2004 2003 2002
The Journal of Pediatrics
143, Issue 6 , December 2003, Pages 707-712
Newborn screening for cystic fibrosis: ensuring more good than harm
Michael H. Farrell MD, Philip M. Farrell MD, PhD
|CF-2004-001 The Journal of Pediatrics Volume 143, Issue 6 , December 2003, Pages 707-712|
Newborn screening for cystic fibrosis: ensuring more good than harm
Michael H. Farrell MD,
Philip M. Farrell MD, PhD
From the Departments of Pediatrics and Internal Medicine, Yale University Medical School, New Haven, Connecticut, and the Department of Pediatrics, University of Wisconsin-Madison, Madison, Wisconsin, USA
|Abbreviations: CF, Cystic fibrosis; CFF, Cystic Fibrosis Foundation; CFTR, Cystic fibrosis transmembrane conductance regulator; IRT, Immunoreactive trypsinogen; Pa, Pseudomonas aeruginosa|
"We must work
to ensure that every child has an optimal chance for a healthy start in
Cystic fibrosis (CF) has many characteristic signs and symptoms, but establishing a timely diagnosis remains a major challenge.[1., 2., 3. and 4.] With conventional methods of identification, children with CF often have symptoms for months or years before definitive therapy is begun. [2. and 3.] About half of these patients have severe malnutrition or chronic lung disease at diagnosis. [4. and 5.] Because these delays have not decreased with decades of professional and public education, the CF community has long searched for methods to identify affected children earlier. [6., 7. and 8.] After favorable results from the Wisconsin CF Neonatal Screening Project [3. and 9.] and other studies, [10., 11. and 12.] Farrell [3.] asserted that "the burden of proof is on those who argue against CF neonatal screening." Since then, examination of outcome data has suggested that screening has the potential for both benefit and harm. Historically, political forces have often pushed for the rapid implementation of screening programs, sometimes resulting in troubling biomedical or psychosocial complications. Clearly, it would be best to avoid such problems during implementation of CF neonatal screening, especially given the likely appearance of many more screening technologies from the Human Genome Project. We have written this Commentary to inform programs that are now implementing or considering CF neonatal screening, especially those who are still unsure about the associated benefits and risks.
Cystic fibrosis is caused by mutations such as F508 in the CF transmembrane conductance regulator (CFTR) gene, leading to defective chloride channel functioning[13.] and the classic clinical triad of pancreatic insufficiency, chronic suppurative pulmonary disease, and salt loss in sweat. Gastrointestinal disease can develop in utero in the second trimester and lead to meconium ileus in ~20% of newborns with CF. [14.] Potentially lethal protein-energy malnutrition develops in some infants. [15.] By 4 years of age, ~85% of patients have some degree of malabsorption. [9., 14. and 15.] Regardless of nutritional status, excessive salt loss in sweat can be fatal for patients exposed to moderate heat or during prolonged hot weather.
The pulmonary aspects of CF vary widely in age of onset and rate of progression and are often the most difficult to treat.[16.] Newborns with CF have histologically normal respiratory systems, but within a few months the epithelial chloride channel defect may lead to abnormal respiratory secretions, bronchopulmonary infection, and airway obstruction. Most patients with CF develop chronic infections with unusual respiratory pathogens such as Staphylococcus aureus and Pseudomonas aeruginosa (Pa). The severity of lung disease generally determines prognosis, but other factors include sex, genotype and the impact of malnutrition and Pa infection.[17., 18., 19. and 20.] Although the median survival of patients in the United States seems to have plateaued at ~35 years, various projections suggest that those diagnosed early may have a lifespan exceeding 50 years. [20. and 21.]
The traditional method of CF diagnosis is the "sweat test," which uses pilocarpine iontophoresis to produce sweat for chloride analysis. Sweat chloride concentrations in patients with CF generally exceed 60 mEq/L, although values as low as 30 mEq/L are consistent with the diagnosis in children.[22.] Unfortunately, the sweat test is expensive [23.] and is only effective when the clinician has a high enough index of suspicion to order it. Delays in obtaining a sweat test are common. The average age of diagnosis in the United States is about 3 years of age, [1.] and many children with CF develop chronic sequelae by this time. [4. and 5.] These delays can be disparately greater for certain populations, including females, children with limited access to care because of economic or geographic factors, and in certain ethnic groups due to the myth that CF is solely a disease of Caucasians. [4.] Other reasons for delays include the relative rarity of new diagnoses of CF in individual practices and the fact that many sweat tests are typically ordered for every patient actually found to have CF. [23.] Thus, although sweat testing remains the "gold standard" method for diagnosis of CF, [24.] its use coupled with clinical judgment is suboptimal with regard to overall population health and the medical outcomes of many patients.
