Airway surface pH in subjects with cystic fibrosis

Cystic fibrosis (CF) is characterised by thick, viscous secretions, chronic bacterial infection and inflammation of the airways. The clinical phenotype is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. One function of CFTR is as a cyclic adenosine monophosphate-dependent chloride channel expressed by many epithelial cells 1. However, how this or the many other described functions of CFTR relate to pathophysiology is unclear. The low volume hypothesis considers that the lack of chloride secretion plus excessive sodium absorption leads to impaired mucociliary clearance and hence chronic bacterial infection 2. In contrast the high salt hypothesis suggests that abnormally elevated sodium and chloride concentrations in airway surface liquid (ASL) impair innate defence mechanisms 3.

In addition to its role in ion transport, CFTR has many other functions. CF cells express increased levels of asialoglycolipids, such as asialoGM1, a receptor for Pseudomonas aeruginosa and Staphylococcus aureus 4, 5. This increased bacterial adherence has been shown to lead to a proportionately increased release of pro-inflammatory cytokines 6, a characteristic hallmark of CF. Furthermore, secreted glycolipids and mucins are over-sulphated, the degree of which correlates with both CFTR mutations 7 and severity of lung disease 8.

CFTR-mediated bicarbonate secretion has recently been identified 9–11. CFTR mutations could thus alter ASL pH which in turn could alter a number of innate defence processes. For example, both ciliary function 12, 13 and mucus viscosity have been shown to be pH dependent 14. Optimal ciliary function is favoured by an alkaline rather than an acidic environment 12. Additionally, mucin solubility is enhanced by a high pH 15 and inhibition of bicarbonate secretion has been shown in an animal model to result in total occlusion of the submucosal glands by mucin 14. Such occlusion, with associated glandular hypertrophy, is one of the first signs of lung pathology in newborn children with CF 16, although this has not been a consistent finding in all studies 17. In addition to these effects on mucociliary clearance, an acidic environment has the potential to adversely affect other aspects of host defence and inflammation. Activation and adhesion of neutrophils 18 and interleukin‐8 production by epithelial cell lines 19 have been shown to increase in conditions of low pH. Inflammatory cells are more likely to undergo necrosis than apoptosis in an acid environment 20, 21, which could contribute to tissue damage from the release of toxic cell contents. In addition to effects on the host, environmental pH has been shown to influence bacterial virulence factors. Lipopolysaccharide (LPS) possesses increased pro-inflammatory properties and Gram-negative bacteria have increased resistance to antimicrobial peptides when grown at low pH. Interestingly, such modifications of LPS have been found in P. aeruginosa cultured from the lungs of CF patients 22. Thus potentially, alterations in airway surface pH may explain many of the phenotypic hallmarks of CF and may lead to new treatment options.

Airway surface pH has been previously measured in the nose, and both the equipment and methodology are well established 23, 24. Previous studies have reported that nasal pH varies with race 25 and the presence of rhinitis 26. Acidification of airway vapour condensate has been shown to occur in acute asthma and is reversible with steroid therapy 21, suggesting a link with inflammation. Further pH measurements in the lower airways have been undertaken in a study assessing the degree of tracheal aspiration secondary to gastro-oesophageal reflux 27. This group has also used this technique to detect micro-aspiration in CF adults, although mean baseline pH was not reported precluding comparison with their reported non-CF values 28. Finally, in vivo murine studies have reported no difference in ASL pH between CF (CFTR-null) and non-CF (wild type) mice 29. However a comparison has not been made in humans.

The possibility that the lack of CFTR-mediated bicarbonate secretion may play a key role in the pathophysiology of CF, led the current authors to compare the airway surface (AS) pH in CF and non-CF subjects. Here, five linked studies comparing two pH probes, measuring the effects of acute hyper- and hypocapnea on AS pH and comparing upper and lower AS pH, are reported.

