INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: cystic fibrosis transmembrane conductance regulator; cystic fibrosis; membrane potentials, physiology; sweat, chemistry

When carefully performed under standardized conditions (1, 2), the sweat test clearly remains the most useful tool for diagnosing cystic fibrosis (CF). In a suggestive context, the presence of a sweat chloride concentration of more than 60 mM strongly supports the diagnosis.

As long as 20 years ago, however, rare cases were described in which CF had to be diagnosed on the basis of clinical evidence, despite normal or so-called borderline sweat chloride values (3, 4). The cloning of the cystic fibrosis transmembrane conductance regulator (CFTR) gene in 1989 (5) raised the hope that these difficult situations could be more easily resolved. In fact, it greatly expanded the CF clinical spectrum. A few mutations that might be associated with intermediate sweat chloride concentrations (30-60 mM) have now been described (6- 17). It has been estimated that an "atypical" phenotype, with suggestive features but pancreatic sufficiency and sweat chloride values below 60 mM, will be found in 2% of patients with CF (18).

We wondered what the incidence of CFTR mutations might be among a group of children with intermediate sweat chloride concentrations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Group

The results of 2,349 consecutive first sweat tests performed at two Belgian university hospitals were reviewed (Cliniques Universitaires Saint-Luc, 2,002 tests over six years; Gasthuisberg, 347 tests over 10 months). At each hospital, the sweat test had been carried out by the same experienced technicians, using the Gibson-Cooke method (1). One hundred and four tests (4.2%) revealed intermediate chloride concentrations, that is, between 30 and 60 mM. The percentage of such tests was similar for the two hospitals. Sixty-eight of these tests had been performed on subjects younger than 18 years of age. With the agreement of the local ethics committee, 43 of these (63%) could be traced, and in each case the parents gave informed consent for their child to take part in the study. A blood sample for genetic analysis was obtained from each subject. Further investigations, including whenever possible a detailed study of nasal potential difference (19, 20), were carried out on all subjects found to be carriers of two CFTR mutations.

DNA Analysis

After amplification by polymerase chain reaction (PCR), a careful analysis of all exons of the CFTR gene and their intronic flanking regions was performed by denaturing gradient gel electrophoresis (DGGE) and sequencing of anomalous DGGE patterns. The database of the Cystic Fibrosis Genetic Analysis Consortium, available on the Web (http://www.genet.sickkids.on.ca/cftr/), was used as reference. A screening was also performed for the 3849 + 10 kb C T mutation, and the alleles at the Tn locus in intron 8 of the CFTR gene were determined. The 5T allele in intron 8 was considered as a mutation with variable, partial penetrance (16, 21). When two mutations were found in a subject, a blood sample was taken from his/her parents to rule out transmission on the same CFTR gene.

Nasal Potential Difference Test

Potential difference (PD) of the nasal respiratory epithelium and activity of ion channels sensitive to amiloride, isoprenaline, and low chloride content were investigated as previously described (19, 20) in nine of 10 subjects found to be carriers of two CFTR mutations. One study was performed under ketamine anesthesia for cardiac catheterization (Subject 7). Briefly, a flexible umbilical vessel catheter was placed under the inferior turbinate of the nose (exploring electrode). A reference electrode was placed on the inner forearm after lightly abrading the skin. Recordings of baseline voltages were made while perfusing the nasal membranes with an isotonic basal solution (NaCl, 140 mM). Solutions were applied in the following sequence: amiloride (100 µM) in basal solution, amiloride in low-chloride (6 mM) solution, and finally isoprenaline (10 µM) in low-chloride solution containing amiloride. Raised baseline voltages and absent or reduced chloride secretion, corresponding to cumulative changes after perfusion with low-chloride solution plus isoproterenol in the presence of amiloride, have been shown to optimally discriminate between subjects with CF and control subjects (19). In our laboratory, mean (±SD) values from 27 healthy subjects are 17 (±1.3) mV for maximal potential difference (PD) and 14.7 (±1.3) mV for chloride secretion. Cutoff levels are set at 30 and 10 mV, respectively.