Recognizing the limitations of sweat testing and potential value of early diagnosis, Shwachman et al advocated for screening as early as 1970.[6.] Initial attempts to test meconium for evidence of CF were unsuccessful. In 1979, Crossly et al [25.] showed that dried newborn blood specimens could be tested for immunoreactive trypsinogen (IRT) levels, which are increased even in patients with intact pancreatic function. Initial findings from Australia [26.] and Colorado [27.] were promising, but a Cystic Fibrosis Foundation (CFF) consensus group recommended more research. [28.] Plans already underway for a randomized controlled trial of neonatal screening in Wisconsin were then finalized, with funding from the CFF and the National Institutes of Health. After discovery of the CFTR gene in 1989, Wisconsin and other regions implemented a second tier of testing using DNA (F508) analysis of specimens found positive on the IRT assay. [29. and 30.] The Wisconsin study has since become the largest prospective pediatric research project since the 1954 polio vaccine field trials. Modern CF screening methods utilize analysis of dried blood specimens collected at about 1 to 3 days of age using a heel stick and the "Guthrie card." [31.] Specimen cards are mailed to regional laboratories, most of which are organized and funded by government agencies such as state divisions of health. Unfortunately, screening program policies and procedures vary widely, especially in practices for follow-up and communication of results. [32.] In fact, the US programs have never had national policies, although the Centers for Disease Control and Prevention has recently assumed a role in promoting quality monitoring of laboratory procedures.
IRT analysis is the first step for all CF neonatal screening tests (Table I) and the sole method in some regions. A single IRT test, or the IRT/IRT combination, is 85% to 90% sensitive[2., 27. and 33.] and is associated with a relatively large number of false-positive results. With both methods, infants who have positive IRT screens are referred for a sweat test to determine which infants actually do have CF. Several programs now use DNA-based testing in place of a second IRT. [31.] Originally, F508 was the sole mutation for these tests, but some regions have begun implementing CFTR multimutational analysis. [31.] Specimens with the highest IRT results go on to DNA testing, although the specific "cutoff" value varies from the highest 1% to 5% of IRT results. [29., 30. and 31.] Even in cases when the DNA analysis is negative, a very high IRT is still considered suspicious for CF. In the Wisconsin program, [30.] infants with an IRT >180 ng/mL are "recommended for sweat testing." All of these methods are potentially less expensive than the traditional method of diagnosis. [2. and 23.] There are several advantages of two-tier IRT/DNA analysis over IRT testing alone: (1) affected infants are identified earlier, especially when a DNA-based diagnosis obviates the need for a sweat test; (2) it provides a better sensitivity, specificity, and positive predictive value than IRT alone [29.]; (3) the DNA step also reduces the number of false-positive results among African American infants or newborns with low APGAR scores [34.]; (4) the two-tier strategy avoids arranging for repeat specimen collection (as is done with IRT/IRT screening), saving time and money, as well as "lost" cases [27.]; and (5) DNA testing allows identification of some heterozygous infants whose families may benefit from genetic counseling.
Table I. Methods
used for CF neonatal screening*
Biomedical benefits and risks of CF neonatal screening
The first benefit observed in studies of neonatal screening was a more rapid referral to an accredited CF center, with high-quality respiratory, nutritional, and genetic counseling services.[20.] Although referrals in the Wisconsin trial were prompted by the study protocol itself, even the opportunity for early referral cannot be proactively assured without screening. In addition, the opportunity to participate in trials of early intervention cannot occur without early diagnosis.
The most consistently observed benefits have been nutritional. In the Wisconsin trial, infants diagnosed with CF through screening were more likely to have normal height and weight measures than those identified through traditional methods, after many experienced prolonged malnutrition.[3. and 9.] Similar findings were observed in Australia [10.] and France. [12.] Head circumference values were also significantly higher in the Wisconsin trial screened group compared with the standard diagnosis group. [9.] Since these observations, differences in growth and nutritional status have been observed in children as old as 10 to 15 years. [3. and 9.] In addition, screening allows prevention of prolonged deficiencies of micronutrients such as vitamin E and the resulting complications, such as hemolytic anemia [35.] and abnormalities in cognitive function. [36.] The only potential risk of screening related to nutritional care is intestinal stricture, which has been rare in young children, but may occur when excessive dosages of pancreatic microsphere enzymes are consumed. [37.]
