INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although inhaled corticosteroid therapy has established benefit in treatment of asthma, use of inhaled corticosteroids in patients with chronic obstructive pulmonary disease (COPD) is less well supported. Improvement in airflow obstruction with inhaled steroids was demonstrated in previous studies that did

not distinguish between subjects with asthma and those with COPD (1). Studies of inhaled steroids in patients with irreversible airflow obstruction have provided conflicting results (4). Four trials of chronic inhaled steroid therapy in COPD have been published recently (8). All four demonstrated no significant decrease in the annual rate of decline in FEV₁ when compared with placebo, though in two of the four studies, the mean FEV₁ remained significantly higher throughout the trial in the steroid therapy group.

The safety of long-term, high-dose inhaled corticosteroids has not been well established. Inhaled steroids have been implicated in causing adrenal suppression (12), cataracts (13), glaucoma (14), and osteoporosis (15). Therefore, it is important to know if there is significant benefit for patients with irreversible airflow obstruction.

We assessed the pattern of prescription of inhaled beclomethasone in our institution in a retrospective study. We found that 35% of patients for whom inhaled beclomethasone was prescribed did not have a diagnosis of asthma or other conditions associated with asthma. Thus, it appeared that a significant number of patients might not require inhaled corticosteroid therapy.

Because of the cost and potential side effects of inhaled corticosteroids and because withdrawal of inhaled corticosteroids caused deterioration of lung function in asthmatics (16), we designed a trial to test the hypothesis that withdrawal of inhaled steroids in elderly patients with severe irreversible airway obstruction did not cause a deterioration in respiratory function. We compared a 6-wk withdrawal period from inhaled steroid with 6-wk of monitored inhaled steroid therapy. Beclomethasone dipropionate (BDP) by metered-dose inhaler (MDI), 336 µg/d, was studied since it is the most commonly prescribed inhaled steroid at that dose at our institution. We assessed a number of parameters of respiratory status, including spirometry, a questionnaire assessing quality of life, exercise capacity by a six-minute walk test, and sputum markers of airway inflammation. We report a prospective, double-blinded, randomized, placebo-controlled, crossover trial in patients with stable COPD to determine if withdrawal of their inhaled steroid has any effect on lung function, exercise capacity, respiratory symptoms, or quality of life.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was performed in the outpatient pulmonary department of a university-affiliated Veterans Administration Medical Center (VAMC) in Providence, RI. The study protocol was approved by the institutional ethics committee, and informed written consent was obtained from all study subjects.

In the retrospective study of prescription patterns of inhaled corticosteroids, we identified all patients with an active prescription for inhaled BDP from hospital pharmacy records. From this population, we reviewed the medical records of a random sample of 252 patients for predefined markers suggesting a diagnosis of asthma or reversible airways obstruction. We then assessed the percentages of patients who fulfilled one or more of these predefined criteria for asthma.

For the prospective study, patients were screened from a pharmacy listing of all patients with an active prescription for BDP by MDI. Medical records and pulmonary function test results were reviewed for identification of patients with stable, irreversible airflow obstruction.

Prospective Study Design

An overview of the double-blind, crossover prospective study design is shown in Figure 1. The subjects were interviewed and examined by a pulmonary physician initially and at each 3-wk follow-up during the study period of 12 wk. After discontinuing their prescribed steroid inhaler, subjects were randomized by a clinical pharmacist to either 6 wk of placebo MDI followed by 6 wk of BDP versus 6 wk of BDP followed by 6 wk of placebo MDI. The subject and the pulmonary physician were blinded to the treatment regimen. Placebo and drug MDI canisters were identical, and the placebo mist was flavored to make the treatments indistinguishable. Randomization was performed by the clinical pharmacist who randomized an odd number dice roll to placebo and an even number dice roll to drug. The clinical pharmacist

View larger version (46K): [in this window] [in a new window]   Figure 1.   Study design.

instructed and assessed inhaler technique with use of a spacer at each visit and weighed canisters to assess compliance.

