Effects of intranasal fluticasone and salmeterol on allergen-induced nasal responses


Dr K. Parameswaran
Firestone Institute for Respiratory Health
St Joseph's Healthcare
50 Charlton Avenue East
L8N 4A6


Background:  Combination of inhaled steroid and long-acting beta-agonists has synergistic effects in asthma.

Objective:  To investigate whether nasal corticosteroid and long-acting beta-agonists have synergistic effects on allergen-induced nasal responses.

Methods:  The effects of intranasal treatment with fluticasone p-MDI (50 μg bid), salmeterol p-MDI (25 μg bid), their combination, and placebo, on nasal symptoms, eosinophil differential cell count and albumin in nasal lavage fluid (measures of inflammation and leakage respectively) and nasal electrical potential difference (measure of epithelial integrity) were studied in 11 atopic subjects with rhinitis, in a randomized, partially-blinded, 4-period, cross-over study. The measurements were made at baseline, at the end of 1 week of treatment, and immediately after a nasal allergen provocation.

Results:  Allergen-induced sneeze, postnasal drip and nasal obstruction were significantly reduced by fluticasone, but not by salmeterol. Eosinophil count in postallergen nasal lavage fluid was significantly less after fluticasone (median 1.9%, IQR 4.6) and salmeterol treatment (median 2.5%, IQR 8.5) compared with placebo (median 12.5%, IQR 27.9). Compared with placebo, both fluticasone and salmeterol attenuated allergen-induced change in nasal potential (mean change from baseline −18.5, +0.4 and −7.2% respectively) and the increase in nasal albumin (median 154, 119 and 130 ng/ml respectively). Combination treatment did not have any additional benefits over the individual therapies.

Conclusions:  Although salmeterol has anti-inflammatory properties, intranasal salmeterol or its combination with fluticasone do not offer any added benefit over intranasal fluticasone alone for allergen-induced nasal responses.

Allergic rhinitis is associated with nasal mucosal inflammation characterized by disruption of the lining epithelium, inflammatory cell infiltration and increased micro-vascular leakage (1). Inflammatory cellular infiltration can be reliably studied by nasal lavage (2). Disruption of the lining epithelium and consequent alteration in the epithelial barrier resistance and ion transport can be assessed relatively noninvasively by studying the nasal trans-epithelial electrical potential difference (3, 4). Nasal allergen instillation has been shown to decrease the potential difference across the nasal epithelium (5). Intra-nasal corticosteroids decrease the recruitment and activation of cells, release of mediators and cytokines and vascular permeability (6). Single dose of intra-nasal salmeterol decreases allergen-induced increase in nasal vascular permeability without affecting the allergen-induced cellular activation (7). Although combination therapy with inhaled corticosteroid and salmeterol has been reported to improve asthma control (8, 9), there is no information on the protective effects of the combination therapy on allergen-induced changes either in the nose or in the lung. Also, a direct comparison of the anti-inflammatory effects of intra-nasal corticosteroid and intra-nasal salmeterol on allergen-induced nasal inflammatory changes has not been made. In particular, there is no information on effects of treatment on the nasal trans-epithelial electrical potential difference or the nasal mucosal membrane-stabilizing effects and how this relates to measurements of cell counts and vascular permeability in nasal lavage fluid.

As both salmeterol and fluticasone have some anti-inflammatory properties, we hypothesized that pretreatment with both salmeterol and fluticasone before a nasal allergen inhalation will attenuate the allergen-induced increase in symptoms, nasal inflammation and decrease in trans-nasal electrical potential difference, compared to pretreatment with a placebo. Also, the combination of fluticasone and salmeterol may offer greater protection than either drug alone. Our objectives were (i) to compare the effects of 1 week of treatment with intra-nasal fluticasone 50 μg bid, intra-nasal salmeterol 25 μg and their combination on allergen-induced changes in nasal trans-epithelial electrical potential, (ii) to correlate the allergen-induced changes in nasal trans-epithelial electrical potential with measurements of inflammation in nasal lavage fluid.



