INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Plasma protein exudation is one of the aspects of airway inflammation in asthma (1, 2). In previous studies, we have reported that in patients with asthma the degree of nonspecific bronchial hyperresponsiveness correlated significantly with the degree of plasma protein exudation, assessed either in bronchoalveolar lavage (BAL) fluid or in sputum (3, 4).

Corticosteroids are effective in controlling asthma symptoms and in reducing airway obstruction and bronchial hyperresponsiveness (5). These beneficial effects of inhaled corticosteroids are attributed to their antiinflammatory action on the bronchial mucosa (6, 7). Corticosteroids have been reported to inhibit vascular permeability (8). Fluticasone propionate (FP) is a relatively new inhaled synthetic corticosteroid, and has been reported to have a higher topical antiinflammatory effect than other inhaled steroids (9). No information is available on the effects of FP on the blood-to-air space permeability for plasma proteins in patients with asthma.

In the present double-blind, parallel-group study we compared the effect of FP and salbutamol (Sb) on inflammatory indices in BAL fluid, and on clinical parameters in patients with asthma. The primary outcomes that were monitored included plasma protein leakage into the airway lumen, cell differentials in BAL fluid, and bronchial hyperresponsiveness to histamine. Significant clinical changes in the primary outcomes were defined as a  50% decrease in blood-to-air space permeability and percentage of eosinophils in BAL fluid, and more than one doubling dose increase in PC₂₀histamine (the concentration of histamine that causes a fall in FEV₁ of 20% of the baseline value). The sample size of the study was based on the results of previous studies (3) and a power of 0.80 to detect the predetermined changes. Secondary outcomes included daily symptoms, morning and evening peak expiratory flow (PEF), and frequency of rescue medication used (salbutamol, 400 µg). In addition, correlations between primary outcomes were analyzed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Thirty-three patients (19 men and 14 women) with mild to moderate asthma were recruited from patients attending the outpatient clinic of pulmonology of the Academic Medical Center (University of Amsterdam, The Netherlands). Asthma was diagnosed according to the American Thoracic Society (ATS) criteria and included a history of recurrent episodes of wheezing, chest tightness, and dyspnea and a normal lung function between asthmatic attacks (10). Asthma severity before institution of treatment ranged from Grade I (episodic) to Grade III (moderate persistent), according to the staging as proposed in the Global Initiative for Asthma (11).

The following inclusion criteria were used during the run-in period (study design shown in Table 1): (1) documented history of asthma; (2) age between 18 and 60 yr; (3) at Visit 1 or 2: FEV₁ > 60% of predicted, having withheld inhaled bronchodilators for at least 8 h; (4) at Visit 1 or 2: if 60% < FEV₁ < 80% predicted: a reversibility of the FEV₁ of at least 15% of predicted, 15 min after inhalation of 400 µg of salbutamol; (5) at the end of the run-in period: an airway responsiveness to histamine: PC₂₀histamine < 8 mg/ml.

Exclusion criteria at the start of the run-in period were as follows: (1) serious or unstable concomitant disease; (2) upper and/or lower respiratory tract infection during the 2 mo before entry; (3) hospitalization for any aspect of airway disease or for an acute exacerbation of symptoms, requiring a change in medication within the 2 mo before entry; (4) use of oral corticosteroids at maintenance dose levels during the 12 mo before entry or short courses of oral steroids during the 2 mo before entry; (5) use of inhaled corticosteroids, sodium cromoglycate, nedocromil sodium, oral antihistamines, or methyl xanthines during the 2 wk before entry; (6) if patient smokes, variable smoking habits; (7) use of -receptor antagonists; (8) hypersensitivity to -receptor agonists or to any other component of the study treatment; (9) pregnancy or lactation; (10) use of an investigational drug within 1 mo before entry; (11) inability to use a peak flow meter correctly. The study was approved by the Internal Review Board of the Academic Medical Center, and was performed after written informed consent had been obtained. Characteristics of patients fulfilling the entry criteria are shown in Table 2.

Study Design

The study was a double-blind, single-center, parallel-group study in which patients were randomized to receive either 250 µg of FP twice a day or 400 µg of Sb twice a day via rotadisk powder inhaler (Glaxo Wellcome, Zeist, The Netherlands). Stratification was done for nonsmoking versus ex-smoking and smoking. The study design is shown in Table 1. After entering the study a 4-wk run-in period preceded a 12-wk treatment period. The treatment period was followed by a 2-wk follow-up period. Patients on maintenance treatment withheld medication from 2 wk before the start of the run-in period onward, and were provided with rescue salbutamol rotadisks (400 µg per blister) for symptomatic relief. Patients were asked to withhold this medication for 8 h before each visit.