Outcome Measures

Changes in respiratory symptoms were assessed by objective and subjective methods. Physical examination and inquiry into exacerbation of disease was performed by the pulmonary physician at each visit. Quality of life was assessed by the Chronic Respiratory Disease Questionnaire (CRQ) (17) at baseline and at each 3-wk follow-up along with spirometry prebronchodilator and postbronchodilator.

A standardized six-minute (6-min) walking test (18) followed by an evaluation of dyspnea was performed at the initial visit and at the 6-wk completion of each treatment phase.

A sputum analysis for total cell count and differential was performed at the same intervals (19).

Statistical Analysis

All data are expressed as the mean (± SE) unless otherwise stated. The means were compared using a two-tailed paired t test for normally distributed continuous data. The chi-squared test was used to compare frequency of exacerbations. A p < 0.05 was considered statistically significant for all tests. Instat (GraphPad Software, Inc., San Diego, CA) statistical package was the method of analysis.

Data were analyzed only from patients who completed both stages of the study, i.e., those who followed protocol; no patient data were analyzed according to intention to treat. (See online data supplement.)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We performed a retrospective study of the prescribing patterns for inhaled corticosteroids. We considered the following to be appropriate indications for the use of inhaled steroids: history of asthma, reversible airflow obstruction with postbronchodilator increase of  12% (and at least 200 ml) of the baseline FEV₁ or FVC, positive methacholine challenge, peripheral blood eosinophilia, current use of oral steroids, a history of steroid responsiveness, and an elevated serum immunoglobulin E (IgE) level. There were 661 patients receiving inhaled BDP at our institution at the time of the study. A random sample of 252 patients was selected to assess their indications for inhaled steroids. The number of patients meeting at least one of our criteria was 65% of the random sample, whereas 35% did not. Most of the patients meeting criteria had a diagnosis of asthma (56%) or evidence of PFT reversibility (39%). Remaining patients had peripheral blood eosinophilia, elevated IgE levels, or a documented response to inhaled steroids. Most of the 35% of the random sample that did not meet criteria had COPD with irreversible airflow obstruction.

Therefore, we found that a significant proportion of patients did not have a clear indication for the use of inhaled steroid therapy. We asked what would happen if we were to discontinue this therapy. We designed this prospective study to answer that question.

Of all eligible patients screened who were using a steroid inhaler for COPD, 24 consented to be in the study. The mean age for the 24 patients was 67 yr, with an age range from 40 to 79 yr. All patients were male and had at least a 20 pack-year tobacco history, with seven patients being current smokers. Patient demographics and baseline PFT measurements are shown in Table 1. The mean FEV₁ was 1.61 ± 0.1 L (47% of predicted). All patients were receiving BDP at the time of study entry and were also using albuterol and ipratroprium bromide delivered by MDI. The number of patients receiving a theophylline preparation was four. Home oxygen therapy was used by two patients, and nasal continuous positive airway pressure (CPAP) was being used by three patients for obstructive sleep apnea. A history of congestive heart failure or coronary artery disease was found in seven patients. These patients were not symptomatic from their cardiac disease during this trial.

Compliance

Compliance was assessed by the clinical pharmacist at each 3-wk follow-up visit. Canister weights were measured initially and throughout the 12-wk study. If compliance was considered low, the pharmacist would interview and advise the patient in order to increase their use of study inhaler. The pharmacist found that none of the study subjects had poor compliance with either placebo or BDP MDI. Most subjects adhered to their treatment regimen throughout the study, and canister weights showed them to use at least 85% or greater of the amount of inhaler prescribed.

Withdrawals

Of the 24 patients who were entered into the study, six were lost to follow-up after the initial visit. Five of them had difficulty obtaining transportation, and one underwent emergency

surgery soon after entering the study. There was an increased frequency of exacerbations of COPD while using a placebo inhaler as compared with inhaled corticosteroids (three of 18 versus zero of 16, placebo treatment period versus BDP treatment period); however, this difference was not statistically significant (p = 0.23). The three patients who experienced an exacerbation of their COPD had to be withdrawn from the study because of the severity of their symptoms during the exacerbation. Only two of these patients had spirometry done while using the placebo inhaler, with one of the two having a significant decline in the FEV₁ during the placebo period (30% decrease in FEV₁ with a total 220-ml decline from baseline).