Methods

Results

Nasal pH measurements

Study 1

In 12 subjects (six CF; six non-CF), values at the edge of the nostril were similar in both groups (CF pH 5.6±0.1; non-CF pH 5.5±0.1). As the probe was advanced along the floor of the nose, the pH rose (fig. 1⇓). Following this, nasal pH measurements were obtained with the antimony probe in 25 CF and 10 non-CF adults. At 4–5 cm, the site at which nasal PD is measured, the mean pH in the CF group was significantly lower than that of the non-CF group (CF 6.2±0.1; non-CF 6.7±0.13; p<0.05).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.—

Nasal pH measured with an antimony probe at different distances from the nares. •: non-cystic fibrosis (n=6); ▴: cystic fibrosis (n=6).

Study 2

Because of the potential confounding influence of the external reference electrode, nasal pH was measured in a further 38 CF (19 males, mean age 26 yrs, mean forced expiratory volume in one second 45% predicted) and 36 non-CF (9 males, mean age 31 yrs) adults at 4–5 cm using the gold probe. In contrast to the above findings, no significant difference was seen (CF 6.8±0.10; non-CF 6.6±0.1) (fig. 2⇓). Similar to other studies, there was no sex difference (males 6.6±0.1; females 6.7±0.1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.—

Nasal pH measured with a gold probe at a depth of 4 cm in cystic fibrosis (CF) (n=38) and non-CF (n=36) adults. No significant difference was found between the groups (CF 6.8±0.2, non-CF 6.6±0.1).

Study 5: lower airway pH

Lower airway pH was successfully recorded in 11 children (table 1⇑). In one child lack of manoeuvrability of the combined bronchoscope and pH probe would not allow passage beyond the larynx.

The mean pH for the lower airways was 7.1±0.2 for CF subjects, and 7.1±0.1 for non-CF subjects. The coefficient of variability for all readings was 2.4% (range 0–8.4%) and did not differ significantly between the groups. However in four children (one CF; three non-CF) a more marked variability was noted in the recordings at different sites (fig. 3⇓). Interestingly, these patients had the most acidic pH values obtained by the group.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.—

Lower airway pH was measured bronchoscopically with the gold probe. •: non-cystic fibrosis; ▴: cystic fibrosis. Measurements for each individual have been plotted.

There was a significant correlation between nasal and lower airway pH measurements (n=11, r²=0.7, p<0.01) for all data (fig. 4⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.—

Correlation of upper and lower airway pH. A positive correlation between nasal pH and mean lower airway pH was identified. r²=0.7, p<0.01.

Discussion

This study suggests that in vivo there is no significant difference between CF and non-CF nasal or lower airway pH. The genotype of the CF subjects did not appear to be an important variable. Airway surface pH remained stable despite acute changes in serum acid/base balance. Finally, a difference in pH measurements made by antimony and gold pH probes was demonstrated, suggesting that in the context of an altered potential difference in CF, only the gold probe provides relevant measurements.

Nasal pH measurements were made with an oesophageal pH probe. Placement of this probe could result in stimulation of the nasal mucosa with subsequent release of alkaline nasal secretions onto the airway surface. To reduce the impact of this, measurements were only recorded after a period of stability, which could take up to a minute to achieve. A further potential limitation of this technique is that the probe could be inadvertently placed into a nasal mucus plug. All patients were asked to clear their nose of secretions by blowing it prior to any measurements. Data were excluded from three subjects as the tip of the probe was found to be covered in mucus upon removal from the nasal cavity. There is a possibility that mucus may be removed from the probe upon its withdrawal from the nasal cavity and that these numbers, therefore, underestimate this problem. However the authors believe that this is unlikely since in both groups of subjects, nasal pH tended to be acidic rather than alkaline. Additionally, the range of pH values observed in the nose and indeed the lower airways compares well with previously reported values from patients with tracheostomies 27.