The pulmonary benefits of CF neonatal screening, anticipated intuitively for many years,[6.] have been more challenging to establish because of the variable pattern and age of onset of lung disease, the difficulty in quantification of pediatric lung disease, and variations in genotypic and environmental influences. Observational studies from Australia, [10.] the Netherlands, [11.] and France [12.] suggest better outcomes in screened groups compared with various CF patients diagnosed conventionally. Unfortunately, these studies could have been affected by selection bias and lack of systematic or controlled respiratory care. The randomized design of the Wisconsin trial avoided these problems but encountered confounders and unexpectedly mild pulmonary dysfunction in both the screened and standard diagnosis (control) groups. [38. and 39.] Quantitative scoring of chest radiographs in Wisconsin revealed a mixed picture; the scores were significantly better at the time of diagnosis in the screened group, but became similar and then worse once Pa infections developed in the screened patients.[39.] However, chest radiographic observations at diagnosis suggested a potential benefit; half of the control group patients diagnosed at a mean age of 108 weeks had radiographically irreversible lung disease compared with 28% of those diagnosed through screening, ie, half of those with delayed diagnosis had already progressed past the "point of no return". [39.]
The risks of presymptomatic diagnosis relative to the respiratory system have only been investigated to a limited extent. Most concerns are related to overly aggressive or inappropriate treatment with antibiotics, aerosol inhalations, and chest physiotherapy (which is an unnecessary burden in patients without lung disease). In prospective studies, premature acquisition of Pa seems to be the only potential harm; however, Pa infections can have a greater adverse impact on patient outcomes than the effect of neonatal screening itself.[39. and 41.] Therefore, methods to delay Pa acquisition are important. Premature Pa acquisition was originally hypothesized to result from repeated use of oral antibiotics in children perceived by clinicians to be more vulnerable to upper and/or lower respiratory infections because of CF.[2.] In the Wisconsin trial, outcome data suggest that the risk for early acquisition of Pa was largely attributable to attendance at a clinic where screened infants were exposed in the waiting room to older Pa-infected patients,[40. and 41.] but earlier introduction of oral antibiotics [18. and 19.] and aerosol exposure [19.] were also potential risk factors. Thus, it seems that this pulmonary risk is not related to screening per se, but is the secondary, potentially avoidable consequence of environmental exposures associated with healthcare services. This is consistent with recent experience in South Australia, which demonstrated increased mortality for children with CF diagnosed as newborns and exposed to Pa acquisition/infection through environmental cross-infection circumstances, ie, person-to-person transmission.[42.] These observations have added to the controversy on cohorting (segregating) Pa-infected patients.[43.]
Psychosocial benefits and risks of CF neonatal screening
Any consideration of the good and harm of screening must examine the psychological and social dimensions, as well as biomedical outcomes. Speculation about the psychosocial effects of CF neonatal screening preceded the development of a practical test.[44.] Apart from a few recent studies, [45., 46. and 47.] much of the concern over psychosocial adversity has been based on experience with other genetic conditions such as Tay-Sach's disease and sickle cell hemoglobinopathy. [48.] The major psychosocial benefit of early diagnosis through screening is reduction of the anxiety often experienced by families as they endure tests, clinic visits, or hospitalizations with little idea of what is wrong with their baby or what the future will hold. [49.] Universal neonatal screening may reduce possible disparities in anxiety and uncertainty by eliminating diagnostic delays due to the infant's sex, ethnicity, and limits in access to care ( Table II).
Adverse psychosocial effects are more controversial for the families of heterozygous infants who gain no apparent benefit from screening.[44., 45. and 46.] When IRT/DNA screening identifies a single CFTR mutation, some parents may develop anxiety over the uncertainty of diagnostic outcome or while waiting for the sweat test results. [43., 44., 45. and 46.] When a negative sweat test confirms heterozygous status, a sense of relief may develop, but some parents may also continue with a grieving process, mourning the loss of the perfect child. Anxiety and grief reactions associated with carrier-state diagnoses are thought to place families at risk for impaired parent-child bonding, disrupted relationships, personality problems, and the development of psychogenic symptoms or some variant of the vulnerable child syndrome. [50. and 51.]
