October 2011: Volume 37, Number 10
How Newborn Screening Algorithms Detect Disease
By Stanley F. Lo, PhD
Cystic fibrosis (CF) is an autosomal recessive disease that primarily affects young children. Considered the most lethal genetic disease in Caucasians, it affects multiple systems, including the respiratory, gastrointestinal, genitourinary tracts, pancreas, liver, and sweat glands. Affected children commonly experience pancreatic insufficiency, chronic respiratory infections, and elevated sweat concentrations of chloride. In fact, exocrine pancreatic insufficiency is present in >85% of CF individuals. Left untreated, the disease causes severe malnutrition, decreased growth, and eventual death within the first decade of life. Pancreatic enzyme replacement therapy and vitamin supplementation can help minimize gastrointestinal symptoms and nutritional complications.
The inability of the respiratory system to prevent infections is responsible for most of the morbidity and mortality associated with CF. Persistent respiratory infections, especially from Pseudomonas aeruginosa, may eventually lead to bronchiectasis. Children with decreased pulmonary function often progress to intractable respiratory failure. However, some treatments relieve the pulmonary disease. For example, physical techniques for airway clearance and medical therapies (mucolytics) help clear respiratory secretions, and antibiotics and other medications reduce the frequency of pulmonary infections and inflammation. Lung transplantation is also a potential consideration for treatment. With the availability of these treatment strategies, today the life expectancy of individuals with CF has reached a median age of 37 years.
This article will review the genetics of CF disease and describe in detail the laboratory testing algorithms used for CF screening of newborns, as well as some of the advantages and disadvantages of each.
The Genetics of CF
In 1989, researchers identified the gene responsible for CF disease. A 250-kb gene composed of 27 exons encoding 1,480 amino acids, the transmembrane conductance regulator (CFTR) protein, is located on the long arm of chromosome 7. Today we know that the prevalence of mutations in this gene is highest in white populations of northern Europe, North America, Australia, and New Zealand. Of the more than 1,800 mutations reported, the most common results in deletion of phenylalanine at position 508. This mutation is found in roughly 50% of all homozygous CF individuals, and >80% of CF carriers within the European Caucasian population. Nearly all of the remaining mutations have a prevalence of no more than a few percentage points; however, some exceptions exist in specific populations. For example, in Ashkenazi Jews, the W1282X mutation has a prevalence of 60%, and in the Celtic population, the G551D mutation is quite common.
Yet the CFTR genotype is usually not predictive of the clinical phenotype of this complex disease. Researchers have, however, separated the genetic mutations into different functional classifications. The most common classification scheme has five classes depending on the defect in the protein’s function (Table 1). Class I mutations result in a lack of protein production; mutation stop codons are examples of this class. Defective protein processing describes Class II mutations in which little or no CFTR protein function is present. Both Class I and II mutations are typically classified as “severe,” and affected individuals commonly have pancreatic insufficiency, as well as more progressive pulmonary problems. Class III, IV, and V are less severe due to the presence of some CFTR function, although still diminished in activity, and generally results in a less severe CF phenotype.
Classes of CFTR Gene Mutations
CFTR protein defect
Type of mutation
|CFTR function abolished|
|Defective processing and maturation
||CFTR function abolished|
||CFTR function abolished|
||Residual expression and function|
|Reduced function/ synthesis
|Residual expression and function|
|Adapted from: Castellani C, Cuppens H, Macek Jr M, et. al. Consensus on the use and interpretation of cystic fibrosis mutation analysis in clinical practice. J Cyst Fibros 2008;7:179–96.|
Genotype to Phenotype
Exactly how mutations in the CFTR gene result in the clinical phenotype remains hypothetical. In the so-called low-volume hypothesis, the CFTR protein regulates chloride transport by affecting multiple ion channels, including an epithelial sodium channel. The consequence of a non- or partially functioning CFTR protein is an increase in the activity of the sodium channel. This results in increased sodium transport across the epithelial membrane, which in this case is the airway lumen into the submucosa of the lungs. The imbalance leads to an increase in cellular water uptake because water follows the salt movement, and subsequent dehydration of the airway mucosal lining. As a result, lung secretions become more viscous and difficult to clear. Alternatively, the high-salt hypothesis proposes that excess sodium and chloride occur on the airway surface due to decreased CFTR functional protein. But overall, the exact pathogenesis of the disease is unknown.