Spirometry and Exercise Capacity

The results of spirometry measured during each treatment period are shown in Table 2. All spirometric values were prebronchodilator values. There was a significant decrease in the mean FEV₁ at 6 wk while using the placebo inhaler. There was no significant change in the mean FEV₁ while receiving BDP. There was a decrease in the mean percentage change in FEV₁ (FEV₁ at 6 wk while using the study inhaler-baseline FEV₁/ baseline FEV₁ × 100) for the study subjects during the placebo-treatment period as compared with the BDP-treatment period. There was no significant decrease in the mean FVC while using the placebo inhaler, or any difference between the mean percentage change in FVC for the patients during the active drug- and placebo-treatment periods. As this was a crossover trial, statistical analysis was performed to assess for both a period effect and a carryover effect. Neither were found to be present.

The individual variations in FEV₁ measured during the placebo and treatment phases, respectively, are demostrated in Figures 2 and 3. More patients had a decrease in their FEV₁ while using the placebo inhaler (six of 15, 73%) as compared with the BDP inhaler (six of 15, 40%).

View larger version (15K): [in this window] [in a new window]   Figure 2.   FEV1 at baseline and after 6 wk of placebo treatment. Each patient is plotted individually. All spirometric values were obtained prebronchodilator.

View larger version (16K): [in this window] [in a new window]   Figure 3.   FEV1 at baseline and after 6 wk of BDP treatment. Each patient is plotted individually. All spirometric values were obtained prebronchodilator.

There was no significant change in distance walked during the 6-min walk test during the placebo and inhaled steroid-treatment periods. Mean distance walked during placebo treatment was 734 feet (95% CI, 366 to 1,101 feet) compared with 713 feet (95% CI, 350 to 1,075 feet) at baseline (n = 7, p = 0.35). The mean distance walked during BDP treatment was 770 feet (95% CI, 371 to 1,169 feet) compared with 717 feet (95% CI, 370 to 1,065 feet) at baseline (n = 7, p = 0.25). Oxygen saturation with exercise did not differ significantly between baseline and 6 wk during either treatment period. The Borg scale assessment of dyspnea used with the walk test increased while receiving placebo (4.14 [95% CI, 2.90 to 5.39] versus 3.29 [95% CI, 2.59 to 3.98], 6 wk versus baseline, n = 7, p < 0.05), and there was a decrease while receiving BDP (3.29 [95% CI, 2.59 to 3.98] versus 4.57 [95% CI, 3.39 to 5.75], 6 wk versus baseline, n = 7, p < 0.05).

Subjective Measurements

The Chronic Respiratory Disease Questionnaire (CRQ) was compared during each six-week treatment period, and results were compared with the subject's baseline, which immediately preceded that treatment period. The CRQ was subdivided into four dimensions for assessment of dyspnea, mastery, emotional function, and fatigue. As shown in Table 3, there were no significant changes in the mean scores of the four dimensions of function measured by the CRQ during the placebo or inhaled steroid-treatment periods. CRQ scores showed minimal variation for most patients throughout the study period.

Sputum Analysis

Sputum was analyzed for total cell count and differential in five of the 15 patients. As shown in Table 4, the total number of cells in the sputum did not vary significantly during either the placebo or the inhaled steroid-treatment period. Also, mean total neutrophil counts did not differ significantly between placebo- and BDP-treatment periods as compared with their baseline values. Most patients did show an elevation in sputum neutrophils over what would be expected for normal, nonsmoking persons. Only three patients had eosinophils identified in expectorated sputum. There was an increase in the number of eosinophils at 6 wk while receiving placebo compared with that at baseline (4.86 [95% CI,  1.79 to 11.51] versus 0.21 [95% CI,  0.38 to 0.80], 6 wk receiving placebo versus baseline, n = 5, p = 0.12), with a decrease during the BDP-treatment period (0.42 [95% CI,  0.75 to 1.59] versus 2.38 [95% CI,  2.05 to 6.81], 6 wk BDP versus baseline, n = 5, p = 0.27).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This report describes the early effects of discontinuation of inhaled corticosteroids in elderly patients with severe irreversible airway obstruction. Our results did not support our hypothesis that withdrawal of inhaled steroids in elderly patients with severe irreversible airway obstruction would not lead to a deterioration in respiratory function. This study found that there was a deterioration in lung function when patients with COPD discontinued their inhaled steroids for a 6-wk period. This is consistent with the previously demonstrated deterioration in lung function after discontinuation of inhaled corticosteroids in asthmatic patients (16). We also found that dyspnea scores with exercise deteriorated after discontinuation of inhaled beclomethasone in patients with COPD, and that there was an increased frequency in exacerbations of COPD. To our knowledge, this is the first prospective, randomized study to look at the effects of discontinuation of inhaled steroids in patients with irreversible airflow obstruction.