Statistical Analysis

The Mann-Whitney and Fisher tests were used. Statistical significance was considered to be demonstrated by a two-tailed p value of less than 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study of 86 chromosomes disclosed 24 putative CFTR mutations, including three 5T allele in intron 8. Three subjects were found to be heterozygotes, carriers of a single mutation (F508, L165S, and G576A). Their sweat chloride concentrations were 32, 36, and 45 mM, respectively. Ten subjects (23.3%) were found to carry a CFTR mutation on both chromosomes 7 (Group A). Tables 1, 2, and 3 summarize the clinical and biological profiles of these subjects, whose mean age (±SD) at the time of writing was 8.9 (±4.2) years. Only one of these children was pancreatic insufficient. The mean sweat chloride concentration (±SD) in Group A (39.4 ± 5.4 mM) was slightly higher than the corresponding value in children not carrying any CFTR mutation (Group B: 35.2 ± 4.4 mM, n = 30) (Figure 1). In Group A, the mean sweat chloride concentration was higher than the sodium concentration (Cl/Na ratio: 1.10 ± 0.31 versus 0.85 ± 0.21 in Group B, p < 0.01). Nasal PD measurements highly suggestive of CF were found in five cases.

View larger version (33K): [in this window] [in a new window]   Figure 1.   Sweat chloride levels and genotype.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have investigated children with sweat chloride concentrations between 30 and 60 mM. Such values are usually referred to as borderline (40-60 mM) or normal (< 40 mM). The main observation to emerge from this study is that nearly one in four children in this group (10 of 43) appeared to carry putative mutations on both CFTR genes. This frequency is higher than expected. Desmarquest and coworkers studied a smaller group of children with borderline sweat chloride values (22). When including the 5T allele (IVS8-5T), these authors found a similar proportion of subjects carrying a CFTR mutation on both genes (five of 23).

In our study, the average sweat chloride concentration of these children is higher than in those without CFTR mutation, but the overlap is such that this difference is of no use in practice (Figure 1). A Cl/Na ratio > 1 has been considered as suggestive of CF in patients with borderline sweat test, although a ratio < 1 did not suffice to exclude the diagnosis (23). Accordingly, a ratio > 1 was observed in 50% of children displaying two CFTR mutations and in three of the four patients with abnormal nasal PD but in only 10% of subjects from Group B. The incidence of a single CFTR mutation (three of 43) is not significantly different from that estimated for the general population in Belgium, which is about 3.9% (24). This is not unexpected because, although it has been shown that the mean sweat chloride concentration was higher in heterozygous infants than in control subjects, this difference was only modest (14.9 ± 8.4 versus 10.6 ± 5.2 mM) and of no practical significance (25).

Can CF be diagnosed in the 10 children carrying two putative CFTR mutations? Clearly, sweat chloride values lower than 60 and even 40 mM do not exclude this diagnosis, especially in young children. Although useful in most cases, the cutoff value of 60 mM is inappropriate in infants (25) and has been questioned for older patients (26). In the latter study, however, data are likely influenced by the frequency of the A455E mutation in the Dutch population. This mutation is one of the few that have been occasionally and at times more consistently associated with borderline or normal sweat electrolyte levels (10, 11). Other examples are the 3849 + 10 kb C T (6-9), R347H (12), G551S (13), D1152H (14), R117H (15, 16), and R117C (17) mutations. Sweat chloride values as low as 16 mM have been reported (7). Current guidelines for the diagnosis of CF (18) suggest that a patient must present both clinical evidence of the disease, such as a compatible clinical picture, family history, or positive neonatal screening, and evidence of abnormal in vivo CFTR function or identification of two CF-causing mutations.

