INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: inflammation; nitric oxide; hydrogen peroxide; inhaled corticosteroids; COPD

There is substantial and growing evidence that endogenous nitric oxide (NO) plays a key role in the physiological regulation of airway inflammation and is implicated in the pathophysiology of various lower airway diseases (1). Exhaled NO (ENO) can be measured noninvasively and accurately in normal volunteers and patients with various respiratory disease. Patients with untreated asthma exhale higher levels of nitric oxide than normal subjects and these levels decrease significantly when inhaled corticosteroids are administered (2).

Patients with chronic obstructive pulmonary disease (COPD) who have stopped smoking also appear to have elevated ENO levels as compared with normal subjects (2, 5, 6), although levels are only mildly elevated as compared with these measured in patients with untreated asthma (2). ENO levels in COPD may be correlated with disease severity (7) and may increase further with exacerbation (8). Little is known of the impact of systemic inhaled corticosteroid therapy on ENO levels in COPD.

Hydrogen peroxide in the breath condensate of subjects with lung disease has also been described as a possible marker of airways inflammation and oxidation. Reactive oxygen- derived species (ROS) have been implicated in the pathogenesis and pathophysiology of tobacco smoke-induced chronic obstructive pulmonary disease (9). Hydrogen peroxide is a harmful ROS because it is relatively stable, it can cross membranes readily due to its small size and its lack of electrical charge, and it can generate the highly reactive hydrogen radical in the presence of superoxide anions and iron. Recent studies show that increased H₂O₂ levels are present in the frozen airway moisture exhaled by subjects with stable COPD (9) and that these levels are further increased at times of exacerbation (9). These results suggest a potential role for hydrogen peroxide measurements to characterize and quantify airway inflammatory processes in the investigation of lung disease.

Although inflammatory processes clearly play a role in the pathogenesis of tobacco-induced COPD, the role of inhaled corticosteroid therapy for COPD remains controversial. There is growing consensus that inhaled corticosteroids have little impact on airway obstruction in COPD, but there is some evidence that inhaled corticosteroid may reduce the frequency and/or severity of exacerbations in selected patients. Traditional oral steroids trials monitored by spirometry appear to be poorly predictive of benefit with inhaled antiinflammatory therapy (12).

Previous studies have described ENO or H₂O₂ levels in populations of patients with COPD, some of whom were taking inhaled corticosteroids and some of whom were not. We believe that the present study is the first randomized placebo-controlled trial that evaluates prospectively the effect of inhaled corticosteroid in those levels. We undertook the following study to determine if these newer noninvasive measures of lower airways inflammation are altered in patients with COPD who inhale conventional doses of beclomethasone.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Population

We recruited 20 stable former smokers with COPD from the ambulatory respiratory clinics of the University Health Network. All participants in the study gave written informed consent. The protocol was reviewed and approved by the Human Subjects Review Committee of the University of Toronto.

All patients had COPD as defined by American Thoracic Society (ATS) guidelines (13) and, in addition, reported smoking history of at least 20 pack-years with abstinence for at least 6 mo. Potential study subjects were excluded if they suffered from a respiratory tract infection within 6 wk of the study, had any significant clinical instability (as shown by need of increase of medication, emergency care, or hospitalization), suffered from other significant medical illnesses known to affect ENO measurements, used systemic corticosteroids within the month preceding recruitment, or had a known history of asthma.

Potential patients using inhaled steroids at the time of the recruitment were included in the trial only after inhaled corticosteroid had been withheld in unblinded fashion for 4 wk without clinically significant worsening.

Study Design

Patients were allocated in randomized double-blind, placebo-controlled, crossover fashion to two 2-wk treatment periods with twice daily inhaled beclomethasone or matching placebo, followed by a 2-wk washout period, after the second period treatment. Randomization was performed with sealed envelopes.

Study medication was inhaled beclomethasone administered by pressurized metered dose inhaler in a dosage of 250 µg per puff (Beclofort, Glaxo Wellcome Canada Inc, Mississauga, Ontario, Canada) or matching placebo administered in a dosage of two puffs twice daily. Each patient continued use of his or her usual medications in unchanged dosage during the trial except for any inhaled or systemic corticosteroid.

