INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cystic fibrosis (CF) is one of the most frequent autosomal recessive disorders of the Caucasian population (1), affecting approximately 1 in 2,000 live births in Belgium. The CF gene is located on chromosome 7 and has been identified in 1989 (2). Over 800 mutations have been recognized in the cystic fibrosis transmembrane conductance regulator (CFTR) (3). The most common mutation, F508, is found in 70 to 80% of all CFTR genes in Northern Europe (4, 5).

Considerable efforts have attempted to find genotype-phenotype correlations in CF. Such correlations appear to exist for pancreatic function, age at diagnosis, and pulmonary function. The mutation F508, associated with pancreatic insufficiency and diagnosis at younger age, is classified as a "severe" mutation (6), whereas others such as R117H and A455E are considered to be "mild" mutations. A455E also correlated with better lung function (7); a correlation between F508 and lung function was not found (8, 9).

Genotype-phenotype correlations for paranasal sinus disease have not been found either by the Cystic Fibrosis Genotype-Phenotype Consortium (6) or by Hubert and coworkers (10) who used a functional mutation classification. However, in most studies paranasal sinus disease was not addressed specifically. In reports that focused on the nose and paranasal sinuses, indications for a genotype-phenotype correlation for the paranasal sinuses were found. Indeed, Moss and King (11) reported a trend for more severe mutations in their paranasal sinus surgery group: F508 homozygosity was found more frequently in the surgical CF patients than in a control CF population: 58% versus 48%. This finding was further confirmed by Kingdom and colleagues (12) in a cross-sectional study of a U.S. data base. They reported a significantly higher prevalence of the F508/F508 and F508/G551D genotypes in patients with nasal polyps requiring surgery: 57.5% versus 49.9% and 12% versus 8%, respectively. Recently, De Gaudemar and coworkers (13) reported a higher frequency of the F508 allele in the nasal polyposis group in comparison to the nonpolyp group: 82% and 66%, respectively.

Upper airway manifestations of CF include maxillary sinus opacification in virtually all patients, symptoms and signs of chronic or recurrent sinusitis, and polyps. Both paranasal sinus surgery (11, 12) and polyps (6, 10, 13) have been used for staging upper airway disease in CF. Polyps can be investigated using anterior rhinoscopy (6, 10), but nasal endoscopy (13) is far more reliable. Polyps are rare under the age of 5, but are found in 25 to 40% of CF patients above the age of 5 (11, 13- 16). Clinical classifications other than surgery (yes/no) and nasal polyps (yes/no) have not yet been reported.

In this study we investigated whether a correlation existed between the nature of particular CFTR mutations and the phenotype for paranasal sinus disease, determined endoscopicallyevaluating polypsand by a clinical classification, including surgery. Most attention was given to F508. These genotype-phenotype correlations may have practical, clinical implications since they may lead to closer follow-up and to more intensive treatment in the predisposed group.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CF Population

Of a population of 129 patients with CF, 113 patients, followed at the Department of Paediatrics of the Leuven University Hospital, were included after examination by an ear, nose, and throat (ENT) surgeon at normally scheduled control visits during the period September 1995 till June 1996. All patients had repeated sweat tests of more than 60 mEq/L (17) and except for 11 unidentified mutations in 10 patients, their genotype was known. Seven patients were not included because genetic data were not available. Nine patients were not included because they were not examined by the ENT surgeon during this period. The demographics are presented in Table 1.

Pulmonary Function

Pulmonary function tests were done using a pneumotachograph (Jaeger, Masterscreen, Hoechberg, Germany) and according to European Respiratory Society (ERS) guidelines (18). FVC and FEV₁ are expressed as percentage of the predicted value for sex and height according to Zapletal and colleagues (19). Children younger than 6 yr are unable to reliably perform pulmonary function maneuvers. Data were available on 71 CF patients (Table 2A).

Pancreatic Function

Patients were considered pancreatic sufficient when fat absorption was > 93% as determined by a fat balance study and when pancreatic enzyme supplementation was considered to be unnecessary.

