INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality (1). Since many patients show some reversibility of airflow limitation (2), bronchodilator drugs are the cornerstone of therapy for COPD. Inhaled short-acting -receptor agonists improve airflow limitation, exercise capacity, functional capacity, and quality of life in patients with COPD (5). Although studies of inhaled -agonists have shown that large doses are required for maximal bronchodilation (3, 8), data suggest minimal if any benefit in terms of functional status, dyspnea, or quality of life from administering regular doses above the equivalent of 200 µg of albuterol four times daily (11).

The merits of regular versus as-needed short-acting -agonist therapy for airflow obstruction have generated considerable controversy. Studies have primarily examined regular versus as-needed -agonists in asthma, but little is known about their comparative effects in COPD. The controversy began with a randomized crossover trial of regular versus as-needed inhaled fenoterol in asthmatic patients (14). Of 64 patients in the study, 57 had a significant difference in asthma control during the regular fenoterol versus the placebo period. In 17 (30%) patients, asthma was better controlled during regular treatment, whereas in 40 (70%) patients, control was better during use of placebo (p = 0.003). A second double-blind randomized trial of regular versus as-needed albuterol as monotherapy in 255 patients with mild asthma showed no benefit of regular administration of albuterol with respect to peak flow, FEV₁, symptoms, quality of life, supplemental use of albuterol, or airway responsiveness, the last of which worsened significantly during the early part of the study (15).

A much shorter randomized crossover trial of regular albuterol at 800 µg daily for 2 wk versus albuterol as-needed in 341 asthmatic patients found that peak flow rates were no different with the two regimens, but asthma symptoms and supplementary bronchodilator use were significantly less severe or frequent when albuterol was given regularly (16). The most recent trial of inhaled -agonist therapy randomized 983 mildly asthmatic patients receiving 2,000 µg or less of inhaled corticosteroids receive either 1,600 µg albuterol regularly or albuterol as-needed (17). Morning peak flow and exacerbations were similar in the two groups. The peak evening flow was significantly greater and the use of rescue bronchodilators was significantly less in the regular-use group, but overall this group used substantially more bronchodilator.

The only study to date that may provide some data for effects of inhaled -agonists therapy in COPD was a trial by van Schayck and colleagues (18) in patients with moderate airflow obstruction caused by asthma or bronchitis. Patients received either regular or as-needed bronchodilators, with regular treatment consisting either of 1,600 µg albuterol or 160 µg ipratropium bromide daily in a crossover design. Among 144 patients completing the 2-yr study, quality of life in the two study groups was similar, but the decline in FEV₁ was 72 ml/yr in the regular treatment group and 20 ml/yr in the as-needed group, irrespective of the drug used (p < 0.05). The difference in decline in FEV₁ was similar in patients with asthma and those with bronchitis. However, a 4-yr follow-up study of the non- steroid-dependent subpopulation showed no difference in peak flow rates, exacerbation rate, or reported health between patients using bronchodilators regularly and those using them on an as-needed basis (19).

To test the hypothesis that patients with COPD as distinct from asthma may derive no benefit from regular versus as-needed inhaled short-acting -agonist therapy, we conducted a randomized, double-blind, crossover trial to evaluate the effects of regular versus as-needed albuterol on spirometric measures, peak flow, functional exercise capacity, and health-related quality of life of patients with moderately severe COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Setting

Patients were recruited from respirology practices and rehabilitation programs in Hamilton and Toronto, Ontario. Institutional ethics review boards approved the protocol. All patients gave written consent for participation in the study.

Study Eligibility

We evaluated patients aged  50 yr who had a clinical diagnosis of COPD (20) and a smoking history of > 20 pack-years, FEV₁  70% predicted, and FEV₁/VC ratio of  0.7 after inhalation of 200 µg albuterol. Eligible patients identified three activities of daily living in which they were limited by exertional dyspnea. We excluded patients with a diagnosis of asthma and those who had an increase of  15% over baseline in FEV₁ after inhaling 200 µg of albuterol, again to exclude patients with asthma. We also excluded patients who had exertional dyspnea caused by other conditions; limited exercise capacity for reasons other than dyspnea; inability to complete quality-of-life questionnaires; and tachyarrhythmias requiring treatment; as well as those in whom oral corticosteroids or theophylline preparations could not be discontinued during the run-in phase of the study; those hospitalized within 2 mo of the study; and those with a change in respiratory medications in the previous month.