Eleven atopic subjects (mean age 28 years, males = 5) with seasonal allergic rhinitis (without asthma) who had not used any nasal or oral corticosteroid, anti-histamine, cromone or decongestant in the previous 6 weeks were recruited from the staff and students of McMaster University. Subjects were excluded if they had a viral or bacterial upper respiratory tract infection in the past 4 weeks. They were studied outside of their allergy season. All subjects gave written informed consent to participate in the study and the Research Ethics Committee of Hamilton Health Sciences Corporation approved the study.


This was a randomized, placebo-controlled, partially-blinded, four-period, cross-over study. On the first day (baseline), the nose was examined and patients recorded their nasal symptoms. Nasal potential difference was measured and a nasal lavage was then performed. The study drug was administered daily for the next 6 days, including the morning of the seventh day. Symptoms and nasal potential difference were then recorded, followed immediately by a nasal allergen challenge (as described below). Thirty minutes later, nasal potential difference measurement and nasal lavage were done and symptoms recorded again. After wash-out periods of 2 weeks, the same procedure was repeated for the next three treatment arms. The treatments were placebo (two placebo inhalers bid), fluticasone-HFA 50 μg + placebo (bid), salmeterol-HFA 25 μg + placebo (bid) and fluticasone-HFA 50 μg + salmeterol-HFA 25 μg (bid) in random order. A nasal adapter (GlaxoSmithKline, Uxbridge, UK) was used to deliver the medications intra-nasally. Salmeterol and fluticasone inhalers looked identical. However, the placebo canister (no active drug, only an HFA propellant) looked different from the active drugs. Hence the blinding was only partial. Maintaining the computer-generated randomization codes off site ensured allocation concealment. The laboratory measurements were made blinded to the clinical details.

Measurement of nasal responses

Symptoms of nasal itchiness, obstruction, nasal drip and sneezing were recorded on a 10 cm visual analogue scale (0 being no symptom and 10 being the worst symptoms). At subsequent visits, patients were not shown their responses at previous visits. Measurements of nasal potential (3) and nasal lavage (2) were performed as previously described. Briefly, the nasal potential was measured across a reference bridge (21-gauge needle filled with physiological saline in 4% agar) inserted into the subcutaneous space of the forearm and an exploring bridge (thin polyethylene tube, marked at 0.5 cm intervals and perfused with Ringer's solution at a rate of 0.2 ml/min) placed 1 cm posterior to the anterior tip of the inferior turbinate. Both bridges were linked by an Ag/AgCl electrode to the input of a high impedance voltmeter. Nasal lavage was performed by the modified Grunberg method (2). Cell counts in the nasal lavage fluid were performed by modifying the method described by Pizzichini et al. (2, 10) for sputum examination (the volume of the dispersing agent, dithiotreitol, was 1/10th of the volume of the nasal lavage fluid). Albumin in the thawed supernatant was measured by ELISA (Bethyl laboratories Inc, Montgomery, TX).

Allergen provocation

Allergen nasal provocation was performed with the allergen to which the patients were clinically allergic. Allergy skin prick tests (cat, dog, horse, Dermatophagoides pteronyssinus, Dermatophagoides farini, tree, grass, ragweed, cockroach, alternaria, cladosporium) were done with 10-fold dilutions of the extract to be used. The starting concentration to be used for the nasal challenge was two 10-fold dilutions below the lowest found to give a wheal of 3 mm. If a subject had positive skin test to two of the allergens tested, the one with the largest wheal size and clinical allergy was selected. Thus two subjects were challenged with grass allergen, two with ragweed allergen, two with cat allergen, one with birch allergen, three with dust mite (D. farini) and one with dust mite (D. pteronyssinus). After baseline assessment of symptoms and rhinoscopy, one puff (0.2 ml) of starting concentration of allergen was sprayed into each nostril from a metered dose pump spray. If there was no response, the procedure was repeated with a 10-fold higher dose until the highest dose was given or a positive response occurred (a minimum change of 50% in two of the symptoms scores from baseline).