Paranasal Sinus Clinical Classification

A clinical classification was made based on a questionnaire, the patient's histories, and their files. The questionnaire included items assessing the frequency and severity of nasal obstruction, the number of common colds per year, the frequency of acute sinusitis episodes, the quantity and quality of nasal secretions, the frequency and severity of headache, the presence of postnasal drip, and the association between upper and lower airway disease. Also the use of topical medication and systemic antibiotics for upper airway disease was recorded. For the number of common colds, a comparison was made with the generally accepted number of upper respiratory tract infections related to age: up to 10 for preschool children, up to 6 for the age group 6 to 10 yr, and 2 to 4 for adolescents and adults (20).

Class "0" indicated that the CF patient did not have more upper respiratory tract problems (upper airway infections, nasal secretions, nasal obstruction complaints) than accepted for that age. Neither the patient nor the parents reported any current or previous specific upper airway related problems more than one could expect according to the age of the patient. They can be called paranasal sinus "asymptomatic" patients.

When specific paranasal sinus problems were present, e.g., upper airway infections exceeding the number stated for class 0, the CF patient was classified in class "1," but only if the paranasal sinus disease was not of major concern for the patient, his family or pediatrician and if the problems did not include chronic nasal obstruction, chronic problematic purulent nasal discharge, or chronic headache.

Class "2" included the CF patients with "severe" paranasal sinus disease: recurrent acute sinusitis episodes, chronic nasal obstruction, chronic purulent nasal discharge, and/or chronic headache. The paranasal sinus disease was considered a significant health problem.

Finally, class "3" is the paranasal sinus surgery group. Every CF patient who had undergone paranasal sinus surgery was assigned to class 3. Adenoidectomy was not considered paranasal sinus surgery. Indications for surgery were made in principle by the pediatricians after consulting an ENT surgeon. Medical therapy before surgery consisted mainly of prolonged oral antibiotic treatment, topical (nasal) and oral steroids, and intense rinsing of the nasal cavity. This decision was made without taking into account the F508 status, which at that time was largely unknown. The standard surgical procedure was an endoscopically controlled intranasal ethmoidectomy (functional endonasal sinus surgery [FESS]) with opening and clearing of all paranasal sinuses. A wide middle meatal antrostomy was created. Ninety percent of all surgical procedures were done by the same surgeon (M.J.).

Two ENT surgeons independently evaluated the assignment of the patients to class 0, 1, and 2 and had to come finally to an agreement. This final classification was made without prior knowledge of the nasal polyp status and the F508 status.

Polyps (Ethmoidal Polyps and Nasal Polyps)

All 113 patients underwent nasal endoscopy using either a rigid endoscope (2.7 mm diameter, 30° angle, Panoview, Wolf, Knittlingen, Germany) or a flexible endoscope after local anesthesia. The presence or absence of polyps was noted and they were classified according to their inferior border. The classification used here is a reduction of the routine grade 0 to 5 classification. Polyps within the ethmoid, eventually coming down to the inferior border of the middle turbinate, but not beyond, are ethmoidal polyps. Nearly always ethmoidal polyps can only be seen using an endoscope. Nasal polyps are those that come out of the ethmoid. Only the larger ones (down to the inferior turbinate) can easily be seen using anterior rhinosopy. Both nasal cavities were evaluated in each patient but only the one with the highest grade was used for calculation and statistical analysis (Table 2B).

CFTR Mutations

The nature of the mutations was determined using the INNO-LiPA CF assay (Innogenetics, Zwijndrecht, Belgium) for the eight most frequent mutations found in the Belgian population. In addition, sequencing of the complete coding region and the exon/intron junctions of the CFTR gene was performed in most patients where no mutations were identified using the INNO-LiPA CF assay (24) (Table 3).