At the beginning of the run-in period a full medical history, physical examination, lung function tests, routine blood tests, and urine analysis were performed. Blood was obtained for a radioallergosorbent test (RAST). Furthermore, patients received a peak flow meter and diary cards to record throughout the study period: nighttime and daytime symptoms of breathlessness, general well-being, morning and evening PEF, volume and color of sputum, and use of rescue medication. Patients scored symptoms on a scale from 0 (none at all) to 4 (severe). During the whole study period, at each visit, lung function was tested and diary cards were checked. A histamine provocation test was performed 3 to 4 d before the treatment period. At the end of the run-in period a blood sample was taken to obtain serum and ethylenediaminetetraacetic acid (EDTA)-plasma and a bronchoscopy with BAL was performed.

At start of the treatment period patients were randomized to receive double-blind treatment with FP or Sb. All patients were provided with rescue salbutamol rotadisks for symptomatic relief during the treatment period. After 1 wk of treatment patients were visited at home to check whether the study medication, peak flow meter, and diary cards were being used correctly. Patients were instructed not to use study medication within 12 h before each visit. During the treatment period, patients were reviewed after 4, 8, 10, and 11 wk of treatment. At 11.5 wk a second histamine provocation test was performed, followed by a second bronchoscopy with BAL after 12 wk of treatment. Before each bronchoscopy, a blood sample was taken to obtain serum and EDTA-plasma.

After the second bronchoscopy patients restarted their regular medications and returned to the clinic for a final control 2 wk later.

Lung Function Tests

The forced expiratory volume in 1 s (FEV₁), maximal midexpiratory flow (MMEF), forced expiratory flow at 25-75% (FEF_25-75%), peak expiratory flow rate (PEFR), forced expiratory vital capacity (FEVC), and inspiratory vital capacity (IVC) were measured at each visit with a dry rolling-seal spirometer (SensorMedics BV, Bilthoven, The Netherlands) according to standardized guidelines (12). All values are expressed as the percentage of the predicted value. Bronchial reactivity to histamine was determined by a 2-min tidal-breathing method (13). The PC₂₀histamine was defined as the interpolated concentration of histamine that caused a fall in FEV₁ of 20% of the baseline value.

Patients received a mini-Wright peak flow meter to measure peak flow rates every morning and every evening, performing three measurements on each occasion and entering the highest value on the diary card. Patients were carefully instructed in the use of the mini-Wright peak flow meter (sitting in an upright position). As far as possible all measurements were performed at the same time of day, and before taking the study medication (7.00 A.M. and 7.00 P.M.).

Fiberoptic Bronchoscopy, Lavage Fluid and Cell Processing

Bronchoscopy was performed as previously described (3). Briefly, after premedication with atropine and codeine and after local anesthesia with lidocaine, a flexible fiberoptic videobronchoscope (p200; Olympus, Norwood, MA) was wedged into a subsegment of the middle lobe. Seven successive 20-ml aliquots of prewarmed NaCl (154 mM) were instilled and each was aspirated immediately by low suction. Transcutaneous oxygen saturation was monitored throughout the bronchoscopy.

The first three aliquots were combined and named Pool 1. Aliquots 4-7 were combined and named Pool 2. The pools were immediately centrifuged at 500 × g and 4° C. The cell-free supernatant was taken and Pool 2 was divided into two equal portions. To one of these portions EDTA (Merck, Darmstadt, Germany) was added (final concentration, 2.8 mM). All supernatants were stored at 80° C until analysis. On the basis of previous results, all measurements described in this study were performed using Pool 2 (3).

Cells were resuspended in phosphate-buffered saline (140 mM NaCl and 10 mM sodium phosphate, pH 7.4) containing 0.5% (wt/vol) bovine serum albumin (Boseral DEM; Organon Teknika, Boxtel, The Netherlands) (PBA). The number of erythrocytes was counted with Daecie suspension (1.2% [wt/vol] trisodium citrate · 2 H₂O plus 0.4% formaldehyde, 37%) in a Fuchs-Rosenthal counting chamber. BAL fluid samples containing more than 2.5 × 10⁵ erythrocytes per milliliter were excluded from analysis of plasma protein leakage in order to exclude increased protein levels resulting from significant bleeding during bronchoscopy. The total cell number was determined by counting manually in a Bürker counting chamber. Cells were cytocentrifuged at 500 rpm for 2 min in a Shandon (Pittsburgh, PA) Cytocentrifuge (model Cytospin 2) and stained with Romanovsky (Diff-Quik; Baxter, McGaw Park, IL) and Jenner-Giemsa. For differential cell counts, a total of 500-1,000 cells was enumerated. Epithelial cells, macrophages, lymphocytes, neutrophils, and eosinophils were identified.