The sample size was calculated to demonstrate a 50% difference in the nasal potential difference or a 50% difference in the percentage of eosinophils in nasal lavage between placebo and the active treatment arms. From a previous study (5) that demonstrated a 51% difference in nasal potential difference following nasal allergen inhalation with a standard deviation of 14%, 10 subjects are sufficient to demonstrate an overall treatment effect (between placebo and active treatment arms) in a cross-over study, assuming an alpha of 0.05 and a beta of 0.2. The outcome measures were the differences in allergen-induced changes from baseline in symptoms and nasal potential, and the postallergen eosinophil count (%) and albumin (ng/ml) in the lavage fluid and they were analyzed by repeated measures analysis of variance with treatment, sequence, period and subjects as factors. The first two outcomes represented a change or delta, and were therefore normally distributed. The measurements in lavage fluid were non-normally distributed and were therefore log-transformed before analysis. Correlation between changes in nasal potential and inflammatory markers in nasal lavage fluid was examined by the Pearson's test. Repeatability of nasal potential measurements (on the placebo treatment arm) was examined by intra-class correlation coefficient (ICC). All analyses were performed using the Statistical Package for Social Sciences, version 10.0 (Chicago, IL).


Effect on nasal symptoms

Allergen inhalation preceded by placebo treatment caused an increase in all four components of nasal symptoms. Fluticasone treatment significantly decreased allergen-induced increase in symptoms of nasal drip (Fig. 1) and nasal obstruction (Fig. 2), whereas salmeterol did not. There was no additional benefit with the combination treatment. None of the treatments had any significant effect on postallergen sneezing or nasal itching or on preallergen challenge symptoms.

Figure 1.

 Effect of treatment on postnasal drip. The open bars indicate changes in symptoms from pretreatment baseline with 1 week of treatment and the closed bars indicate changes observed after allergen provocation. None of the treatments had effects on baseline symptoms. Fluticasone significantly decreased allergen-induced symptom. Combination treatment did not have any additional benefit (*P < 0.05 compared with placebo). Data are expressed as mean and SD.

Figure 2.

 Effect of treatment on nasal obstruction. The open bars indicate changes in symptoms from pretreatment baseline with 1 week of treatment and the closed bars indicate changes observed after allergen provocation. None of the treatments had effects on baseline symptoms. Fluticasone significantly decreased allergen-induced symptom. Combination treatment did not have any additional benefit (*P < 0.05 compared with placebo). Data are expressed as mean and SD.

Effect on nasal potential difference

Allergen challenge, preceded by placebo, caused a mean change of 18% in nasal potential difference compared with preallergen value. This was completely prevented by fluticasone treatment (P < 0.05) and attenuated by 60% (which did not reach statistical significance) by salmeterol. The combination treatment did not have any additional benefit over individual treatments (Fig. 3). There were no significant effects of treatment on the preallergen values.

Figure 3.

 Effect of treatment on nasal potential difference. The open bars indicate changes in symptoms from pretreatment baseline with 1 week of treatment and the closed bars indicate changes observed after allergen provocation. None of the treatments had effects on baseline potential difference. Allergen caused a significant change in nasal potential difference. This was attenuated by both fluticasone (statistically significant) and salmeterol. Combination treatment did not have any additional benefit (*P < 0.05 compared with placebo). Data are expressed as mean and SD.

The nasal potential measurements were repeatable with an ICC of 0.89. The change in nasal potential following allergen challenge correlated modestly with the increase in the proportion of eosinophil (rs 0.3, P < 0.05) and the increase in fibrinogen in the nasal lavage fluid (rs 0.4, P < 0.05). The measurement was also responsive to change. The mean change following allergen inhalation on the placebo treatment arm was −18.5 mV and on the fluticasone treatment arm was +0.4 mV. The responsiveness index was 0.15 (calculated as the ratio of the mean treatment effect (14.9 mV) to the root mean square of the placebo effect).