Statistical Analysis

Statistical analysis was done using the genotype as the independent variable. Analysis was done for the presence or absence of polyps, for the four clinical classes, and for surgery alone. Age was considered as an additional parameter. Regression and nonparametric analyses were done when indicated. Chi-square and when necessary Fisher exact test were used. The Fisher exact test was used whenever 25% of the values was less than five. The test used for statistical analysis is always indicated. The level of significance was put at 5% (p < 0.05).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The 113 patients with CF were equally distributed over the four clinical classes. The surgical group comprised 25% of all CF patients (Table 2A). There were no differences in age and sex between the different classes. Four patients were classified as pancreatic sufficient. Three of them had a mild mutation (R117H or A455E) and were clinically classified as class 0 (n = 2) or class 1 (n = 1). The fourth patient had undergone sinus surgery and was F508 negative (G542X/unidentified). The mean and standard deviation for FVC and FEV₁, expressed as percentage of the predicted value, are given in Table 2A. Pulmonary function in the surgical group (class 3) was at least as good as in all other groups.

F508 was present on 171 (75.7%) of all 226 CF chromosomes analyzed. Homozygosity for F508 was found in 69 patients (61%), 33 patients (29%) were compound heterozygotes, and 11 (10%) were F508-negative. The list of non-F508 mutations is presented in Table 3. In 215 CF chromosomes (95%) a mutation was identified. In one patient no mutation was found. Six patients were F508 compound heterozygotes, but the second mutation was not identified. Three patients had a mutation other than F508 on one chromosome, but the second mutation was not identified.

Polyps were found in 56% of the CF patients (see Table 2B). The smaller, purely ethmoidal polyps were found in 17% of the CF patients; the larger, nasal polyps were present in 39%; in 25% the polyps could be considered as being large. The patients with pure ethmoidal polyps accounted for 30% of all patients with polyps (nasal + ethmoidal). Considering the mean polyp score (left + right) instead of the maximal polyp score did not alter the results (data not shown).

There was a correlation between the F508 status and the clinical classification (p < 0.02, Fisher exact) (Table 4A). Clinical classification "0" was more frequently associated with compound heterozygosity and F508 negativity. Only 38% of these "asymptomatic" patients were F508 homozygous, compared with 79% in the surgical group. There was a linear increase in the percentage of F508 homozygous patients with increasing clinical classification (p < 0.002, Mantel-Haensel chi-square). Also, when comparing the surgical group (class 3) with the nonsurgical group (class 0, 1, and 2), F508 homozygosity was more frequently present in the surgical group (p < 0.03, Fisher exact) (Table 4B).

The polyps correlated also with the F508 status (Table 5). There was a statistically significant difference between the polyp and the nonpolyp group (p < 0.05, Fisher exact), but only when considering all polyps, i.e., including the smaller, purely ethmoidal polyps. When considering only nasal polyps, the ones extending beyond the middle meatus, no difference was observed (p = 0.43, chi-square).

Approximately one-third (22/69) of the F508 homozygous CF patients had undergone paranasal sinus surgery, compared with one-seventh (6/44) of the compound heterozygous and F508-negative CF patients. The relative risk for sinus surgery in the described CF population is statistically significantly higher (p = 0.02, Fisher exact) for the F508 homozygous subgroup than for the non-F508 homozygous subgroup: × 2.33 (95% confidence limits: 1.030 to 5.305).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first report of a clear correlation between CFTR mutations and various clinical manifestations in the nose and paranasal sinuses. Specifically F508 homozygosity was found to correlate not only with paranasal sinus surgery, but also with the clinical presentation and with polyps, but the latter only when the pure ethmoidal polyps were included. The relative risk for paranasal sinus surgery in cases of F508 homozygosity is 2.33 times that of non-F508 homozygous CF patients.

Previous and recent investigations failed to find such a correlation at the level of nasal polyps (6, 13), or reported only a trend when considering surgery (11): more F508 homozygosity in the surgical group. Only Kingdom and colleagues (12) recently reported a correlation between surgery for nasal polyps and F508. The absence of an upper airway genotype- phenotype correlation in the Cystic Fibrosis Genotype-Phenotype Consortium report (6) and in the publication by Hubert and coworkers (10) may be explained by the fact that they only considered the larger, nasal polyps. A similar lack of correlation for larger nasal polyps was also found in the present study. In addition, the reported prevalence of nasal polyps in patients with CF (6) was only 15%. This is much lower than the 25 to 40% seen with nasal endoscopy (11, 13- 15), but within the 25% range of largereasily diagnosed by anterior rhinosopynasal polyps.