Permeability of the Blood-Airway Lumen Barrier

Plasma protein leakage into the airway lumen was analyzed by measuring albumin (Alb) and ₂-macroglobulin (A2M) in BAL fluid and paired serum samples.

To account for variable dilution of epithelial lining fluid (ELF) during lavage, we calculated the concentrations of proteins in the ELF as described by Rennard and coworkers (14). In this method it is assumed that the concentration of urea in the blood is in equilibrium with that in ELF. The concentration of a protein in ELF (cProtein) can be estimated by the formula: [cProtein] = ([protein] in BAL fluid) × ([urea] in serum)/([urea] in BAL fluid).

By comparing the levels of a large protein (A2M, 725 kD) with that of a smaller protein (Alb, 67 kD) it is possible to measure protein leakage and size selectivity of the blood-airway lumen barrier, without the influence of variable dilution and without the need for the urea correction procedure. We hereto calculate the relative coefficient of excretion (RCE): RCE = QA2M/QAlb, where Qprotein equals ([protein] in BAL fluid)/([protein] in serum) (15).

Albumin was measured by an immunoturbidimetric assay with a Cobas Bio analyzer (Roche Diagnostics, Nutley, NJ). Antiserum for albumin was obtained from Dako (code A001; Glostrup, Denmark). As a standard we used N protein standard serum for nephelometry (Behring, Marburg, Germany). The levels of A2M were measured by enzyme-linked immunosorbent assay (ELISA) (4).

Urea in serum and BAL fluid was determined by the glutamate dehydrogenase method, using an Eppendorf Epos 5060 GT analyzer (Eppendorf, Hamburg, Germany) (18).

Statistical Analysis

The quantification of all parameters analyzed was performed in a blinded manner and the treatment code was not broken until all results were submitted for statistical analysis. Mean diary card symptoms, use of rescue medication, and morning and evening PEF were calculated for the 4-wk run-in period (baseline), and for three 4-wk treatment periods. In addition, for morning and evening PEF, the first 4 wk of treatment were analyzed separately. For the comparison of the change from baseline (percentage for lung function and absolute for other parameters) of diary card variables and lung function parameters, we used repeated measurement analysis of variance (RMANOVA). To obtain a normal distribution the data on changes in PC₂₀histamine and percentage eosinophils were transformed logarithmically.

Comparisons within groups were done using the paired t test. Differences between the treatment groups were analyzed using the unpaired t test. Analysis of covariance was used to correct for differences in baseline values. Spearman's correlation coefficients (Rs) were used to assess relations between variables. A two-tailed p value  0.05 was considered significant. RMANOVAs were done using BMDP statistical software (Berkeley, CA). Other statistical analyses, including analysis of covariance, were performed with the SPSS/PC+ statistical package (version 6.1; SPSS, Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thirty-three patients agreed to participate in the study and started the run-in period, after having stopped their regular medication for 2 wk; three did not fulfill the inclusion criteria during the run-in period. Two of them had an FEV₁ < 60% predicted at Visit 1 and 2, and the other had an exacerbation, requiring additional medication. Thirty patients were randomized to receive FP (n = 15) or Sb (n = 15). At the start of the treatment period no significant differences were observed in baseline characteristics between the two treatment groups, with respect to patient demographics and pulmonary functions (Table 2).

In the FP group one patient was withdrawn for noncompliance after 4 wk of treatment (failure to attend scheduled visits). In the Sb group one patient did not tolerate the bronchoscopy and another patient refused the second bronchoscopy. Remaining investigations were completed successfully in these two patients. In neither of the two groups were patients withdrawn for worsening of asthma. No drug-related side effects were observed in the FP group. In the Sb group, two patients complained of a persistent tremor throughout the treatment period.