Effect on nasal inflammation

Allergen inhalation, preceded by placebo, caused a significant increase in the proportion of eosinophils in the lavage fluid. This was significantly attenuated by fluticasone (P < 0.05) and attenuated, but not significantly, by salmeterol (P = 0.06). The combination treatment did not have any additional benefit over fluticasone or salmeterol alone (Fig. 4). Similarly, allergen inhalation caused an increased in the level of albumin in the lavage fluid. This was significantly attenuated by both fluticasone (P < 0.05) and by salmeterol (P < 0.05). The combination treatment did not have any additional benefit over the individual treatments and surprisingly, seemed to be less effective than the individual treatments (Fig. 5).

Figure 4.

 Effect of treatment on allergen-induced eosinophilia in nasal lavage. Allergen-induced eosinophilia was attenuated by all three active treatment arms (not statistically significant by salmeterol), but the combination treatment did not have any additional benefit over individual treatments (*P < 0.05 compared with placebo). Data are expressed as median and interquartile range. The closed circles represent outliers.

Figure 5.

 Effect of treatment on postallergen albumin levels in nasal lavage. Allergen-induced increase in albumin was attenuated by both fluticasone and by salmeterol, but the combination treatment did not have any additional benefit (*P < 0.05 compared with placebo). Data are expressed as median and interquartile range. The closed circles represent outliers.

All subjects completed the study without any adverse events.


This study has made three novel observations. First, treatment for 1 week with a combination of an inhaled corticosteroid and a long-acting beta-agonist did not exhibit an additive anti-inflammatory effect on allergen-induced nasal inflammation. Secondly, salmeterol, while not as effective as nasal fluticasone, attenuated allergen-induced nasal eosinophilia and nasal fluid albumin. Thirdly, nasal electrical potential difference measurement is a valid, repeatable and responsive outcome measurement of allergen-induced nasal pathophysiology. The study thus confirms the anti-inflammatory effect of nasal corticosteroids and raises questions related to the lack of additive anti-inflammatory effect with a long-acting beta-agonist, as has been reported in the airway.

The addition of a long-acting beta-agonist to inhaled corticosteroids produce clinical benefits more than that observed with the individual drugs in majority of patients with asthma. This is widely believed to be beyond their sustained bronchodilator effect, and due to their potentiation of the anti-inflammatory effect of corticosteroids as recently reviewed by Johnson (11). Additive or synergistic anti-inflammatory effects have been reported on in vitro pro-inflammatory cytokine and mediator release from cultured human airway epithelial cells (12), macrophages (13), fibroblasts (14) and smooth muscle (15). The molecular mechanisms may be by facilitation of glucocorticoid receptor translocation to the nucleus (observed in cultured fibroblasts) (14) or synchronization of the activity of the glucocorticoid receptor and CCAAT-enhancer binding protein (C/EBPα), which in turn achieves an optimum antiproliferative action via p21Waf1/Cipl (in smooth muscle cells) (15). In contrast to these in vitro effects, this synergy is less convincing in clinical studies that have assessed airway inflammation. Li et al. (16) made the only published observation of a reduction in the number of activated eosinophils in bronchoalveolar lavage fluid and in bronchial mucosal biopsies when salmeterol was added for 12 weeks to inhaled corticosteroids. Our observations of a lack of additive or synergistic effect in the nose may be due to a number of reasons. First, the in vitro effects may not translate to in vivo efficacy in the nose because of differences in drug concentrations related to the dosage and the short duration of treatment. However, we used dosages that are recommended and clinically relevant. Secondly, it could be argued that the study was underpowered to detect differences between treatment arms. This is true because the study was powered to detect a difference between active treatment arms and placebo, but not between the different treatment arms. However, while fluticasone clearly showed an attenuating effect, not even a trend towards synergy was observed with the combination. Thirdly, nasal tissue may have a fundamentally different response to the combination of a beta-agonist and corticosteroid compared with bronchial tissues and cells that have been used for most of the published studies. This is unlikely because they share a common embryology and developmental chronology. The only differences between the upper and airway are the keratinization of the epithelium, increased vascularity of the lamina proporia and absence of smooth muscle in the upper airway (17). The latter difference, along with the fact that the most convincing molecular mechanism of synergy has been in cultured smooth muscle cells, may suggest that the clinical benefits of the combination in the airway and not in the nose may simply be a reflection of sustained bronchodilatation and not a true anti-inflammatory effect.