Diagnosis of CF
In 1996, the Cystic Fibrosis Foundation published a consensus statement for diagnosis of CF. In the statement, experts recommended that diagnosis be based on identification of one or more clinical features, such as a history of a sibling with CF or a positive newborn screening test. In addition, the diagnostic guidelines included laboratory evidence of CFTR dysfunction identified by two abnormal sweat chloride concentrations obtained on separate days, the presence of two disease-causing mutations in CFTR, or an abnormal nasal potential difference measurement.
The Sweat Chloride Test
The sweat chloride test is considered the gold standard for diagnosing CF; however, it is not a perfect test, because approximately 1% of diagnosed CF patients have normal sweat chloride concentrations. Researchers have found that many of these patients possess uncommon genotypes. Furthermore, individuals who have other unusual conditions, such as eczema, anorexia nervosa, congenital adrenal hyperplasia, adrenal insufficiency, and Munchausen’s syndrome by proxy, can have false-positive results. Similarly, edema, hyponatremia, and CFTR mutations with preserved sweat-duct function produce false-negative results.
The Clinical Laboratory Standards Institute’s comprehensive guideline, C34-A3, “Sweat Testing: Sample Collection and Quantitative Chloride Analysis,” provides valuable information for laboratories that conduct sweat chloride testing. Table 2 presents the cutoff values used for diagnosing infants with CF. Although sweat chloride values are generally ≥60 mEq/L in infants with CF, lower values including concentrations <30 mEq/L can occur. For babies with intermediate or borderline results, the lab should repeat the sweat chloride test and the clinician should refer the baby to a center with expertise in the diagnosis of CF.
Today, labs are receiving an increasing number of requests for sweat chloride testing as a result of the implementation of mandatory state newborn screening programs for CF. Often the infants identified by such programs do not have the same clinical features exhibited by older children, increasing the importance of the sweat chloride test in diagnosing infants with CF.
Guidelines for Interpreting Sweat Chloride Results
Sweat Chloride (Cl–)
|Source: CLSI Guideline: C34-A3, “Sweat Testing: Sample Collection and Quantitative Chloride Analysis.”|
Newborn Screening for CF
Early detection of CF by newborn screening programs allows infants to live longer and healthier lives. In addition, these infants can be put on special diets to prevent nutritional deficiencies, resulting in improved growth and possibly cognitive function. Early diagnosis may also be helpful in obtaining genetic counseling and decreasing the chances of a long and drawn-out diagnosis.
In recent years, several groups have recommended screening for CF, including the American College of Medical Genetics (ACMG), the Centers for Disease Control and Prevention, and the U.S. Cystic Fibrosis Foundation. As of 2009, all 50 states and the District of Columbia now screen newborns for CF. Screening for CF is not limited to the U.S., however. Programs for newborn CF screening also exist in Australia, New Zealand, Canada, and many European countries.
Immunoreactive Trypsinogen in Screening
Screening newborns for CF relies on detecting elevated levels of immunoreactive trypsinogen (IRT), an enzyme made in the pancreas. Babies born with CF often have high levels of IRT in their blood, but other conditions can also cause the enzyme to be elevated. To understand how CF newborn screening works, an understanding of IRT and CF mutation frequency within specific populations is helpful.
In 1979, researchers first reported elevated levels of IRT in dried blood spots obtained from newborn infants with CF. High concentrations of IRT in newborns were suggestive of pancreatic injury, but not specific for CF disease. IRT concentrations decrease with age, however, such that a negative result is common at 8 weeks of age, limiting the utility of the IRT test to newborns.
The most common methodology for detecting IRT is a direct sandwich immunoassay using either colorimetry or fluorometry. The ancinar cells of the pancreas secrete two major isoforms of IRT, IRT1 (cationic trypsinogen) and IRT2 (anionic trypsinogen). Both are secreted in an inactive form, and they are activated to trypsinogen upon cleavage of a hexapeptide from the N-terminus.
Most manufacturers have not determined their assay’s specificity for the two IRT isoforms. Furthermore, IRT methods currently are not harmonized, so clinical validation for each method is necessary before implementing it in a screening program. A multiplex assay for each isoform also is available commercially; however, its role within screening laboratories remains unclear.
The IRT/IRT Algorithm: Pros and Cons
All methodologies for newborn screening for CF begin with determining the concentration of IRT. This first stage of testing requires a dried blood spot sample obtained during the newborn’s first week of life. High IRT concentrations, typically set in the 99th–99.5th percentile range, indicate an increased risk of CF and initiate a second round of testing for which an additional blood spot is collected 1–3 weeks later. Due to limited data, it is difficult to set the cutoff for the second IRT; however, labs typically use between the 95th and 97th percentile. Newborns who continue to show elevated IRT levels on the second tests are referred for sweat chloride testing.
