Editor—Mutations of the cystic fibrosis transmembrane conductance regulator (CFTR) gene have an extremely wide phenotypic spectrum.1 The “classical” severe form of cystic fibrosis (CF) is characterised by pancreatic insufficiency and chronic endobronchial infection.2 Milder forms may show pancreatic sufficiency, though the degree of pulmonary involvement varies.3 4 Several other pathologies have now been linked with mutations in the CFTR gene, including congenital bilateral absence of the vas deferens, liver disease, pancreatitis, and disseminated bronchiectasis.5-7

There are now over 850 documented mutations in theCFTR gene.8 The most common is ΔF508, which appears on 66% of CF chromosomes in western Europe.9 The pancreatic status of patients has a strong correlation to the genotype, whereas the severity of lung disease shows little or no relation to genotype.1 10-12 Generally, mutations that result in no CFTR protein, such as truncating mutations, or those, such as ΔF508, which result in mislocalisation of the protein, result in a severe phenotype with pancreatic insufficiency. Missense mutations, particularly in the transmembrane domains, result in a milder, more variable disease.1 13 14 The sweat test, long regarded as the gold standard diagnostic test for cystic fibrosis, may give normal results in these milder forms.

Neonatal screening for CF relies on an increased immunoreactive trypsinogen (IRT) concentration in blood of affected babies during the first two months of life.15 However, this method has low specificity, particularly with samples taken in the first week of life, so mutation analysis is increasingly being used as a second tier. In the Trent region of the United Kingdom (UK), where ΔF508 accounts for over 80% of CF mutations, we currently use a three stage IRT-DNA-IRT protocol, which has a low requirement for second blood samples and for sweat testing.15 Any initial blood sample with IRT readings above the threshold is analysed for the ΔF508 CFTR mutation. Subjects who are heterozygous for ΔF508 are resampled at 27 days; if the IRT level is again above the threshold the child is referred for sweat testing. The families of heterozygous neonates with a second IRT level below the threshold are referred for genetic counselling on the assumption that such neonates are unaffected carriers.

It has previously been shown that neonates with transient hypertrypsinaemia carry the ΔF508 mutation at a higher frequency than the general population.16 Additionally, IRT remains on average higher in the repeat blood sample in babies who are heterozygous for ΔF508 than in non-ΔF508 babies.15Thus, it is possible that a proportion of infants with transient hypertrypsinaemia and ΔF508 may have a second (mild) CF mutation. We have examined this hypothesis in a cohort of infants assessed through the Trent (UK) neonatal screening programme.

A cohort of 88 ΔF508 heterozygous neonates with transient hypertrypsinaemia born between April 1991 and November 1996 were available for testing.15 All had exceeded the IRT cut off value for the 6th day blood sample (usually 90 ng/ml but adjusted periodically to select approximately 0.5% of tested babies) and had an IRT concentration below 80 ng/ml in the 27th day sample. Cases with meconium ileus were excluded. There was a sample available from every subject fulfilling the selection criteria for the trial period. All samples were irreversibly anonymised but the IRT data were retained and the subjects in the cohort were known. In addition, three transiently hypertrypsinaemic babies who would otherwise have qualified for the cohort had already been identified as compound heterozygotes for ΔF508/R117H CF mutations through extended mutation analysis of their parents.

DNA for PCR was obtained from the blood spot by elution in 100 μl of 50 mmol/l NaOH for 30 minutes at room temperature. The samples were then boiled for five minutes, cooled, and neutralised with 15 μl of 1 mol/l Tris-HCl (pH 8.0).

Primers used for amplification comprised both published18(nomenclature as reference) and unpublished sets. PCR was performed in 25 μl reactions under standard conditions. The IVS8-nT tract was amplified and identified as described elsewhere.19 Exon 9 was amplified using the IVS8-nT protocol and the primer sets 9 (1) followed by 9 (2). This excluded the two highly polymorphic regions before the exon (TG(n) + IVS8-nT) from amplification and greatly simplified mutation analysis. Full details of primer sequences and annealing temperatures can be found in table 1.

Amplified fragments were screened for sequence alterations by single stranded conformational polymorphism analysis (SSCP)/heteroduplex analysis (HA) using standard conditions and procedures as in table 1with 14% v/v (59:1 acrylamide:bisacrylamide) polyacrylamide gels and electrophoresis at 12°C in 0.5 × TBE except exon 11, which used a 14% (37:1 acryl:bis) polyacrylamide gel. Intron 19 used non-denaturing electrophoresis on 8% (29:1 acryl:bis) polyacrylamide gels in 1 × TBE. Any samples displaying shifts (indicative of a sequence alteration) were investigated further by direct DNA sequencing (as per manufacturer's instructions). This methodology identifies >95% of all mutations in our laboratory.

Second mutations were identified in 20 subjects (table 2) giving a compound heterozygote frequency of 22%. R117H was the most common second mutation found, constituting 45% of the compound heterozygotes identified. Fig 1 shows the range of second IRT reading displayed by non-ΔF508 babies, the putative ΔF508 heterozygotes, and the compound heterozygotes. The IRT distributions of the putative ΔF508 heterozygotes (19) and particularly the compound heterozygotes are shifted to the right. Forty one percent of ΔF508 heterozygote neonates with >25 ng IRT/ml in the 27th day blood sample possessed a second mutation compared to ∼6% for those with <25 ng/ml, an enrichment factor of 2.35.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Distribution of immunoreactive trypsin concentrations in 27th day blood spots. The data from the non-ΔF508 sample have been normalised to the number of ΔF508/N samples.