Effect of Corticosteroids on Primary Outcomes

Bronchial hyperresponsiveness. Paired analysis of bronchial hyperresponsiveness was possible in 14 patients of the FP group and 15 patients of the Sb group. Within the FP group the PC₂₀histamine increased significantly from baseline (mean increase, 2.4 doubling doses; p < 0.0001). After 12 wk of treatment, the inhalation of the histamine diluent (0.9% NaCl) caused a fall in FEV₁ of more than 20% of the baseline value in two patients from the Sb group. They had pretreatment PC₂₀histamine values of 0.18 and 0.12 mg/ml, respectively. In the statistical analysis their posttreatment PC₂₀histamine was allotted 0.015 mg/ml, which represents half of the lowest concentration of histamine used. Within the Sb group the PC₂₀histamine decreased significantly from baseline (mean decrease, 0.9 doubling doses; p = 0.03). The difference between the treatment groups was significant (p < 0.0001) (Figure 1).

View larger version (17K): [in this window] [in a new window]   Figure 1.   PC20histamine before and after treatment in the fluticasone group and in the salbutamol group. Mean change in the fluticasone group, 2.4 doubling doses (p < 0.0001); mean change in the salbutamol group, 0.9 doubling doses (p = 0.03). Significance of the difference in the mean changes: p < 0.0001.

Plasma protein leakage. Paired analysis of plasma protein leakage was possible in 12 patients from the FP group and 13 patients from the Sb group. Two BAL fluid samples from the FP group (one pre- and one posttreatment) and one from the Sb group (pretreatment lavage of the patient who refused posttreatment bronchoscopy) contained more than 2.5 × 10⁵ erythrocytes per milliliter, and were therefore excluded from analysis of plasma protein leakage.

The percentage of BAL fluid recovered differed neither within treatment groups nor between the two treatment groups (percentage recovery before and after treatment, mean ± SEM: FP group before, 66.5 ± 15.2; after, 74.2 ± 12.0; Sb group before, 71.1 ± 18.9; after, 69.0 ± 14.5; p > 0.1).

Levels of proteins in BAL fluid and in ELF from the two patient groups before and after treatment are shown in Table 3.

Protein levels in BAL fluid. At baseline, mean levels of A2M were similar in both groups, but levels of Alb were higher in the Sb group than in the FP group (p < 0.01). Within both groups no significant changes were observed in the BAL fluid levels of Alb or QAlb and A2M or QA2M; neither did the differences in changes between the two groups reach significance. Analysis of covariance with baseline protein level as a covariate did not change the results for Alb. However, changes in QA2M between the FP and Sb group were significantly different (p = 0.03).

Protein levels in ELF (urea correction). At baseline, mean levels of cA2M were similar in both groups, but the level of cAlb was higher in the Sb group than in the FP group (p < 0.01) (Figure 2A). Levels of cAlb did not change significantly during treatment in either group, but a significant decrease was observed in the cA2M (p = 0.05) in the FP group only (Figure 2B). The difference between the treatment groups did not reach significance, either corrected or uncorrected for baseline values.

View larger version (13K): [in this window] [in a new window]   Figure 2.   (A) cAlb, (B) cA2M, and (C) RCE before and after treatment in the fluticasone group and in the salbutamol group.

Relative coefficient of excretion. At baseline, the RCE was not different between the two treatment groups. In the FP group a significant decrease was observed in the RCE after treatment (p = 0.03). No significant change was observed in the Sb group. The difference between the two treatment groups was significant (p = 0.02) (Figure 2C). Correction for baseline levels of Alb and A2M did not change this result (analysis of covariance [ANCOVA] with baseline protein levels as covariates: difference between groups, p = 0.03).

Cell number and differentials. At baseline, no differences were observed in cell numbers and differentials between the two treatment groups. The percentage of eosinophils decreased in the FP group during treatment (p = 0.02), but there was no significant change in the Sb group (Table 4). The difference in change in the percentage of eosinophils (relative to total leukocytes) between the two treatment groups did not reach significance (p = 0.09). No significant changes were observed in the percentage of lymphocytes, neutrophils, or macrophages in either treatment group.

Correlations between Primary Outcomes

Baseline levels of Alb and cAlb correlated significantly with baseline levels of A2M and cA2M, respectively (n = 27, Rs = 0.48, p = 0.01 and Rs = 0.56, p = 0.002). No significant correlations were observed between baseline levels of Alb and lung function parameters (histamine threshold and FEV₁ [% predicted]). Baseline levels of A2M and cA2M correlated with the FEV₁ (n = 27, Rs = 0.39, p < 0.05 and Rs = 0.37, p = 0.06, respectively), but not with the PC₂₀histamine (p > 0.1). The baseline percentage of eosinophils in BAL fluid correlated significantly with the FEV₁ (n = 29, Rs = 0.56, p < 0.01) and with the PC₂₀histamine (n = 29, Rs = 0.47, p < 0.01).