Noninvasive measurements and routine flow-volume spirometry were performed at baseline (recruitment) visits and at 1, 2, 3, 4, 5, and 6 wk follow-up visits. These visits correspond to posttreatment measurements at 1 and 2 wk of active or placebo therapy, and at the end of each week of the 2-wk washout period.

Pulmonary Function Tests and Noninvasive Measurements

All patients performed flow volume spirometry both before and after four puffs (100 µg per puff) of albuterol via pressurized dry suspension metered dose inhaler (Ventolin, GlaxoWellcome Canada, Inc., Mississauga, Ontario, Canada). Spirometry and diffusion (by single breath method-Morgan MDAS) were performed according to ATS standards (13).

We measured ENO in all subjects by the chemiluminescence technique (using a Sievers 280 NO Analyzer, Boulder, CO) as described in detail previously (14). We recorded the mean value of three ENO measurements with acceptable plateaus varying by less than 5%. The ENO measurement was always performed before spirometry (15) and at the same time of the day.

Breath condensate was collected using a cryocond breath condenser (Boeringher Ingelheim, Burlington, Ontario, Canada). The fluid phase in exhaled breath condensate was collected by cooling the exhalate, thereby causing condensation and the formation of a frozen sample. The exhaled breath was always collected before spirometric testing, at approximately the same time every week. The patients were asked to breathe at a normal frequency and tidal volume, wearing a noseclip, in a non rebreathing system, over 5 min. Frozen samples were collected in a foil at 30° C and was subsequently stored at 70° C. Duplicate samples were collected during a 15-20 min sampling session.

Hydrogen peroxide (H₂O₂) was measured as described previously (16). Briefly, 100 µl of 420 µM 3',3,5,5'-tetramethylbenzidine (dissolved in 0.42 M citrate buffer pH 3.8) and 10 µl of 52.5 U/L of horseradish peroxidase (Sigma P-2088, 5000 Units) were reacted with 100 µl of condensate for 20 min at room temperature. Subsequently, the mixture was acidified to a pH of 1 with 10 µl of 18 N sulfuric acid. The reaction product was measured spectrophotometrically at a absorbency of 450 nmg using an automated microplate reader (model Ceres UV90 HDI, Bio-Tek Instruments, Winooski, VT). The detection limit was approximately 0.1 µM H₂O₂.

Statistical Analysis

A test for interaction with order of treatment was done. We calculated the changes from baseline for each 2-wk treatment period (beclomethasone and placebo) and compared them using a pooled standard error as follows (mean difference [beclomethsone2_group1  baseline_group1]  mean difference [placebo2_group1  baseline_group1]  (mean difference [beclomethasone2_group2  baseline_group2]  (mean difference [placebo2_group2  baseline_group2]).

The t test p value for this test was not significant (p = 0.565), meaning that the effect was not significantly different between subjects who received placebo first as compared with those who received beclomethasone first, as there was no "carryover effect." The order of the administration had no significant effect; hence, all data were analyzed together.

The distribution of the data by level of time was tested for normality using the Kolmogorov-Smirnof test. If significant, data were transformed on the natural log scale and retested. When the log transformation failed to "normalize" the distribution of the data, further analysis was performed using nonparametric methods.

t tests and repeated-measures ANOVA were used for normally distributed data. When data were not normally distributed, we used the Mann-Whitney rank sum test. The Spearman or Pearson correlation coefficient was used to assess relationships.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nineteen patients completed the 6-wk study with one patient missing his final visit. The mean age of the patients was 69.1 ± 6.5 yr. They had smoked an average 51.9 ± 27 pack-years and had quit smoking 10.8 ± 6.4 yr before the trial. Mean postbronchodilator FEV₁ was 1.53 ± 0.67 L (55% of predicted), FVC 2.95 ± 0.99 L (88.4% of predicted), and diffusing capacity of the lung for CO (DL_co 15.7 ± 5.91 L/min (68% of predicted).