Table 3 summarizes data relevant for the diagnosis in the 10 children of Group B, with reference to these guidelines. It illustrates the complexity of these situations. In several cases, the diagnosis remains debatable. Potential limitations of each criterion contribute to these ambiguities. Specifically, (1) suggestive symptoms in subjects with mild forms of the disease could emerge with time; (2) more than 900 putative CFTR mutations have now been reported but some of them are not CF-causing mutations as defined by a consensus panel (18); and (3) a positive nasal PD study strongly supports the diagnosis of CF but caution is advised in interpreting a negative test.

Nasal PD measurements were considered abnormal in five patients (Subjects 1, 4, 5, 6, and 10). Data on the R117C mutation (Subjects 5 and 6 are twin sisters) and the S977F mutations (Subject 4) are limited but suggest that a mild form of the disease is associated with the former (27). D1152H is a common mutation in men with congenital bilateral aplasia of the vas deferens (CBAVD) (28-31). It might also be associated with variable although usually mild respiratory involvement (32). Pulmonary symptoms are significant in Subject 1 and typical in Subject 10. Nasal PD was normal in four children (Subjects 2, 7, 8, and 9). Two of them are carrying a classic CFTR mutation on one gene and the R117H mutation on a 7T background on the other (Subjects 2 and 7). A similar genotype was found in the single subject without nasal PD study (Subject 3). According to the consensus panel (18), this combination (R117H-7T) does not meet the criteria for a CF-causing mutation and a demonstration of CFTR abnormality by sweat testing or nasal PD testing is required to support a diagnosis of CF in such cases. It has usually been associated with an atypical presentation such as CBAVD (33, 34). The repeated presence of Pseudomonas aeruginosa and/or Staphylococcus aureus in the sputum of Subject 6, who also suffers from recurrent and serious pulmonary infections, is suggestive of CF, but his complex congenital cardiac malformation and the frequent hospitalizations it has necessitated may be a contributory factor.

Little is known about the S1235R mutation (Subject 9) (35, 36). An in vitro study shows that it has no measurable impact on the CFTR chloride channel function, suggesting that it could not be a CF-causing mutation (37). However, it cannot be excluded that the regulatory properties of CFTR are affected and in this way cause disease.

The 5T variant in the intron 8 polythymidine tract (IVS8-5T) is a common allele with a frequency in the general population of about 5% in several countries (16, 38, 39). IVS8-5T has consistently been found in increased frequency among male patients with CBAVD (38-41). It has also been reported in excess among patients with pancreatitis (42, 43), in newborns with hypertrypsinemia and normal sweat test (44), and in patients with atypical sinopulmonary disease (39). This variant has been shown to downregulate CFTR expression. However, its effect is highly variable among different individuals and between different organs of the same individual, which provides the molecular basis of the partial penetrance of its clinical expression (21). Interactions with other polymorphisms in the CFTR gene may also explain the variability of the associated phenotypes (45). Without evidence of CFTR dysfunction, the 5T mutation alone is not considered a CF-causing mutation (18). A homozygous carrier of the 5T variant of intron 8, Subject 8 is unlikely to develop respiratory symptoms of CF, although this has been reported (46). Infertility due to obstructive azoospermia could be a problem (38, 39).

These discussions are not simply semantic. Every time the diagnosis remains questionable, and perhaps even in some truly asymptomatic patients, the psychological burden of the announcement of such a diagnosis (mentioning the uncertainties) must be weighed against the potential advantages: improved adherence to regular follow-up, genetic counseling, and institution without delay of appropriate treatment should symptoms appear.

In the vast majority of cases, the sweat test remains the essential diagnostic tool that, in a suggestive context, makes it possible to confirm a diagnosis of CF. Although the threshold of 60 mM for the sweat chloride concentration has proven to be discriminating and useful in clinical practice (47, 48), we observed in this study that almost one of four children with sweat chloride values between 30 and 60 mM was a carrier of a putative CFTR mutation on both chromosomes 7. This confirms the necessity of a less rigid approach to the interpretation of sweat test results. Further specialized investigation, including exhaustive genetic analysis and measurement of nasal potential or intestinal current (49), is justified in children with symptoms suggestive of CF and such sweat chloride concentrations.