Similar to the observations of Proud et al. (7), salmeterol reduced allergen-induced increase in nasal lavage fluid albumin suggesting an effect on vascular permeability. In addition, we observed that salmeterol treatment for 7 days (in contrast to the single dose effect in the earlier study) attenuated allergen-induced eosinophilia compared with placebo, but the effect was less than that of nasal corticosteroids. This may have been by the effect of salmeterol on vascular permeability or on adhesion and diapedesis. In vitro studies on eosinophils isolated from peripheral blood of atopic subjects have demonstrated salmeterol to attenuate β2-integrin-mediated adhesion probably by blockade of cytosolic phospholipase A2 (18). We did not investigate this mechanism in this study. However, the decrease in airway albumin levels and eosinophil numbers were not accompanied by any improvements in clinical symptoms. As we did not measure nasal patency by nasal peak flow or rhinometry, we cannot comment on whether these would have improved with salmeterol treatment.

We employed measurement of nasal electrical potential difference as a surrogate of epithelial integrity. As reported before (5), allergen caused a change in nasal potential difference, presumably by disrupting the epithelial integrity and affecting ion transport. We did not measure the magnitude or the direction of net ion transport by studying the effect of amiloride or indomethacin. However, we observed that both salmeterol and fluticasone attenuated the allergen-induced change suggesting a protective effect on the epithelium. The effect of salmeterol is supported by two previous observations. Prophylactic administration of salmeterol reduced the risk of high-altitude pulmonary edema, particularly in subjects with leaky sodium ion channels in airway epithelium identified by low nasal electrical potential (19). This implies that salmeterol may be able to restore epithelial integrity. More recently, Coraux et al. (20) demonstrated that Pseudomonas aeroginosa toxin could disrupt the junctional barrier of cultured human airway epithelial cells. Salmeterol treatment was able to restore this by increasing the functional activity and expression of Zonula occludens-1, a tight junction associated protein. Our observations may have two clinical implications. First, nasal potential measurements, in addition to phenotyping patients with cystic fibrosis, may have an application in clinical trials investigating airway epithelial integrity. Assuming a 50% attenuation in nasal potential to be clinically significant, 20 paired observations will have 80% power to detect this difference with 95% confidence. Secondly, although salmeterol did not seem to have a benefit over fluticasone in the treatment of allergic rhinitis, its effects in nonallergic and infective rhinitis warrants investigation.

In summary, although salmeterol has anti-inflammatory properties such as reducing the number of luminal eosinophils and restoring epithelial integrity, intranasal salmeterol or its combination with fluticasone do not offer any added benefit over intranasal fluticasone alone for the treatment of allergic rhinitis or controlling nasal inflammation. Further studies are necessary to investigate clinically relevant anti-inflammatory synergy of corticosteroids and long-acting beta-agonists in upper airway.


We are grateful to L. Ellis, Hospital for Sick Children, University of Toronto, and Dr L. Janssen, McMaster University, for helping us to set up the nasal potential measurements, T. Rerecich and K. Radford, McMaster University for assisting with the nasal fluid measurements, Dr N. Stepner and B. Burgess of GlaxoSmithKline Canada, and Dr G. Berlyne, McMaster University, for supplying us with the placebo inhalers and nasal adapters.

The study was supported by an unrestricted educational grant from GlaxoSmithKline Canada. Salmeterol and Fluticasone pressurized metered-dose inhalers were kind gifts from GlaxoSmithKline Canada. Dr Parameswaran was supported by a postdoctoral fellowship from the Canadian Institutes of Health Research.