This approach to newborn screening for CF has several disadvantages. It is sometimes difficult to collect a second dried blood spot because not all parents bring their infants back for the follow-up test. Furthermore, tracking specimens and accurately linking them to the newborn’s first specimen can present logistical complications. As mentioned previously, it is also difficult to set good cutoff values for the second IRT test. In some cases, the additional testing is triggered by false-positive results that have been associated with prenatal asphyxia and other stresses. On average, African-American infants have higher IRT levels compared to Caucasian infants, causing them to be referred to the second IRT testing more frequently. Finally, the test is not good at detecting heterozygous carriers of CTFR mutations who have decreased levels of IRT.
But there also are advantages to the IRT/IRT newborn testing algorithm. For example, it avoids the problems associated with mutation testing—detecting mutations of unknown clinical significance and performing a comprehensive screening for the large numbers of known CFTR genetic mutations. It may be possible to improve the specificity of this algorithm by using a scheme that tests a third specimen after a normal IRT level on the first test and an increased level on the second test. For this model to be successful, it is critical that the second and possibly third blood spot collections are properly tracked and identified.
The IRT/DNA Algorithm: Pros and Cons
While the IRT/IRT testing algorithm avoids the complexities of genetic testing for multiple CFTR mutations, more than 90% of newborn screening programs in the U.S. and Europe and all the programs in New Zealand and Australia use the IRT/DNA algorithm. This approach starts with IRT testing on the standard dried blood spot specimens collected from newborns. The algorithm first tests for IRT, and babies whose concentration exceeds the 96th –99th percentile cutoff have a second DNA test for CFTRmutations using material from the same blood spot.
With more than 1,800 mutations identified in this gene, selection of the CFTR mutations for the DNA test has been controversial, and many commercial kits are available for identifying panels of CFTR mutations. In the U.S., the ACMG has recommended a pan-ethnic panel for screening that consists of 23 CFTR mutations (Table 3), which includes roughly 80% of patients with CF. Clearly, selecting representative mutations is critical to establishing an acceptable screening specificity, and the test panel should consider the most prevalent mutations within the population being screened. For example, the most common delta F508 mutation is found in about 72% of the U.S. non-Hispanic Caucasian population; however, other ethnicities have lower percentages: Hispanic Caucasian, 54%; African American, 44%; Asian-American, 39%; and Ashkenazi Jewish, 31%.
American College of Medical Genetics Pan-ethnic Newborn Screening Panel for CFTR Mutations
This panel consists of 23 CFTR mutations that are common in most ethnic groups. The mutation nomenclature system for simple DNA lesions has now been adopted broadly by the medical genetics community.
For a further explanation, see Ogino S, Gulley ML, den Dunnen JT, Wilson RB. Standard mutation nomenclature in molecular diagnostics: practical and educational challenges. J Mol Diagn 2007, 9:1–6.
Amino acid abbreviations: A–alanine; C–cysteine; D–aspartate; E–glutamate; F–phenylalanine; G–glycine; H–histidine; I–isoleucine; K–lysine; N–asparagine; P–proline; R–arginine: T–threonine; W–tryptophan; and X – unspecified.
To improve the sensitivity of screening, the more prevalent mutations commonly found in the population being screened can be added to customize a CFTR mutation panel for the local population. As an example, California screens for 37 CFTR mutations due to the heterogeneity of its population.
The IRT/DNA algorithm typically identifies infants with one or two CFTR mutations. Infants with an elevated IRT level and a DNA mutation are also recommended for sweat chloride testing. Overall, the advantages of this approach include: improved sensitivity; elimination of false-positive IRT results sometimes found in premature, stressed, and African-American infants; and obtaining results from only one dried blood spot, thereby obviating the need for parents to bring their babies back for collection of a second blood spot sample.
Occasionally, this screening algorithm finds infants with very high IRT levels but no DNA mutations. In essence, such a finding acts as a safeguard to identify infants with rare mutations, and screening programs differ in how they handle this scenario. Some consider this a positive test and recommend follow up sweat chloride testing, while others recommend following the baby for clinical signs and symptoms.