Previous studies of the effects of inhaled corticosteroids in COPD had conflicting results. Auffarth and colleagues (6) found no improvement in lung function in patients with COPD receiving 1,600 µg/day of budesonide dipropionate, whereas Weir and colleagues (1, 2) found a significant improvement in lung function with BDP doses of 1,500 µg/d and 3,000 µg/d. Nishimura and colleagues (3), in a recent trial of patients with COPD, found a minority of patients (five of 30) with significant improvement in FEV₁ receiving 3,000 µ/d of BDP after a 4-wk treatment period. However, some of these responders had a positive bronchodilator challenge, as well as an elevated serum IgE level or eosinophil count suggestive of an asthmatic component to their airflow obstruction. Weiner and colleagues (7) found that in patients with COPD without significant bronchodilator response, there was no significant difference in lung function after 6 wk of 800 µg/d or 1,600 µg/d of inhaled budesonide versus placebo.

The incidence of acute exacerbations has also been studied in patients with COPD treated with a steroid inhaler (20). No significant difference was found in total number of COPD exacerbations when patients with COPD were using fluticasone propionate 500 µg twice a day compared with placebo over a 6-mo period. However, the severity of exacerbations was less in the fluticasone-treated group. This study also demonstrated an improvement in FEV₁, FVC, symptom scores and 6-min walking distance with fluticasone. These findings are in contrast to those of Bourbeau and colleagues (21), who also investigated the effects of inhaled steroids for 6 mo in patients with irreversible airway obstruction. They found no significant difference in the number or severity of COPD exacerbations, nor in the change in FEV₁ from baseline between the treatment and placebo groups; exercise capacity, dyspnea after exercise, and quality of life were also found not to differ between the treatment groups.

Long-term prognosis in patients with COPD receiving inhaled steroids was evaluated by a meta-analysis of previous inhaled steroid trials with exclusion of the asthmatic patients in each trial (22). Prebronchodilator FEV₁ was increased by 0.039 L/yr over a 2-yr period of treatment with inhaled corticosteroid, as compared with placebo. However, no benefit on the exacerbation rate of COPD was found. More recently, Pauwels and colleagues (8) reported that subjects with mild COPD who continued smoking exhibited a small onetime improvement in lung function, but the rate of decline in FEV₁ over 3 yr with long-term treatment with budesonide was not significantly different when compared with placebo. Another recent report that also looked at the long-term effects of inhaled steroids in mild and moderate irreversible airway obstruction came to the same negative conclusion, i.e., there was no significant difference in the rate of decline of FEV₁ compared with placebo during 3 yr of treatment (9). Of note, both of the latter studies involved patients with predominantly mild COPD. In contrast, both the ISOLDE trial from Britain (10) and the Lung Health Study (11) investigated the effects of chronic therapy with inhaled steroids in patients with moderate to severe COPD. Again, both demonstrated no significant difference in the rate of decline of FEV₁ with chronic therapy; however, they did demonstrate benefits in other clinically relevant outcomes. The ISOLDE trial reported a reduction in the number of exacerbations and a decrease in the rate of decline in health status (10). The Lung Health Study reported a reduction in respiratory symptoms, a decreased use of health care services, and improved airway reactivity (11).