Exhaled NO and Breath Condensate H₂O₂

Median baseline ENO was 26.2 (19.3 to 54.8 ) ppb. ENO levels fell significantly with inhaled beclomethasone but were unaffected by inhaled placebo (see Figure 1). Median differences from baseline during beclomethasone therapy were 10.6 (p = 0.002) and 6.3 ppb (p = 0.013) after 1 and 2 wk, respectively. Median differences during placebo therapy were 1.5 and +1.4 ppb after Weeks 1 and 2, respectively. ENO was not significantly different from baseline during the washout period; median differences from baseline were 0.9 and 0 ppb by the end of the first and second weeks, respectively.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Box and whisker plots showing changes from baseline in ENO values after 1 and 2 wk of placebo (Pl 1, Pl 2), after 1 and 2 wk of beclomethasone (BDP 1, BDP 2), and after 1 and 2 wk of washout (wash 1, wash 2).

Median baseline breath condensate H₂O₂ concentration was 2.6 (1.9 to 3.5) µM and was not significantly affected by beclomethasone (0.2 µM) or placebo (0.3 µM) administration.

Pulmonary Function

FEV₁, FVC, FEV₁/FVC ratio, bronchodilator response, DL_co, and DL_co/VA were unchanged by beclomethasone or placebo inhalation. However, there was a nonsignificant trend toward increased FEV₁ following 2 wk of beclomethasone inhalation, FEV₁ increased by 0.043 ± 0.11 L, p = 0.09.

Relationship among Changes in ENO, H₂O₂, and Pulmonary Function

There was a moderate inverse correlation between changes in ENO and changes in FEV₁ after beclomethasone treatment (r = 0.50, p = 0.02). Similarly, there was a moderate inverse correlation between changes in H₂O₂ and changes in FEV₁ after beclomethasone treatment (r = 0.45, p = 0.04). There was no correlation between ENO or H₂O₂ and changes in any other conventional measure of lung function, including bronchodilator response.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data show that the inhalation of 1,000 µg/d of beclomethasone dipropionate by stable ex-smokers with chronic obstructive lung disease reduces their elevated baseline exhaled nitric oxide levels in as little as 1 wk. This decrease in a noninvasive marker of airways inflammation occurred in the absence of clinically or statistically significant changes in FEV₁. Hydrogen peroxide concentration in frozen breath condensate was unchanged by inhaled beclomethasone.

Our findings confirm our earlier report that ENO levels in stable ex-smokers with COPD are elevated slightly when compared with those found typically in healthy nonsmoking volunteers without asthma (2). Although this increase in ENO is modest when compared with the marked elevation sometimes measured in patients with asthma, their reduction by inhaled beclomethasone demonstrates that a corticosteroid responsive factor contributes to the exhaled nitric oxide levels seen in patients with COPD.

In patients with asthma, the marked elevations of ENO are thought to reflect eosinophilic airways inflammation and stimulation of inducible nitric oxide synthase. In COPD, airway and parenchymal inflammation is characterized by an increase of various cellular elements but predominantly macrophages, T-lymphocytes, and neutrophils (17). Inflammatory cells have a complex interrelationship and influence each other via a vast range of mediators and cytokines. As the disease process progresses, the role of the inflammatory cells may change (20). Recently, Silkoff and colleagues showed significant correlation between the changes in both ENO and sputum interleukin (IL)-8 and changes in sputum neutrophils (21). However, there may also be an increase of eosinophils in some patients, particularly at times of exacerbation. Levels of eosinophilic cationic protein (ECP) and EPO are elevated in the induced sputum of patients with COPD suggesting that eosinophils may be present but are degranulated. Thus, similar mechanisms may be responsible for the response to inhaled corticosteroids seen in patients with asthma and, to a lesser degree, in patients with stable COPD.

Our study differs from two earlier studies that considered the impact of corticosteroids on ENO values in COPD. Robbins and colleagues (22) used three measurement methodologies and did not find elevated ENO levels among patients with COPD nor could they detect an effect of corticosteroid use on ENO. However, the study was cross-sectional and observational; ENO levels might have been suppressed by baseline corticosteroid use or, in some patients, continued smoking. Maziak and colleagues (5) reported increased ENO values correlated with disease severity and/or exacerbations. These investigators reported the counterintuitive finding that corticosteroid use was associated with increased ENO levels. This might be explained as confounding by severity in a study that was also cross-sectional in design.