Other psychosocial risks and benefits are related to genetic information and the parents' awareness of their own CFTR mutations. Understanding this information through genetic counseling provides a beneficial opportunity for informed reproduction. However, there are risks as well as benefits associated with genetic counseling. Although some parents are able to make use of such resources to help inform subsequent reproductive decisions, others can remain confused, and it is unclear to what extent reproductive decisions are influenced by the genetic information.[46. and 52.] In addition, unaffected family members have been reported to experience survivor guilt over their "escape" from a genetic disease or the status of their infant relative. [53.] Genetic screening may also lead to the identification of mispaternity (potentially estranging the father) or to stigmatization of the parents, or the child. With other genetic diseases, the misconception that the heterozygote child is at risk of developing the disorder has sometimes led to difficulties with medical or life insurance, employment discrimination, and even community ostracization or devaluing the child as a potential marriage partner. Most historical examples of this were related to misunderstandings, and therefore should be preventable through high-quality efforts at effective counseling and public education. [46. and 53.] Finally, a false sense of reassurance may develop after a false-negative test result, which usually occurs when an infant's CFTR mutation is not included in the standard screening panel, or in rare cases due to human errors in laboratory procedures, record-keeping, or communication.
Excellent implementation is the key to ensuring more good than harm
The results of the Wisconsin randomized trial[3. and 9.] and other studies [10., 11. and 12.] satisfy the "more good than harm" criteria of Cadman et al [8.] and the World Health Organization guidelines. [54.] Beneficial outcomes, therefore, can be expected after neonatal screening for CF. Patient outcomes, however, depend as much on actions after screening as on the screening program itself. Screening identifies the opportunity to achieve good results, but does not automatically lead to more good than harm.[8.] Two opportunities to ensure benefit are optimized communication and the delivery of improved medical care. After the screening result becomes available, the first major opportunity is the provision of effective communication regarding the test result and its implications. Effective communication is consistent with the ethical principle of autonomy and practices of informed consent. [55.] Good communication is also known to increase patient satisfaction, [56.] improve compliance, [57.] and to enhance health outcomes for patients with chronic diseases. [58., 59. and 60.] Specific to CF, effective communication is necessary for full understanding of the medical information and the implications of the genetic data. [31. and 46.]
Unfortunately, various studies have raised concern about the poor communication skills of physicians.[32. and 61.] In addition, many clinicians have a limited grasp of genetics [62.] and are perceived as deficient by both families [46.] and public health officials. [32.] Studies have also identified a fundamental difficulty clinicians have in achieving nondirective counseling. [7., 31. and 63.] Although state screening programs could conceivably assume responsibility for consistent communication of results and surveillance for psychosocial complications, this would be a major change from the current status characterized by great variability and no apparent best practices. [32.] Physicians can refer family members to a certified genetic counselor, but there are not enough counselors to provide services for all the families who need it. In 1990, the estimated total of 450 certified genetic counselors was grossly inadequate to provide counseling to all families with infants heterozygous for CF. [53.] The approximate four-fold increase of genetic counselors in the past decade [31.] still leaves a significant imbalance between counseling supply and demand. Thus, primary care physicians will continue to be responsible for the bulk of genetic counseling. Consequently, they will need to become more knowledgeable and skilled about both genetic risk information and its communication. We have reviewed the appropriate content and conduct of such communication elsewhere. [31.] This important role for primary care physicians will not be unique to CF screening.
The second prerequisite to achieving optimal outcomes after neonatal screening is a follow-up system to assure that medical care consistently reflects the best available scientific evidence and standards for care. Fortunately, in the United States and most western European countries, there are networks of CF centers with dedicated teams focusing intensive efforts on clinical management of individual patients. However, the outcomes of these centers are quite variable and few are skilled in managing the presymptomatic but very challenging CF patients emerging from neonatal screening programs.[3.] Therefore, quality improvement is needed; this is not a new concept in pediatric care, but systems-based interventions aimed at prevention have not been widely used in children with chronic diseases. Although CF shares little physiologically with many chronic diseases in adults, establishing a national consensus for standards of care could best be guided by quality improvement successes with other chronic diseases. In adult chronic heart failure, for example, establishment of standards of care has benefitted greatly from the recent creation and refinement of two different stages-of-disease models.[64.] Adapted to CF, such an approach would categorize patients into discrete groups based on the standards of care applicable to each group with a special, prevention-oriented stage for presymptomatic and mildly affected children identified by neonatal screening.
Since the characterization of CF as a distinct disorder in 1936, the standard of treatment has evolved from palliative care and "tune-ups" to identification before symptoms arise. With the future advances anticipated in molecular therapy and prevention, the CFF has recognized the importance of an integrated process for diagnosis, primary care, and multidisciplinary care services.[24.]