We deliberately selected patients who not only had absence of clinical features of asthma, but also had no significant airflow reversibility with bronchodilators. A previous study has suggested that some patients with COPD and bronchodilator reversibility on PFT measurements are more likely to have improved lung function on inhaled steroids (7). Our retrospective review of the prescribing patterns for inhaled steroids at our institution showed that inhaled BDP was prescribed for a significant number of patients with irreversible airflow obstruction. Concerns about the widespread use of a therapy in which benefits in this population are controversial, has known side effects, and considerable cost, prompted this investigation.

A 336 µg dose per day of BDP was chosen since this was the most commonly prescribed dosage at our institution. A prior study comparing high to low dose inhaled steroid regimens in patients with COPD showed no difference in lung function (2). The duration of the placebo treatment period of 6 wk was felt to be adequate since airway reactivity has been shown to revert to baseline 1 wk after cessation of inhaled steroids (23). We also used a 6-wk period for the BDP treatment period as we felt that this was adequate to stabilize lung function.

There was very good compliance with the study inhalers in our subjects. Objective measurement of compliance with the study inhaler was performed by weighing canisters at each visit. Inhalation technique, with or without the use of a spacer device, was assessed at each visit and found to be adequate in all subjects. Therefore, differences in compliance or drug delivery were unlikely to have contributed to the deterioration of lung function found when inhaled steroids were withdrawn in these patients.

There was an increased frequency of COPD exacerbation on withdrawal of the inhaled corticosteroids, though the number of subjects was very small. All three exacerbations were of such severity that the patients had to be withdrawn from the study. Of these three patients, one had the lowest prebronchodilator FEV₁ of 0.73 L at the time of entry into the study. This patient was treated for a tracheobronchitis at the time of withdrawal from the study, and did have a 30% decrease in the FEV₁ after 6 wk of placebo inhaler. The second patient withdrawn because of worsening respiratory symptoms did not have a significant decline in lung function, but did have increased symptoms of fatigue as assessed by the CRQ. The third patient had completed the 6-wk BDP-treatment period and suffered an exacerbation 12 d into the placebo-treatment period. Because of transportation problems this patient did not have PFTs done at the time of the exacerbation. These findings are somewhat similar to those reported from the ISOLDE trial during the run-in phase (24). Of the 272 patients enrolled, 160 were receiving inhaled steroids. As part of the run-in phase, they had their inhaled steroids stopped. Over the next 7 wk, 38% of those who had been receiving inhaled corticosteroids suffered an exacerbation, versus 6% of those not previously receiving steroids. However, that report was an observational study, and thus has many inherent limitations; also, only the number of exacerbations was monitored, not lung function, quality of life, exercise capacity, etc.

Of the 24 patients who were entered into the study, there were eight who did not complete it. Six were lost to follow-up after the initial visit, and three were withdrawn because of significant exacerbations in their underlying COPD. All data that were analyzed were only those obtained from patients who completed the study. That is, no data were analyzed according to "intention to treat." We performed a simulation, extrapolating data from existing data for the "missing patients," to determine how these dropouts may have influenced the outcomes had they completed the study. We performed this simulation only for the outcome FEV₁. The three patients who were withdrawn because of exacerbations were given a final FEV₁ of 20% less than their initial FEV₁ for the placebo period and were given a FEV₁ unchanged for the period when they would have received BDP. For the other six dropouts, initially we matched them with the six best patient outcomes, i.e., assumed they had the same percentage change in their FEV₁ while receiving placebo and while receiving BDP as those six patients that fared best. Then we matched them with the six worst patient outcomes, i.e., had the greatest deterioration in their FEV₁ while receiving placebo. The results demonstrated that when matched with the six best patient outcomes, the FEV₁ decreased from 1.66 L (95% CI, 1.46 to 1.86) to 1.58 L (95% CI, 1.36 to 1.79) after 6 wk receiving placebo (p = 0.04), and increased from 1.62 L (95% CI, 1.39 to 1.85) to 1.63 L (95% CI, 1.42 to 1.84) while receiving BDP (p = 0.71). When matched with the six worst patient outcomes, the FEV₁ decreased from 1.64 L (95% CI, 1.45 to 1.83) at baseline to 1.48 L (95% CI, 1.28 to 1.68) after 6 wk receiving placebo (p < 0.0001). While receiving BDP, the FEV₁ increased from 1.57 L (95% CI, 1.35 to 1.78) to 1.61 L (95% CI, 1.42 to 1.81) after 6 wk (p = 0.24). Thus, using this simulation model, if the dropouts had completed the study, even if they had a favorable outcome, the overall group would have had a clinically and statistically significant decrease in the FEV₁ when the steroid inhalers were withdrawn.

