A paediatric respiratory specialist reviews the research to demonstrate that screening babies for cystic is cost-effective in certain populations, if systems are put in place for processing positive results and dealing with families sensitively
Kevin W Southern
Reader and honorary consultant in paediatric respiratory medicine
Institute of Child Health
Alder Hey Children’s Hospital
Cystic Fibrosis (CF) is the most common life-shortening inherited condition in the population derived from northern Europe. The incidence of CF in newborns from this population makes a screening protocol feasible and two randomised studies have investigated the impact of this intervention on nutritional and respiratory outcomes for children with CF.
The results from these studies (reviewed in detail in previous work) provide some evidence to support newborn screening (NBS) but this is not overwhelming.[2,3] In an article reviewing the ‘human rights’ of screening newborns for CF, Phil Farrell concluded that screening was indicated in a population with a high incidence of the condition if healthcare resources could provide appropriate levels of care after the diagnosis.
All CF newborn screening programmes use measurement of immuno-reactive trypsinogen (IRT) in the first week of life as the initial screening test. In the 1970s, Crossley and her colleagues in New Zealand identified that IRT was significantly raised in babies with CF. This seminal work stimulated a number of groups to establish newborn screening protocols. The advantage of the IRT measurement is that it can be undertaken on a blood spot from the standard newborn screening card.
A problem with measuring IRT in the first week of life is that it may be raised in infants born pre-term and in infants that are unwell. Also, data suggest that IRT may be higher in infants from a non-Caucasian origin and infants with chromosomal disorders. These factors need to be taken into account when designing a NBS protocol.
IRT measurement in the first week of life (IRT-1) is an extremely sensitive test to detect infants with CF. However, it does not have the necessary specificity to exist as a stand-alone screening test and therefore a second tier of testing is required. Prior to recognition of the CF gene defect, this was generally a repeat measurement of IRT after three or four weeks of age, at which age a raised IRT is more specific for CF.
Many protocols have been based on this IRT-IRT strategy. However, after identification of the cystic fibrosis transmembrane conductance regulator (CFTR) gene in 1989 and the recognition that phe508del is by far the most common CF-causing mutation, an increasing number of protocols have moved to using DNA analysis as the second-tier test. The advantage of this is that the analysis can be undertaken on the same blood spot as IRT-1 and that infants who are homozygous for phe508del are recognised earlier (in the second or third week of life).
Second-tier DNA analysis
In a predominately northern European population, most patients with CF will be homozygous or compound heterozygous for phe508del. Therefore, simply examining for phe508del at this stage will recognise most patients and this strategy was initiated by a number of regions.
However, many other CF-causing mutations have been recognised, all much less common than phe508del; therefore, it is difficult to exclude CF with confidence with a protocol based on DNA analysis. To counter this, protocols have been organised to include sweat testing for infants with one phe508del mutation. Increasing the number of CFTR mutations on the screening panel does diminish this problem, but not significantly. Most infants with one mutation will have a low sweat chloride (or conductivity) and will be termed a healthy carrier.
CARRIER RECOGNITION: Disadvantages
The recognition of carrier status affords the opportunity for parents to undertake cascade screening – screening family members for the putative gene defect. Cascade screening can give parents and extended family members the opportunity to make informed decisions with respect to future pregnancies.
However, studies suggest that, on the whole, the general public does not regard carrier recognition positively and issues remain around informing the index case later in life. Studies also suggest that the period of assessment (sweat test) following a positive screening result is intensely stressful for families. Another potential disadvantage of DNA analysis is the recognition of non-paternity, if the infant is recognised to have a mutation that is subsequently not identified in either parent.
The most important strategy for reducing carrier recognition is good management of the IRT assay, taking into account factors such as setting appropriate cut-offs and assessing poor-quality samples. Such experience in undertaking the assay should result in the recall of fewer babies for repeat testing or clinical assessment. This is particularly important, as studies have demonstrated that carriers have higher IRT levels than non-carriers, therefore small margins may increase carrier recognition disproportionately.
Other strategies include reducing the number of CFTR mutations in the initial screening panel, on the assumption that, after a certain point, increasing the number of mutations in the panel does not result in any further increase in case recognition, but will result in a small increase in carrier recognition. In the UK, in an attempt to reduce the stress associated with clinical assessment and sweat testing, a second blood spot is taken at 21-28 days for a repeat IRT measurement on infants with one mutation recognised. Only infants with a raised second IRT are invited for sweat test and clinical assessment; for the other infants, parents receive information on carrier status.
Designing a NBS protocol for CF
A protocol to screen newborn infants must reflect the target population. There are populations (for example, in Africa and south-east Asia) for whom it is inappropriate to screen for CF, as the balance between case recognition and the impact of medicalising families unnecessarily is not sufficiently positive.
For protocols using DNA analysis, it is generally considered that the initial CFTR panel should provide in the region of >90% coverage of the CF population for that population. However, all programmes should reflect the ethnic diversity of the population. Some programmes have incorporated a second IRT measurement on infants with a very high IRT-1 but no mutation recognised on DNA testing, in order to provide a ‘safety net’. However, there has been concern that this may not be a valid strategy, as infants with uncommon CF mutations may have a less classic phenotype and possibly equivocal IRT levels.