The change in cA2M during treatment correlated significantly with the change in FEV₁ (n = 25, Rs = 0.44, p = 0.03), but not with the change in PC₂₀histamine (n = 25, Rs = 0.33, p = 0.11). The change in the RCE during treatment correlated significantly with the change in FEV₁ (n = 25, Rs = 0.54, p < 0.01) and with the change in PC₂₀histamine (n = 25, Rs = 0.48, p = 0.02). Changes in the percentage of eosinophils during treatment did not correlate significantly with the change in FEV₁ (Rs = 0.31, p > 0.1) or with the change in PC₂₀histamine (Rs = 0.23, p > 0.1).

Effect of Fluticasone on Secondary Outcomes

Spirometry. At baseline, no significant differences were observed for any of the parameters between the two treatment groups. Within the Sb and FP groups, there was no significant change in FEV₁, MMEF, FEF_25-75%, and PEFR during treatment.

Daily symptoms, rescue medication, and peak flow. At baseline, no significant differences were observed for any of the parameters between the two treatment groups. Within the FP group, morning and evening PEF increased from baseline from 2 wk of treatment onward. The average change from baseline (percentage) was significantly different between the treatment groups (RMANOVA: both p < 0.001). The mean difference between the two treatment groups was 11.6% (SEM, 3.0%) for the morning PEF and 8.8% (SEM, 2.4%) for the evening PEF.

The overall change from baseline in daily symptoms during treatment within the FP group was (mean ± SEM): nighttime: 0.17 ± 0.06, p < 0.01; daytime: 0.18 ± 0.09, p = 0.03. In the Sb group a significant increase was observed in nighttime symptoms, and a trend to an increase in daytime symptoms (increase from baseline, mean ± SEM): nighttime, 0.15 ± 0.05, p < 0.01; daytime, 0.15 ± 0.08, p = 0.06. The change in nighttime and daytime symptoms in the FP group differed from that in the Sb group (RMANOVA, FP versus Sb: nighttime: p < 0.01; daytime: p < 0.001).

The nighttime and daytime use of rescue medication decreased significantly within the FP group (mean change in percentage of days with rescue medication: nighttime, 22% (p < 0.01); daytime, 23% (p < 0.01). Within the Sb group no significant changes were observed (nighttime, 2%, p > 0.1; daytime, 4%, p > 0.1). The mean differences of these changes between the treatment groups were significant (daytime, p = 0.04; nighttime, p = 0.03).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this double-blind, parallel-group study we have investigated the efficacy of 12 wk of treatment with inhaled fluticasone propionate on plasma protein leakage into the airway lumen. We found that the inhaled glucocorticosteroid therapy significantly reduced plasma protein leakage as deduced from the decrease in QA2M and cA2M, and in the relative coefficient of excretion (RCE) of A2M to albumin.

The analysis of protein levels in BAL fluid is complicated by variable dilution of the epithelial lining fluid by the BAL procedure, which may differ between patients and within patients. One approach to take into account the variable dilution, proposed by Rennard and coworkers (14), is to use a correction factor based on the concentration of urea in serum and BAL fluid to calculate protein levels in ELF. An important drawback of this procedure is that there may be influx of urea into the BAL fluid during the lavage procedure. This has been noticed by Rennard and colleagues (14), and was further substantiated by our own studies (18) and by others (19, 20). Actual proof has been provided with the use of radiolabeled urea in a lavage procedure with 60-ml portions of saline (21). Marcy and coworkers (20) showed that the application of lavage with 60-ml portions made the urea correction procedure less accurate than with 20-ml volumes. This is probably due to a longer dwell time of the larger, 60-ml portions. During repeated lavage with similar recovery, 60-ml portions as opposed to 20-ml portions may result in a larger fluid reservoir in the lung, into which influx of urea can take place. In our studies we have always used 20-ml saline portions. We did find evidence of an enhanced influx of urea in patients with increased permeability of the respiratory membrane (18). On the other hand, we did not observe an increase in urea influx fluid during BAL procedures with deliberately extended dwell time between portions (18). Taken together, there is no doubt that urea influx into BAL fluid may occur during the lavage procedure, but this phenomenon appears to be positively associated with the degree of mucosal permeability. Thus, the urea correction procedure in our study may have resulted in an overestimation of the amount of ELF sampled and an underestimation of the protein levels in ELF. On this basis we conclude that if there was any difference between influx of urea before and after corticosteroid treatment, it was probably less after treatment. Thus, corticosteroid treatment decreases the leakage of A2M from the blood into the airway lining fluid, and the urea correction procedure may have resulted in an underestimation of this decrease.