Other Screening Models
Two variations of the IRT/DNA algorithm exist: the IRT/IRT/DNA and the IRT/DNA/IRT algorithms. In the IRT/IRT/DNA model, the cutoff chosen for the first IRT determination raises the specificity of the test, and a second sample is collected from newborns who test at the top 5–10% of values. Newborns who continue to have high IRT levels are then tested with the CFTR mutation panel, and those that have one DNA mutation are recommended for sweat chloride testing.
In the IRT/DNA/IRT model, newborns with high levels of IRT also have DNA mutation analysis. If the analysis detects one mutation, the infant is retested for IRT from a second dried blood spot sample. Those infants with high second IRT levels are referred for sweat chloride testing.
As with the other testing algorithms, these two methods have advantages and disadvantages. They reduce the number of sweat tests performed on infants, but they also require a second sample for collection and testing and reduce the number of carrier families that are detected. Similar to the IRT/DNA algorithm, the selection of CFTRmutations used for screening in this scheme will affect the number of CF cases detected in different ethnic populations.
Yet another variation of the IRT/DNA model only requires detection of a single CFTR mutation rather than two. To expand the analysis, the model uses a second genetic method, typically DNA sequencing of the gene. Unfortunately, large deletions and insertions may be missed by sequencing methods, so an additional, or fourth tier of testing, may be needed. The advantages of this model are twofold: it reduces the number of infants forwarded for sweat chloride testing, and it can identify mutations that may have variable expression.
Two more testing models are worth mentioning. One is the IRT/pancreatitis-associated protein (PAP) model used in France. This scheme uses only one sample, and preliminary data suggest that it is worthy of further investigation. The other model adds a third stage to the IRT/DNA method. Samples in which only one CFTR mutation is detected are tested by other DNA methodologies, such as gene scanning, to determine the presence of any additional mutations. Depending on the screening program, sweat chloride testing may be recommended for those infants with one or two detected mutations.
Screening for CF: More than Newborns
While much effort has been focused on early identification of newborns with CF disease, screening for CF is not limited to newborns. The American College of Obstetricians and Gynecologists (ACOG) recommends that all women, regardless of ethnicity or childbearing age, be offered CF carrier screening as part of routine obstetric care, preferably before conception. Women identified as carriers, and if possible their reproductive partners, may require additional testing, as well as genetic and reproductive counseling. ACOG also recommends CF screening for individuals with a family history of CF, reproductive partners of individuals with CF, and couples in which one or both individuals are Caucasian and/or European or of Ashkenazi Jewish descent. Screening in these populations would capture the 23 most common CFTR genetic mutations.
A Bright Future
The wealth of knowledge that exists today about the genetics of CF, along with technological advances in molecular testing, have had a significant impact on establishing newborn screening programs for the disease. Only a limited number of states screened for CF at the turn of the century—just 11 in 2002—but now CF newborn screening is a standard test in all 50 states and the District of Columbia, as well as many other countries.
Multiple screening algorithms are now available, each with advantages and disadvantages. The future of newborn screening for CF remains in flux, however, as it is a very involved testing process given the complexities of the CFTR gene. New genetic technologies are continuing to be developed every year, including inexpensive DNA sequencing technologies. Although it may be years away, it is easy to foresee complete DNA sequencing of the CFTR gene becoming cost-effective and integrated into CF screening programs, as well as screening for other inborn errors.
Castellani C, Cuppens H, Macek Jr M, et al. Consensus on the use and interpretation of cystic fibrosis mutation analysis in clinical practice. J Cyst Fibros 2008;7:179–96.
Crossley J, Elliott R, Smith P. Dried-blood spot screening for cystic fibrosis in the newborn. Lancet 1979;1(8114):472–4.
O’Sullivan BP, Freedman SD. Cystic Fibrosis. Lancet 2009;373:1891–1904.
Rosenstein BJ, Cutting GR. The diagnosis of cystic fibrosis: a consensus statement. J Pediatr 1998;132(4):589–95.
Ross LF. Newborn screening for cystic fibrosis: a lesson in public health disparities. J Pediatr 2008;153:308–13.
Sarles J, Berthezene P, Le Louarn D, et al. Combining immunoreactive trypsinogen and pancreatitis-associated protein assays, a method of newborn screening for cystic fibrosis that avoids DNA analysis. J Pediatr 2005;147:302–5.
Stanley F. Lo, PhD, is associate director clinical laboratories, technical director of clinical chemistry, point-of-care, and biochemical genetics, and director of the reference standards laboratory at the Children’s Hospital of Wisconsin, Milwaukee. Email:firstname.lastname@example.org