Another strategy is to undertake extensive DNA testing (even sequencing) in order to significantly increase the number of CFTR mutations examined for. The potential disadvantage of this strategy is the recognition of novel CFTR sequence changes, the molecular (and phenotypic) relevance of which is less clear.
Processing a positive result
One of the biggest challenges that face a NBS programme is the efficient and timely processing of a positive result. This is a considerably stressful time for families. Planning is required as to the most appropriate interface between the screening laboratories, the CF centres and the families. It is essential that periods of uncertainty are kept to a minimum and that the parents/carers receive clear and honest information.
This is dependent to some degree on the health service and CF networks established in the screening region. However, some basic principles should be adhered to:
• It is better to have organised appropriate assessment prior to family contact – a positive CF screen is not a medical emergency in contrast to the recognition of some metabolic conditions such as phenylketonuria.
• Communication between screening labs, primary care health workers and CF teams must be robust.
• Feedback must ensure that all results are
An unfortunate side-effect of the very good sensitivity of the IRT assay to detect infants with CF in the first year of life is the recognition of infants with an equivocal diagnosis of CF, such as infants with persistently equivocal sweat chloride values or two CFTR mutations, the consequence of which is unclear. Consensus statements have advocated a pragmatic approach to these infants with regular, albeit infrequent review, and careful monitoring for clinical signs consistent with CF.[9,10] In addition, there may be a role for additional physiological tests looking for abnormal transepithelial salt transport across the nasal or intestinal epithelia.
Impact on a health service
As with all NBS tests, implementing a CF protocol relies on effective primary care resources around time of birth. Clear information is required for parents and primary care health workers must have a good understanding of the screening process.
Measurement of IRT can be undertaken on a blood spot sample and therefore it is possible to integrate NBS for CF into well-established NBS programmes for phenylketonuria and congenital hypothyroidism. However, the IRT assay requires a very good quality of sample and so some training is required to ensure that an adequate percentage of samples are appropriate for analysis. If this is not achieved, then it is a considerable resource and logistics issue to obtain repeat samples – not to mention the distress this causes parents.
Most screening laboratories cope well incorporating the IRT assay into their schedule, but there are challenges, most notably ensuring adequate quality assurance (QA) and data recording.
Significant extra resources are required to run a CF programme. There is no particularly robust QA mechanism and laboratories need to rely on careful interpretation of their data. For protocols that use DNA analysis as a second tier of testing, a regular number of samples (those with a high IRT) will be sent for further gene analysis. This is generally a separate laboratory, often on a different site. Clear lines of communication between the screening and molecular laboratories are essential. The costs of DNA analysis are dependent on the protocol implemented. In some cases – in which gene sequencing is employed – these costs are significant.
Finally, a CF screening programme requires robust pathways to ensure an appropriate interface between the result and the family. In many countries, this will utilise primary care workers and there are issues of training that need to be addressed. In contrast to other conditions that are screened for in the newborn period, the relatively high incidence of CF means that a positive result is not uncommon and the impact of this programme on healthcare workers is considerable.
Although no studies have examined quality of adjusted life years from this intervention, there are data to support the impression that screening newborns for CF is cost-effective. In an evenly matched cohort of children with CF, costs of care were considerably reduced for those screened compared to children who had been diagnosed clinically. This is consistent with other data from this and other databases suggesting that while screened children may have similar condition to non-screened, this is at the expense of considerably less treatment. From the data available, a screening programme will generate a healthcare saving simply with respect to treatment costs.
There is agreement that screening newborns for CF is a valid strategy in a population that has a significant incidence of the condition: a high percentage of northern European lineage. However, the weight of evidence to support newborn screening is not as great as for other conditions such as phenylketonuria and it is imperative that the infrastructure is in place to care appropriately for infants diagnosed in this manner. It is also important that NBS protocols are designed to reflect local populations and healthcare resources. Robust systems must be in place to minimise the potential negative impact of a NBS programme.
1. Farrell PM et al. Pediatrics 2001;107:1-13.
2. Castellani C et al. J Cyst Fibros 2009;8:153-73.
3. Southern KW et al. Cochrane database of systematic reviews (Online). 2009(1):CD001402.
4. Farrell PM. J Cyst Fibros 2008;
5. Comeau AM et al. Pediatrics 2007;119:e495-518.
6. Southern KW et al. J Cyst Fibros 2007;6:57-65.
7. Bobadilla JL et al. Adv Pediatr 2002;49:131-90.
8. Southern KW. Journal of the Royal Society of Medicine 2004;97 Suppl 44:57-9.
9. Mayell SJ et al. J Cyst Fibros 2009;8:71-8.
10. Farrell PM et al. J Pediatr 2008;153(2):S4-S14.
11. van den Akker-van Marle ME et al. Pediatrics 2006;118:896-905.
12. Sims EJ et al. Lancet 2007;
13. Sims EJ et al. J Pediatr 2005;147(3):306-11.