Although the distance walked in a 6-min walk test did not change after withdrawal of inhaled corticosteroid therapy, dyspnea, as measured on the Borg scale, did increase. The CRQ has been found to be a useful tool in assessing quality of life in patients with COPD (17). No significant deterioration in any of the four dimensions (dyspnea, fatigue, mastery, and emotion) occurred when patients discontinued inhaled steroids.

Sputum markers of airway inflammation have been studied in patients with COPD receiving oral and inhaled steroids. Keatings and colleagues (25) found no significant decrease in the number of neutrophils in patients with COPD receiving inhaled or oral steroids. Confalonieri and colleagues (26) came to an opposite conclusion. In our patients, there was no significant difference in the sputum total cell count or neutrophil number during placebo- or BDP-treatment phases. Airway mucosal or sputum eosinophils have been suggested as markers for responsiveness to corticosteroids in patients with COPD (27). Three patients in our study had a small increase in sputum eosinophil count while treated with placebo. However, none of them displayed a significant improvement in FEV₁ with beclomethasone treatment or deterioration in FEV₁ after placebo treatment. Thus, although the number of sputum samples evaluated in this study was small, our results do not identify a subgroup of patients with COPD whose pulmonary function improves with inhaled corticosteroid therapy.

The decrease in FEV₁ on withdrawal of the steroid inhalers was small, though statistically significant. We believe we have underestimated the true reduction as patient compliance with their steroid inhalers on enrollment into the study was not evaluated. It has been well described that compliance with inhaled medication is poor (28, 29). Thus, some patients who were randomized to first receive placebo inhaler may have had a lower baseline FEV₁ than they would have had had they been taking their medication. This would lead to a reduction in the group baseline FEV₁, thus minimizing the decrease seen on withdrawal of the steroid inhalers. Similarly, all other parameters measured in this study are underestimated.

The response to a bronchodilator challenge is very commonly used to define "reversible" from "irreversible" airway obstruction. It is clear that this test is imperfect. There is a subgroup of patients with COPD, who have a negative bronchodilator challenge, but yet who respond to inhaled corticosteroids. It may be that this subgroup accounts for the conflicting results from the various studies looking at the benefits of inhaled steroids in patients with irreversible airways obstruction. Further investigation is warranted to help develop some method to allow us identify these "responders."

In summary, in a retrospective study, we found that a significant number of patients with irreversible airflow obstruction were prescribed inhaled corticosteroids. In a double-blind, prospective study, we assessed the effects of discontinuation of inhaled BDP in elderly male patients with severe irreversible airway obstruction. BDP was discontinued for 6 wk with comparison of outcome measures between placebo and BDP 6-wk treatment periods. We measured lung function, quality of life by CRQ, exercise capacity, and sputum markers of inflammation. We found that FEV₁ and dyspnea worsened and that there was a trend towards an increased frequency of exacerbations upon withdrawal of inhaled corticosteroids. Whether a similar response would be seen in women, or in patients with milder disease, is unknown and requires further investigation.

We conclude that previous studies assessing the efficacy of inhaled corticosteroids in irreversible airway obstruction have had conflicting results; however, the majority suggest a lack of efficacy. Our study investigated the consequences of discontinuing maintenance inhaled steroids in these subjects. In the short term, these patients suffered a deterioration in their respiratory status. Thus, patients with irreversible airway obstruction who have their maintenance inhaled steroids discontinued, initially need very close follow-up as there is a risk that they may suffer a significant deterioration.