As a completely different approach to obtain information on plasma protein leakage the RCE has been proposed (4, 15- 17). Earlier, a similar approach was made with the distribution coefficient (22). We have used the RCE, on the assumption that a disturbance of the blood-to-air space barrier results in relatively more leakage of large proteins than of smaller proteins (22). The RCE is the ratio of the blood-to-BAL fluid gradient of a large protein (A2M) to the gradient of a small molecule (Alb). With the RCE the variable dilution is corrected without the use of urea concentrations (15, 17, 18). In addition, the RCE provides information on changes in the size selectivity of the blood-to-air space barrier for proteins. In agreement with the urea-corrected cA2M concentrations in BAL, the RCE of A2M to albumin likewise decreased after therapy with corticosteroids.

Thus in this study two independent calculations, which both take into account the variable dilution of BAL fluid, have provided evidence of the inhibition of plasma protein leakage into the airway lumen by inhaled corticosteroids. Levels of albumin did not decrease significantly after corticosteroid treatment. It is probable that the smaller size of albumin, and as a consequence its lower gradient from blood to airway lining fluid, makes it less suited for measuring changes in the blood-to-air space barrier (22).

There is now considerable evidence that measurement of A2M is a valid concept in analyzing the kinetics of plasma exudation after allergen challenge (23). When corticosteroids are applied for a longer period and BAL proteins are analyzed in the steady state it cannot be excluded that the results are partly due to inhibition of local synthesis of A2M. However, there is no evidence of significant local production of A2M in patients with asthma. Furthermore, the significant correlations between albumin and A2M in BAL fluid and ELF are in line with lumenal entry of A2M owing to leakage and argue against significant local synthesis of A2M.

To our knowledge this is the first double-blind study demonstrating a decrease in plasma protein leakage during treatment with inhaled corticosteroids, and it confirms our previous results in open studies in which patients were treated with budesonide, and dexamethasone (3, 26).

For ethical reasons salbutamol was used instead of placebo in the control group. In the FP group ₂-agonist escape medication was used less than twice a week. Patients in the Sb group used salbutamol twice daily and in addition showed a significantly higher use of salbutamol as escape medication. The ₂-agonists have been reported to exhibit inhibitory effects on mediator-induced airway microvascular leakage in animals, although these observations have not been confirmed in human studies so far (27, 28). On the basis of these results the effect of corticosteroids on plasma protein leakage is possibly underestimated by the comparison with ₂-agonists.

Within the corticosteroid group, an increase was observed in the PC₂₀histamine values, with a mean of 2.4 doubling doses. This confirms the study by Weersink and coworkers, who demonstrated an increase in PC₂₀histamine after 6 wk of treatment of patients with asthma with a similar dose of fluticasone (29). In the patients receiving salbutamol, a worsening of the bronchial hyperresponsiveness and other clinical parameters was found. This was not limited to the patients who were on maintenance treatment with inhaled corticosteroids until 6 wk before the start of the study treatment. Although the issue of worsening of bronchial hyperreactivity during regular treatment with ₂-agonists is far from settled (30), the results of this study appear to be in support of such a phenomenon.

Before treatment, a significant correlation was observed between the baseline histamine threshold and the percentage of eosinophils in BAL fluid. Changes in the histamine threshold during treatment did not correlate with changes in the percentage of eosinophils, but did show an association with changes in the RCE. Although these associations do not confirm a causative role, they support the concept that both inflammatory cells and plasma protein leakage are involved in determining the degree of nonspecific bronchial hyperresponsiveness in patients with asthma.

We have confirmed that fluticasone propionate (250 µg, twice a day) is clinically effective in patients with asthma from 2 wk onward. We conclude that after 12 wk of treatment, a decreased lumenal entry of high molecular weight plasma proteins, like A2M, is observed with a significant improvement in the size selectivity of the blood-airway lumen barrier. Furthermore, changes in histamine threshold correlated significantly with changes in the RCE. The analysis of leakage of high molecular weight plasma proteins may therefore be a more sensitive parameter in BAL fluid for determining antiinflammatory properties of inhaled corticosteroids, than changes in levels of Alb or the number of eosinophils.