In our judgment, the potential benefits for CF screening are compelling. The proportion of patients dying with undiagnosed CF has been estimated at 5%[65. and 66.]; hyponatremic dehydration, severe protein-energy malnutrition, and acute respiratory failure can all be fatal, particularly in infants with CF. [66. and 67.] In contrast, expanding evidence points to reductions in mortality, improved nutrition, informed reproductive decisions, and creation of the opportunity for early pulmonary intervention before irreversible lung disease develops. As improved respiratory therapies emerge for young children, such as inhalation of tobramycin, [65.] the opportunity for pulmonary benefits will become even more attractive.
Recognizing that complex improvements in health care are as challenging as they are important, Farrell[68.] has described "ten steps to success in implementing a newborn screening program." Fortunately, many regions of North America and Europe have been implementing CF neonatal screening in a carefully planned fashion. [69.] These efforts are encouraging and will help avoid some of the problems that developed with phenylketonuria and sickle cell disease screening. [31.]
In response to findings of newborn screening follow-up variation,[32.] Holtzman asked "Is public health ready for genetics?" [70.] The answer may vary depending on the systems in place to manage both genetic information generated through screening and the medical care that follows. Now is the time for more effective leadership to achieve the changes that are sorely needed to improve the outlook for children with CF to "ensure that every child has an optimal chance for a healthy start in life," and to ensure more good than harm through consistently early diagnosis using IRT/DNA screening coupled to optimal follow-up care for all.
Corresponding author. Reprint requests: Philip M. Farrell, MD, PhD, Professor of Pediatrics and Dean, University of Wisconsin Medical School, Rm 1217 MSC, 1300 University Ave, Madison, WI 53706.
CF-2002-001 BMC Clin Pathol. 2002 Nov 19; 2(1): 4.
Survey of CF mutations in the clinical laboratory
Huber K, Mirkovic B, Nersesian R, Myers A, Saiki R, Bauer K.
|CF-2002-001 BMC Clin Pathol. 2002 Nov 19; 2(1): 4.|
|Survey of CF mutations in the clinical laboratory|
Huber K, Mirkovic B, Nersesian R, Myers A, Saiki R, Bauer K.
Ludwig Boltzmann Institute for moleculargenetic laboratory diagnostics, Donauspital, Vienna, Austria. firstname.lastname@example.org
|BACKGROUND: Since it is impossible to sequence the complete CFTR gene routinely,
clinical laboratories must rely on test systems that screen for a panel of the
most frequent mutations causing disease in a high percentage of patients. Thus,
in a cohort of 257 persons that were referred to our laboratory for analysis of
CF gene mutations, reverse line probe assays for the most common CF mutations
were performed. These techniques were evaluated as routine first-line analyses
of the CFTR gene status. METHODS: DNA from whole blood specimens was extracted
and subjected to PCR amplification of 9 exons and 6 introns of the CFTR gene.
The resulting amplicons were hybridised to probes for CF mutations and
polymorphisms, immobilised on membranes supplied by Roche Molecular Systems,
Inc. and Innogenetics, Inc. Denaturing gradient gel electrophoresis and
sequencing of suspicious fragments indicating mutations were done with CF exon
and intron specific primers. RESULTS: Of the 257 persons tested over the last
three years (referrals based on 1) clinical symptoms typical for/indicative of
CF, 2) indication for in vitro fertilisation, and 3) gene status determination
because of anticipated parenthood and partners or relatives affected by CF), the
reverse line blots detected heterozygote or homozygote mutations in the CFTR
gene in 68 persons (26%). Eighty-three percent of those affected were
heterozygous (47 persons) or homozygous (10 persons) for the DeltaF508 allele.
The only other CF-alleles that we found with these tests were the G542X allele
(3 persons), the G551D allele (3 persons), the 3849+10kb C-T allele (2 persons)
the R117H allele (2 persons) and the 621+1G-T allele (1 person).Of the fifteen
IVS8-5T-polymorphisms detected in intron 8, seven (47%) were found in males
referred to us from IVF clinics. These seven 5T-alleles were all coupled with a
heterozygous DeltaF508 allele, they make up 35% of the males with fertility
problems (20 men) referred to us.
CONCLUSIONS: In summary, the frequency of CF chromosomes in the cohort examined with these tests was 26%, with the DeltaF508 allele affecting 83% of the CF chromosomes. It is a substantial improvement for routine CF diagnostics to have available a test system for 30 mutations plus the polypyrimidine length variants in intron 8. Our results show that this test system allows a routine first-line analyses of the CFTR gene status.