Recent work has suggested that “polyvariant mutantCFTR genes” may, when combined, result in less functional or pathologically insufficient CFTR.21 In the light of this, we determined the incidence of the intronic poly-T tract, IVS8-nT, which interacts with the R117H mutation. The IVS8-5T allele was present in three of the 91 neonates, none of whom were identified as being compound heterozygotes for a second CF mutation. The 71 remaining subjects are likely to be true heterozygotes with transient hypertrypsinaemia arising from causes other than CF.

In this study, we have confirmed that the combination of heterozygosity for ΔF508 and persisisting mild hypertrypsinaemia in the newborn period carries a substantial risk of that infant having a second CF mutation. In addition, the concentration of IRT in whole blood at 27 days is a biochemical marker to refine this risk further.

All the mutations found in our cohort have been reported previously and for the more common ones there are data on phenotypic presentation. In general, the severity of lung disease is less predictable than the degree of pancreatic involvement.22

The R117H has been associated with a range of phenotypes and the intronic variant polyT tract (IVS8-nT) is known to affect expression. The IVS8-5T variant is an inefficient splice acceptor site and with R117H, in trans with ΔF508, results in a pancreatic sufficient phenotype. The 7T variant and R117H combination in ΔF508 heterozygotes is associated with congenital bilateral absence of the vas deferens and, in some cases, mild lung disease.19 23Some females with this genotype are completely asymptomatic. Three ΔF508:R117H compound heterozygotes with the 7T variant from the cohort have already presented with symptoms of lung disease. An increased frequency of IVS8-5T in non-CF transient neonatal hypertrypsinaemia, with or without heterozygosity for ΔF508, has been reported in two studies.23 24 However, in our cohort, IVS8-5T was not present in any of the compound heterozygotes detected and its overall frequency was not significantly different from that reported in the general population.

Five other neonates screened had missense mutations in transmembrane domains. These have been reported in patients with presenting phenotypes ranging from “cystic fibrosis” to oligospermia, but there have been too few cases described to define genotype-phenotype correlations. The missense mutation F693L is located in the regulatory domain of CFTR and was first identified in a young girl with ΔF508 on her other allele who was diagnosed with pancreatic insufficient cystic fibrosis.26

Three subjects had nonsense mutations which are normally associated with severe disease as they introduce a stop codon, leading to truncated, usually inactive, CFTR protein being transcribed.

The 3849+10 kb C→T mutation, identified in two subjects in our cohort, activates a partially active, cryptic splice site within the intron that causes an 84 bp “exon” to be inserted, in frame, within exon 19.27 Homozygotes for this mutation usually have relatively mild lung disease, while compound heterozygotes may show pancreatic insufficiency and may have more severe lung disease.

Compared to other neonatal metabolic screening tests, that for CF has a relatively low sensitivity; the more successful programmes detect approximately 90% of severe, “classical” CF.28 With the increasing recognition of milder forms of CFTR deficiency, and the tendency to use DNA analysis rather than the sweat test as the ultimate diagnostic arbiter, the perceived effectiveness of neonatal screening is being further eroded. Nevertheless, provided that the prime aim of the programme remains the detection of classical cystic fibrosis, in the way the term was understood 10 years ago, performance using either a two stage IRT method or one of the newer IRT-DNA protocols may be regarded as acceptable.

The difficulties with DNA based protocols arise from the need to provide explicit counselling to families where the baby has been found to be heterozygous for ΔF508 but has given a normal sweat test or had a normal IRT concentration in a second blood sample. In most screening programmes more babies will fall into this group than will be diagnosed with classical cystic fibrosis (representative figures are given in reference 15). The present study shows that such babies, particularly the subgroup with IRT in the second blood sample >25 ng/ml, have a high likelihood of a second CFTR mutation. Should further testing of ΔF508 heterozygotes be restricted to mutations associated with a severe phenotype or should extended mutation analysis, at least in the higher risk group with second IRT >25 ng/ml, become part of routine investigation? Would such investigation improve the accuracy of counselling of such families?

The majority of compound heterozygotes detected in this study have mutations usually associated with milder forms of CFTR disease. However, some of our hypertrypsinaemic cohort have genotypes previously reported in severe forms of CF and may be regarded as having given false negative screening test results. Genotype-phenotype correlation is complicated by the existence of polyvariantCFTR genes and modifier polymorphisms,29 as well as possible modifiers at other loci. It is probably unsound to extrapolate data from clinically selected cases to a cohort which was selected on an entirely different basis. The clinical significance of compound heterozygosity detected through neonatal screening will only become apparent through systematic follow up. Preliminary indications are that some of the expectedly mild cases develop significant early symptoms; six of the compound heterozygotes in our cohort (indicated in table 2) were unblinded by being diagnosed with cystic fibrosis, enabling them to be traced forwards from their neonatal screening results. All have developed symptomatic respiratory infections associated with positive bacterial cultures. A prospective study is under way to determine the clinical spectrum present in such compound heterozygotes.

It is generally agreed that diagnosing adult onset disease during childhood is undesirable if there is no effective preventative treatment. One would not normally wish to predict fertility problems in an otherwise healthy male neonate. However, in so far as early active treatment can reduce the impact of lung disease in moderate CFTR deficiency, previous warning that a child is thus predisposed would be valuable. Thus, the increased possibility of mild CF spectrum disease should be raised during counselling for ΔF508 heterozygosity detected on newborn screening. There may be a case for more active clinical follow up, particularly in the persistent mild hypertrypsinaemia group. However, until we are more confident in predicting outcome from genotype, the prospective identification of “milder” mutations should not be included in routine neonatal screening protocols.