A discrepancy was identified between the antimony and gold pH probe readings. In order to investigate this further, pH measurements were taken sequentially from the same nostril after a minimum of 10 min. Previous researchers have demonstrated crosstalk between nostrils, such that stimulation of the nasal mucosa from one nostril will result in the release of alkaline nasal secretions in the other 23. In this study, the current authors found that the order in which either probe was used, that is either first or second, did not affect the pH recorded. Importantly, it was demonstrated that the antimony pH probe reads a lower pH in the CF group compared to the non-CF group, which was not seen with the gold probe. The abnormal nasal PD of CF patients 30 may render the measurements of the antimony probe inaccurate, and the authors would therefore recommend the gold probe be used for any further studies of airway pH.

Measurement of lower airway pH proved challenging. In preparation for lower airway readings, it was demonstrated that nasal pH is stable despite acute changes in serum acid/base balance possibly related to the presence of buffers on the airway surface. Recent work has suggested that ammonia produced by the glutaminase enzyme system may play an important role in airway pH homeostasis 31. A disturbance of acid/base status could quite conceivably occur during any general anaesthetic and in particular during bronchoscopy when the airway is obstructed. These results confirmed the ex-vivo findings of Kyle et al. 32. Briefly, using excised ferret trachea in which both cut ends were sealed, Kyle et al. 32 demonstrated the stability of ASL pH despite increasing concentrations of CO₂, both within the external bathing fluid and the tracheal lumen. The findings of Kyle et al 32 are in contrast to Jayaraman et al. 33 who used a murine model to assess the effects of acute respiratory acidosis on ASL pH. Hypoventilation of the mice increased arterial P_CO2 significantly, failed to cause a significant drop in arterial pH but reduced ASL pH. Species differences may account for the contrasting conclusions. For example, the diffusion capacities and buffer systems may differ between mouse, ferret and indeed humans. In the study by Jayaraman et al. 33, the resting murine arterial pH was 7.2, considerably more acidic than resting human blood pH. Despite this, both Jayaraman et al. 33 and the current group found the mean lower airway surface pH to be 7.1. The present study is thus in keeping with data from the ferret, whose airways more closely resemble those of man.

Lower airway measurements were made under general anaesthetic. Gas induction with sevoflurane was used in all patients and anaesthesia was maintained with this agent throughout the procedure. There is no evidence to support acidification of the airway secondary to this agent. Furthermore, it is likely that had there been an effect this would be seen in both CF and non-CF subjects. The authors were aware that local anaesthetic applied to the vocal cords prior to instrumentation could cause a change in airway surface pH and for this reason it was avoided in all patients. The increased cross-sectional diameter of the combined bronchoscope and pH probe limited the size of airway in which measurements could be made and also technically made manipulation of the bronchoscope within the airway difficult. pH readings were taken at individual sites after only a 15‐s period of stability, as previously defined, and all readings were completed within 5 min. Nasal mucus was observed to have an alkaline pH during the measurements and during episodes of rhinitis nasal secretions become alkaline 26. Although the current authors tried to avoid areas covered in mucus in the lower airways, accurate placement of the pH probe in the peripheral airways was difficult and this may account for the apparent trend towards an increase in pH observed in the third generation airways of the CF patients. Further measurements are necessary in order to clarify this.

The CF children attending the current authors' unit undergo bronchoscopy infrequently and recruitment for this study was therefore limited. It was also difficult to find matched controls for lower airway pH measurements in this group, as the non-CF children that undergo bronchoscopy may be less likely to have inflamed or infected airways similar to those of the CF children. The four most acidic lower airway measurements were from a mix of children (three non-CF, one CF). Only one out of these three non-CF children was infected as demonstrated by positive bacteriological culture of bronchoalveolar lavage fluid. It is unlikely the lack of difference in pH between CF and non-CF subjects is therefore a consequence of differing levels of inflammation.

In conclusion, no difference in the pH of the upper and lower airways between cystic fibrosis and non-cystic fibrosis subjects was found. The authors would suggest that the cystic fibrosis transmembrane regulator plays a relatively minor role in airway surface pH homeostasis. Finally, it is recommended that the gold probe is used for future airway pH studies.