Eligible patients made an initial visit during which we conducted spirometry at  4 h after administering bronchodilators or caffeine; this approach was maintained throughout the study. Patients whose metered dose inhaler (MDI) technique was not optimal were given feedback until they mastered the technique. If patients were not receiving inhaled ipratropium bromide, or were receiving < 160 µg ipratropium bromide daily (20 µg per puff in 2 puffs four times daily), this regimen was started and patients were followed for a 1-mo run-in phase. If patients were not receiving inhaled corticosteroids as maintenance therapy, or were receiving < 1,000 µg beclomethasone daily (250 µg per puff in 2 puffs twice daily) or < 800 µg budesonide daily (200 µg per inhalation in 2 puffs twice daily), beclomethasone at 1,000 µg daily or budesonide at 800 µg daily was started and continued for a 2-mo run-in period. During this run-in period we also discontinued patient's use of theophylline preparations.

At the beginning of the run-in phase, eligible patients received a symptom and peak flow diary that they completed until the midpoint of their run-in period, at which time they returned for repeat spirometry and confirmation of eligibility. Patients who took < 75% of prescribed bronchodilators as measured by canister weights received reminders about the importance of adhering to the study protocol. At the end of the run-in phase, patients dependent on theophylline, those intolerant of ipratropium bromide or inhaled corticosteroids, or those who were less than 75% compliant were not considered for randomization.

Randomization and Blinding

Patients who completed the run-in phase of the study were stratified by center and randomized in blocks of four in a crossover design to receive either active or placebo albuterol (100 µg per puff in 2 puffs four times daily) for 3 mo each. All caregivers and research personnel were unaware of the randomization schedule and block size. Patients, caregivers, and research personnel measuring outcomes were blinded to patient's treatment allocation.

Drug Preparation, Dispensing, and Administration

Patients received active or placebo albuterol from coded boxes stored at each study site. Each patient had two treatment periods, each lasting 3 mo. In one treatment period, patients self-administered, from an MDI, 2 puffs four times daily of albuterol in addition to open-label albuterol as needed. In the second treatment period, patients received matching placebo in 2 puffs four times daily in addition to open-label albuterol as needed. In each period, at least 1,000 µg (2 puffs twice daily) of beclomethasone or equivalent and 160 µg (2 puffs four times daily) of ipratropium bromide were administered daily. All inhaled medications were delivered from a valved holding chamber (Aerochamber; Trudell Medical, London, ON, Canada). Glaxo Wellcome (Mississauga, ON, Canada) supplied MDI canisters of albuterol, placebo, and beclomethasone. Boehringer Ingelheim, Burlington, ON, Canada) supplied the ipratropium inhalers.

Compliance, Cointervention, and Exacerbations

A research assistant monitored inhaler technique throughout the study. Drug canisters were weighed before dispensing them and at each 6-wk and 3-mo visit. Patients recorded their medication use in daily diaries. We requested that patients not alter their exercise programs for the duration of the study.

We defined an exacerbation of COPD as a respiratory illness during which the patient began taking prednisone or antibiotics, or showed a substantive increase in use of as-needed -agonists. Exacerbations were treated according to a predefined protocol including oral prednisone at 40 to 60 mg or an intravenous equivalent for 5 d, with dosage tapered rapidly over 2-wk as tolerated, intensified bronchodilator therapy, and at least 10 d of antibiotic therapy. If patients were hospitalized, the study continued as soon as they felt that they had returned to their usual state of health. If study medications were discontinued during hospitalization, patients resumed taking them upon discharge. If a 3-mo study treatment period ended during an exacerbation, patients continued in the same treatment arm until they recovered, and were then crossed over to the second arm. Data from any treatment period in which such an illness occurred were included in the data analysis.

Measurements

Patients attended clinic twice in each treatment period, at 6 wk and at the end of each 3-month treatment period. Primary outcomes included spirometric measures (21), results of the 6-min walk test, and quality-of-life measures. At each visit, patients completed three forced expirations into a water spirometer (Warren E. Collins, Braintree, MA) before and after taking 2 puffs of active albuterol. We measured FEV₁, slow vital capacity (SVC), and peak expiratory flow, using the best result for each test. Patients recorded the best of three peak flow measurements (Personal Best; Health Scan Products Inc., Cedar Grove, NJ) in a daily diary in the morning and in the evening before using their inhaled medications (ipratropium bromide, corticosteroids, and study medication). Patients measured predrug peak flow, then took their medications in the order of study albuterol (active or placebo), ipratropium, and corticosteroids. Postdrug peak flow measurements were recorded 30 min after taking of the coded medication. If patients needed their bronchodilators in the morning before taking their peak flow measurements, peak flow was measured 30 min after bronchodilator use.

The 6-min walk test was completed in a quiet, closed corridor 30 m long (22), with standard encouragement (23). At the end of the test, patients rated their maximum dyspnea during the walk on a modified Borg scale. The Chronic Respiratory Questionnaire (CRQ) provided a responsive and valid measure of dyspnea, fatigue, mastery (ability to cope with COPD) and emotional function (24). Patients completed the CRQ at the 6-wk and 3-mo visits. Questions about dyspnea in the day-to-day activities section of the CRQ were included in the daily diaries. We also administered the Medical Outcomes Survey Short Form-36 (SF-36), a generic, health-related quality-of-life instrument (25).

