fact, CF is the most common, life-threatening autosomal recessive disorder among Caucasians, and over 1000 patients per year are diagnosed in the US.

Early diagnosis through IRT/DNA screening provides an opportunity to prevent the potentially fatal salt depletion with sweating, and the severe malnutrition that can cause long term problems such as growth retardation and cognitive dysfunction [14]. In addition, early diagnosis facilitates therapy of CF lung disease, although currently it can’t be prevented [14]. On the other hand, the development of CFTR modulator therapy [81] offers the potential to ameliorate lung disease through enhanced chloride channel functioning.

CF illustrates the challenges that need to be addressed to enhance the quality of newborn screening programs, including: 1) improved testing protocols to optimize both sensitivity and positive predictive value; 2) enhanced quality assurance with consistent policies; 3) better education of families and providers, especially in the pre-screening, prenatal period; and 4) improved risk communication for both true and false positive families, i.e., better genetic counseling. Specific recommendations for improvements include:

7. The future of newborn screening

Although the time frame for future evolution of newborn screening programs is difficult to predict, it is clear that more and better tests will be applied to newborn populations to ensure early diagnosis and hopefully reduce health disparities further. More precise and technologically advanced, though not necessarily more expensive, methods are readily available such as tandem mass spectrometry, next generation sequencing for DNA analysis, and possibly whole genome sequencing (WGS); these procedures are likely to be implemented during the next decade.

Clearly tandem mass spectrometry is an evolution that is already causing a revolution. Not only are more regions implementing MS/MS [58], but it is causing an expansion of testing panels to literally quadruple overnight the number of congenital disorders being screened for routinely. Completion of the Human Genome Project has given us molecular genetics analytical biotechnology capability far beyond what is currently being used in newborn screening programs, and thus more DNA-based tests and possibly whole genome sequencing will emerge over time. Indeed, the CDC [76] referred to newborn screening for CF using IRT/DNA analysis as “a paradigm for public health genetics policy development.”

It is safe to predict that as DNA-based assessment of newborns for CF proliferates, other DNA-linked screening tests will be implemented. For instance, CAH screening programs, because of their low positive predictive value (up to 200 false positives for every diagnosis) [64], could be improved quite significantly with a two-tier strategy using 17-hydroxyprogesterone coupled to detection of 21-hydroxylase mutations [5]. SCID screening provides an excellent example of how public health laboratories can quickly adapt to new, DNA-based technologies [85–88].

In addition to more and better tests, the glaring lack of consistencies among regions needs to be addressed effectively and eliminated through quality improvement methods that address all six components of the newborn screening system (http://mchb.hrsa.gov/screening/) namely: 1) education of professionals and parents; 2) screening, i.e. specimen collection, submission, and testing; 3) follow-up of abnormal and unsatisfactory test results; 4) confirmatory testing and diagnosis; 5) medical management and periodic outcome evaluation; and 6) system evaluation and quality assurance.

The variations relate to policies, financing, testing panels and protocols, follow-up efforts, and education. Consequently, the American Academy of Pediatrics Newborn Screening Task Force has issued “a call for a national agenda on state newborn screening programs.” Some of their most important recommendations follow.

The AAP ‘call for a national agenda’ was followed by a landmark effort under the leadership of the ACMG to address some of the critical issues. Commissioned by the Maternal and Child Health Bureau of the Health Resources and Services Administration (US Department of Health and Human Services), a consensus-developing group of multidisciplinary experts published their analyses, conclusions, and recommendations recently in a document entitled “Newborn Screening: Toward a Uniform screening Panel and System” (http://mchb.hrsa.gov/screening/). This group recommended a core panel of 29 genetic conditions based on predetermined criteria “as primary targets for screening that warrant the full breadth of the newborn screening system” (see Table 1). In addition to congenital hearing impairment, the list includes biotinidase deficiency, classical galactosemia, congenital adrenal hyperplasia, congenital hypothyroidism, cystic fibrosis, three hemoglobinopathies associated with the hemoglobin S allele, five disorders of fatty acid oxidation, six amino acidurias, and nine organic acidurias. Subsequently, screening for severe combined immunodeficiency (SCID) with molecular methods [86–88] was added to the list, so that it now the core group recommended consists of 30 genetic disorders, as shown in Table 1. These new methods for SCID involve quantitating T-cell receptor excision circles (TRECs) using real-time quantitative polymerase chain reaction on DNA extracted from dried blood spots to detect infants with T-cell lymphopenia. Multiplex technologies such as MS/MS will detect blood analyte abnormalities found in 23 of the 30 conditions of the core panel. The group also emphasized that newborn screening “programs are obligated to establish a diagnosis and communicate the result to the health care provider and family” and furthermore to make available the information “when newborn screening laboratory results definitely establish carrier status.”

This AAP list of recommendations describes a wide range of issues that need to be addressed in quality improvement efforts. Enhancing quality assurance is difficult because although newborn screening is a system, its application varies markedly in different regions. In addition, as advances in biotechnology occur, the rate and effectiveness of implementation also vary. Furthermore, the ability of providers and parents to understand the laboratory results, whether they be diagnoses or false positives, is variable at best. Currently, about 5% of newborn screening tests are associated with a reportable result that needs to be explained to the parents. This amounts to 200,000 families per year in the United States or about 4,000 per week (ironically, the same number as the total diagnosed through newborn screening on an annual basis). The communication challenges faced by providers are daunting. The content of the message, how and when to deliver the stress-promoting news to parents, and the kind of follow-up required are all difficult issues. It is ironic that the biotechnology has advanced far beyond the capability of the system to meet the most fundamental of human interactions – communication. Clearly, more research is needed to improve risk communication methods and to evaluate their impact.

Despite the need for multidimensional improvement, it should be emphasized that newborn screening programs have been one of the great public health success stories. Thousands of children have benefited with reductions in mortality and morbidity. Even greater success will undoubtedly be achieved in the future.