Although airways inflammation and its suppression by corticosteroids likely explain our present findings, we note that other mechanisms have been postulated to explain the elevated ENO levels seen in patients with stable COPD. In earlier cross-sectional studies (2, 5, 6), ENO levels in COPD have been correlated with pulmonary function indices of tobacco-induced lung damage such as FEV₁ and diffusion capacity. That is, patients with the greatest impairment show the greatest elevation in exhaled nitric oxide level. This could be explained by ventilation-perfusion (/) mismatch so that there is decreased scavenging of airway and alveolar nitric oxide by circulating hemoglobin. Nitric oxide has actually been used as a marker gas in the measurement of diffusion capacity (23). These findings need not be mutually exclusive; / mismatch and airways inflammation may contribute to the elevated nitric oxide levels measured in patients with stable COPD. The development of cor pulmonale may decrease exhaled NO levels, presumably as a reflection of endothelial injury (24). The degree to which these factors play a role may vary from patient to patient; COPD is a heterogeneous disorder.

We cannot rule out the possibility that small changes in lung function and / matching explain the suppression of nitric oxide we saw with beclomethasone inhalation. We think this is unlikely, however. The changes in FEV₁ were small. In all but 2 of our 20 patients the change in FEV₁ with beclomethasone treatment was less than 200 ml and could not be considered clinically significant. Although variability in airflow obstruction has been reported to correlate with increased ENO levels in COPD (25), our patients showed little response to therapy and we believe that we excluded patients with asthma from the trial. In addition to their lack of response to beclomethasone, at least 14 of our 20 patients had previously undergone 2-wk oral steroid challenges; all had negative responses using as a criterion a 200 ml increase in postbronchodilator FEV₁. Three patients had postbronchodilator improvements in FEV₁ greater than 200 ml at baseline. All of these patients had persisting airflow obstruction following bronchodilator, none had a history of asthma and had decreased diffusion capacity. Excluding the data from these three patients from our analysis did not alter our statistical conclusion. Finally, we note that there was no correlation between responsiveness to inhaled bronchodilators and changes in ENO with beclomethasone.

It has been suggested recently that the role for inhaled corticosteroids in COPD is not to produce acute spirometric improvement or to slow decline in lung function (26). Instead, at least two randomized, placebo-controlled trials have suggested that inhaled corticosteroids can reduce exacerbation frequency and/or severity (12, 26, 30). This potential benefit of inhaled corticosteroid therapy is most evident in patients with chronic cough and severe airflow obstruction. The measurement of ENO might identify a subset of patients most likely to have airway inflammation that is corticosteroid responsive (31, 32) and that may contribute to exacerbation frequency or severity if left untreated. Our findings are consistent with this hypothesis but prospective studies will be required to determine if there is such a role for nitric oxide measurement in COPD management.

It is likely that inflammatory mechanisms as well as oxidative mechanisms play an important role in the development and progression of COPD. Although our study demonstrated a decrease in ENO levels, hydrogen peroxide concentrations were unaffected by the administration of beclomethasone. Hydrogen peroxide is thought to be produced by a number of oxidative mechanisms (9, 11, 33). Neutrophilic processes that result in lipid peroxidation would likely be less responsive to corticosteroid administration than eosinophilic processes. The lack of change that we observed with inhaled corticosteroid intervention is consistent with this.

In summary, we have demonstrated in our previous cross-sectional study that ENO levels are elevated in patients with COPD. In this present study, we demonstrated in a prospective randomized, placebo-controlled, crossover trial that these elevations are responsive to the inhalation of corticosteroid. Eosinophilic or other corticosteroid-responsive inflammatory mechanisms are likely present in at least some patients with COPD. These observations suggest that measurement of ENO could prove useful in the management of patients with COPD by selecting those most likely to benefit from corticosteroid use. Such hypotheses must be tested by prospective clinical trials.