Secondary outcomes included cough and sputum production, side effects of albuterol, and use of as-needed albuterol. Patients rated cough and sputum production on a seven-point scale in their diaries (1 = none, 7 = maximum); measures were also obtained by interviewer-administered questionnaire at the 6-wk and 3-mo visits. Patients counted their as-needed puffs of medication each day, and recorded symptoms that they considered might be caused by study medications. During study visits, we asked patients whether they experienced any adverse reactions to the study medications. Canister weights of all bronchodilators were recorded at each visit.

Statistical Analysis

Our primary analysis for each outcome utilized an analysis of variance model appropriate for crossover designs (26). For data collected at the end of each study treatment period (spirometric data, 6-min walk distance, quality of life scores) we used the mean of the 6-wk and 12-wk observations in one analysis. The results were very similar to those of an analysis done with end-of-period data alone, and we report the latter. Use of inhalers was calculated from canister weight changes. Mean data for the last 6 wk of each study period are reported. For symptoms and peak flow measurements recorded in the patient diary, we used patients' mean values across each entire study period in our analysis. For all outcomes that patients recorded in their diaries, we used only values during the last 2 wk of each period, in order to minimize carryover effects. We formally tested for a treatment effect, a period effect, and an effect of interaction between treatment and period.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Randomization began in October 1994; the last follow-up visit was made in October 1996. Screening of medical records identified 851 patients with COPD, of whom 409 remained apparently eligible after contact with the attending physicians at their institutions, and 105 remained apparently eligible and interested after telephone contact for the study. Of the remaining 304 patients, 125 were not interested, 75 felt they were too sick to participate, 33 declined to change medication, 24 identified transportation problems, 19 felt they were too healthy to participate, 16 were ineligible, nine could not be located, and three had died. Of the 105 patients who attended at least one run-in visit, 19 proved ineligible and 13 declined participation.

Of the 73 patients who began the study, 11 failed to complete the run-in phase. Three of these 11 patients decided that the burden of the study was excessive, six proved unable to complete study requirements, and two became and remained unstable. Nine patients began or required an increase in regular ipratropium, 14 began or required an increase in inhaled corticosteroids, and three discontinued theophylline to meet study entry requirements. Of 62 patients who completed the run-in phase, nine did not complete the study. In the first study treatment period, two patients were receiving active medication (one of whom did not want to risk deterioration and one of whom was noncompliant) and three patients were receiving placebo (two of whom lost interest in the study and one of whom was noncompliant). In the second period, two patients were receiving active medication (one of whom was tremulous and one of whom believed she was receiving placebo) and two patients were receiving placebo (both experienced a severe exacerbation).

Table 1 summarizes the characteristics of the 53 patients who completed the study and of the 20 eligible patients who withdrew either during the run-in period or after starting to take blinded medication. Patients who withdrew were more likely to be female and to have lower (worse) scores on the dyspnea domain of the CRQ but higher (better) scores on the CRQ mastery and emotional function scales and on the SF-36 emotional function scale.

Table 2 summarizes the effects of treatment on spirometric measures and peak flow rates. For all but one measure, results were very similar whether patients were taking active or placebo medication, and there were no significant treatment effects. These analyses revealed no evidence of period effects or of a treatment-by-period interaction. Patient peak flows measured at home increased to a greater degree after the use of coded medication was begun when the medication was active rather than placebo medication.

Table 3 summarizes the effects of treatment on the 6-min walk distance, dyspnea following the walk, and quality-of-life measures. Scores on most measures showed very little difference between active and placebo periods, and no results approached conventional statistical significance. Although we found no treatment-by-period interaction, patients tended to do better during the first than during the second study period in all quality-of-life measures. For the dyspnea and fatigue domains of the CRQ, the difference between the first and second periods was statistically significant. The mean CRQ dyspnea score in the first period was 4.28, whereas that during the second period was 4.04 (p = 0.01); fatigue scores during the first and second periods were 4.33 and 4.07 (p = 0.02), respectively. Table 4 summarizes the effects on symptoms, and shows no evidence of a treatment effect.

Patients taking regular albuterol used twice as much albuterol as during the placebo period, at an average of 12.75 puffs of active albuterol daily (8.07 puffs of coded and 4.68 puffs of open-label medication), whereas patients taking as-needed albuterol used an average of 6.34 puffs of active albuterol daily (all open-label). Open-label use of albuterol was 1.7 puffs per day more during as-needed use, but overall albuterol use was halved. Patients did not use the study inhaler less often during placebo than during active-treatment periods (weight difference: 2.4 mg; 95% confidence interval (CI): 0.78 to 5.6; p = 0.14). This corresponds to an average compliance of 105% during active treatment periods and of 98% during placebo periods. The difference between use of ipratropium and corticosteroids during the active treatment and placebo periods was small and nonsignificant. Average compliance with use of ipratropium was 104% and that with corticosteroids was 102%.