In conclusion, polyps, evaluated using nasal endoscopy, are an important parameter for paranasal sinus disease in CF and correlate with the genotype.

The second parameter for upper airway disease in CF was sinus surgery. Although the indications for surgery may be subjective, the surgical event itself is an objective element. In addition, surgery was decided here in a standardized way by a limited number of physicians, mostly K.D. and M.J., without considering the genotype in all cases and before mutation analysis was carried out in most cases. Surgery was only considered when maximal medical therapy, including long-term antibiotics, topical and oral steroids, and intense nasal lavages, failed to control the symptoms and signs. This created maximal guarantee for objectivity when looking at genotype-phenotype correlations.

In contrast to the surgical decision, subjective interpretations may have played a role in the distinction between class 0 and class 1 and between class 1 and class 2. To overcome these problems or limit their impact on the classification, two ENT surgeons independently and without knowing the genotype classified the patient and provided a consensus advice. The distinction between "no problems" (class 0) and "severe problems" (class 2) was clear and without discussion.

The present study provides a number of arguments to support the genotype-phenotype correlation when looking at sinus surgery or at clinical classification. Not only is there a correlation between F508 homozygosity and paranasal sinus surgery (p < 0.03); linear regression analysis shows a highly significant trend between the clinical severity of paranasal sinus disease and the proportion of F508 homozygosity (p < 0.002). An additional argument can be found in the relative risk calculation for sinus surgery.

However, even with all these consistent statistical findings, the authors realize that in a relatively small CF population (n = 113) small changes or shifts may have a great impact. In particular, additional surgery on a limited number of compound heterozygous patients or F508-negative patients would have resulted in loss of statistical significance in a number of tests. The same could be true if less surgery was done on F508 homozygous patients.

Radiology was not included as a parameter in the present study for a number of reasons. Standard sinus radiographs cannot differentiate different phenotypes. Conventional tomography is ethically no longer acceptable, because of elevated exposure to radiation and the lack of soft tissue differentiation. Computed tomographic (CT) scan would be a better choice, but was limited to surgical cases and therefore could not be used in the present study. Paranasal sinus CT scans are difficult to obtain both in the younger patients (less than 5 yr) and in "asymptomatic" CF patients. Moreover, the prevalence, spontaneous evolution, sensitivity, and specificity of a number of CT parameters, such as maxillary pseudomucocele, sinus development, sinus opacification, and polyps, remain to be determined (26). The value of magnetic resonance imaging (MRI) in CF also remains to be determined.

Paranasal sinus surgery is the second most frequent type of surgery in CF after laparotomy for meconium ileus (29). The incidence of paranasal sinus surgery in CF has not been studied with precision, but is estimated to be higher than 15% (12, 28, 29). One-fourth of the CF patients underwent paranasal sinus surgery in the present study, which seems a high proportion, but may be related to the high frequency of F508 homozygosity. A more conservative attitude concerning the surgical indications and/or more intensive, conservative treatment schedules for the exacerbations and/or the baseline rhinosinusitis, such as daily nasal lavages with aminosides (15, 16), possibly could have reduced the number of surgical interventions, but it is impossible to predict in which patients surgery would have been prevented, and it is unlikely that this would have influenced the overall genotype-phenotype correlation.

One-third of the homozygous CF patients had paranasal sinus surgery in this study, compared with one-seventh of the compound heterozygous patients and F508-negative patients. This clearly illustrates the risk of F508 homozygosity for paranasal sinus disease and surgery. More attention may therefore be given to the paranasal sinus problems, for precise clinical diagnosis and follow-up in this group.

In conclusion, based on the present study F508 homozygosity should be regarded as a risk factor for paranasal sinus problems in CF.