Two patients experienced exacerbations requiring hospitalization during active medication periods. Another 33 patients experienced 44 exacerbations that did not require hospitalization; 22 of these exacerbations occurred while patients were receiving active medication and 22 while they were receiving placebo.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The primary finding of our study was that patients with COPD treated with inhaled ipratropium and corticosteroid do as well by using as-needed inhaled albuterol as by using regular albuterol, and require considerably less medication. Patients used a mean of less than 7 puffs of albuterol per day during the as-needed period of the study, whereas during the regular-use period they used almost 13 puffs per day with no clinically or statistically significant differences in outcome.

Strengths of our study include the concealed randomization of treatment; blinding of patients, caregivers, and study research personnel; and rigorous and comprehensive measurement of all clinically important outcomes. The duration of the study was sufficiently long as to permit observation of chronic effects of regular versus intermittent use of albuterol on physiologic, functional, and quality-of-life outcomes. The CIs around the differences between active treatment and placebo periods effectively exclude differences in favor of regular albuterol exceeding 9 ml in FEV₁, 20 ml in SVC, 11 m in the 6-min walk test, and 0.24 in any dimension of the CRQ (where 0.5 represents a small but clinically important difference) (27). In aggregate, these results effectively exclude an important benefit of regular use of albuterol in patients with stable COPD.

Several issues bear on the application of the study findings to practice. Ipratropium and inhaled corticosteroids were both standard therapy for COPD patients at our centers when the study began. Ipratropium remains a mainstay of COPD management (30, 31). The COMBIVENT Study shows that -agonist therapy is additive to the effects of ipratropium in patients with COPD (32), and we afforded patients in our trial the benefit of regular ipratropium use. Although studies of inhaled corticosteroids do not provide strong evidence of their clinical benefit in COPD, they are still in widespread use. We elected to standardize inhaled corticosteroids by initiating them in the few patients who had not already had corticosteroids prescribed, rather than discontinuing them in a large number of patients. It is possible that the results of our study cannot be generalized to patients who are not using regular ipratropium or inhaled corticosteroids. In the time since our study was begun, long-acting -agonist therapy has been shown to improve symptoms, lung function, and quality of life in asthma (33). The results of our study apply to regular use of short-acting -agonists, and cannot be generalized to patients receiving long-acting inhaled -agonist therapy.

Importantly, we enrolled only a modest proportion of potentially eligible patients; patients who withdrew differed to some extent from those who completed the study. Those who declined to participate might have done so because of a perception that they would suffer symptomatic deterioration if their regular albuterol were withdrawn. This possible heterogeneity of response suggests the use of randomized trials in individual patients (n = 1 trials) as the optimal approach to individualizing -agonist treatment for patients with COPD (34, 35).

We chose a crossover design because the use of patients as their own controls markedly enhances the power of a study, and because of our success with this design in a trial of as-needed versus regular -agonist administration in asthma (14). Although we found no treatment effect, we did observe a significant period effect for dyspnea and fatigue in the CRQ. Possible explanations for this include the progression of underlying disease with consequent functional limitations, or adverse effects of participation in a clinical trial that made substantial demands of its subjects over a period of 6 mo. The play of chance remains an alternative explanation.

Guidelines for the management of patients with COPD offer discordant recommendations regarding short-acting -agonist therapy. The American Thoracic Society recommends that COPD patients with persisting mild to moderate symptoms receive ipratropium aerosol plus a selective -agonist at 1 to 4 puffs four times daily, either as needed, or regularly (36). The Canadian Thoracic Society states that although 2 puffs of -agonist taken four times daily from an MDI is common initial treatment for COPD, greater benefit may be achieved by taking 4 to 6 puffs at each administration (37). The British Thoracic Society recommends short-acting -agonists on an as-needed basis (38), and states that although a possible detrimental effect of regular -agonist use has been suggested by one randomized trial, the evidence is not strong enough to advise against this practice.

Our results provide strong evidence that patients with COPD who are treated with inhaled ipratropium and an inhaled corticosteroid, although not harmed by regular -agonist therapy, do not gain symptomatic or functional benefit from regular short-acting inhaled -agonist use. Our results provide support for a clinical policy of using short-acting -agonist therapy only as needed for stable patients with COPD, and make it difficult to justify recommendations for the regular use of inhaled short-acting -agonists in COPD patients taking regular inhaled corticosteroids and ipratropium bromide.